Preprint
Review
Smart Viniculture: Applying AI for Improved Winemaking and Risk Management
Altmetrics
Downloads
170
Views
112
Comments
0
A peer-reviewed article of this preprint also exists.
Submitted:
06 September 2024
Posted:
09 September 2024
You are already at the latest version
Alerts
Abstract
This review explores the transformative role of artificial intelligence (AI) in the entire winemaking process, from viticulture to bottling, with a particular focus on enhancing food safety and traceability. It discusses AI’s applications in optimizing grape cultivation, fermentation, bottling, and quality control, while emphasizing its critical role in managing microbiological risks such as mycotoxins. The review aims to show how AI technologies not only refine operational efficiencies but also raise safety standards and ensure traceability from vineyard to consumer. Challenges in AI implementation and future directions for integrating more advanced AI solutions into the winemaking industry will also be discussed, providing a comprehensive overview of AI's potential to revolutionize traditional practices.
Keywords:
Subject: Computer Science and Mathematics - Artificial Intelligence and Machine Learning
1. Introduction
The advent of artificial intelligence (AI) has revolutionized many industries by introducing innovative solutions that enhance efficiency, accuracy, and productivity. In sectors ranging from healthcare to finance and agriculture, AI has demonstrated its capacity to analyze vast amounts of data, identify patterns, and make data-driven decisions [1]. These advancements have led to significant improvements in disease diagnosis [2], financial forecasting [3], and crop management [4], illustrating the transformative potential of AI technologies.
In agriculture, AI has emerged as a powerful tool in precision farming, optimizing various aspects of crop production. AI-driven technologies such as machine learning algorithms, computer vision, and robotics have been pivotal in increasing yields, reducing resource use, and minimizing the environmental impact of agricultural practices [5,6]. The integration of AI in agriculture enables farmers to monitor crop health, predict pest outbreaks, and manage irrigation and fertilization more effectively [7,8].
As the wine industry faces increasing pressure to improve quality, ensure sustainability, and increase operational efficiency, the role of AI's is becoming more prominent. Winemaking is a complex process that ranges from viticulture—the cultivation of grapevines—to the intricacies of fermentation and bottling. Each stage presents unique challenges that benefit from the precision and predictive capabilities of AI [9].
In viticulture, AI technologies monitor vineyard conditions, predict disease outbreaks, and optimize resource management. Drones, satellite imagery, and sensors provide real-time insights into grapevines health, enabling proactive interventions that enhance grape quality and yield [10]. During production and fermentation, AI systems monitor key variables such as density, temperature, sugar levels, and yeast activity to ensure optimal fermentation conditions and consistent wine quality [11]Furthermore, in bottling and quality control, AI-driven machine vision systems detect imperfections and ensure label accuracy, maintaining high standards throughout the production process [12].
Beyond operational efficiency, AI applications in winemaking play a critical role in enhancing food safety and traceability. In an industry where product quality and safety are paramount, AI technologies offer sophisticated tools for monitoring and managing microbiological risks. For example, AI can detect mycotoxins and other contaminants to guarantee that the wine produced meets stringent safety standards [13]. Additionally, AI systems can integrate data from different points in the supply chain to provide comprehensive vineyard-to-consumer traceability, which is essential for maintaining consumer confidence and meeting regulatory requirements [14].
In this review, we explore the diverse applications of AI in vineyard management, production processes, and enhancing safety and traceability in winemaking. By examining these topics, we aim to provide a thorough understanding of how AI is transforming traditional practices in the wine industry. Specifically, this review will explore how precision agriculture techniques powered by AI, such as disease prediction models, pest management using drone technology, and automated grape harvesting systems, are revolutionizing vineyard management. It will examine AI systems for monitoring fermentation variables and predictive control to ensure consistency and quality and highlight the implementation of machine vision systems for quality inspection and the use of predictive analytics to maintain standards. It will also explore how AI is improving traceability and managing microbiological risks, with a focus on the identification of mycotoxins and other contaminants. Finally, we will identify current limitations of AI technology within the enological context and future advances that could further transform winemaking.
2. Artificial Intelligence: An Overview (Fundamentals and Applications in Agriculture)
AI represents a transformative shift in modern technology, fundamentally changing how machines perform tasks that traditionally required human intelligence [15]. This section covers the core aspects of AI, tracing its historical development, elucidating its fundamental principles, and examining its significance in various domains, with a particular focus on its applications in winemaking.
2.1. Definition and Basics of AI
AI refers to the ability of computer systems to perform tasks that generally require human cognitive functions, including learning, reasoning, perception, prediction, and decision-making. Essentially, AI involves the creation of algorithms that enable machines to replicate human cognitive processes, such as dynamically processing and adapting to information over time [16].
The field of AI is comprised of a multitude of sub-domains (Figure 1):
Machine Learning (ML): ML algorithms are designed to learn from and make predictions based on data. Techniques such as supervised learning, unsupervised learning, and reinforcement learning fall under this category. In supervised learning, models are trained on labeled data, enabling them to make predictions or classifications. Unsupervised learning involves finding hidden patterns or intrinsic structures in input data without labeled responses. Reinforcement learning trains algorithms through trial and error, using feedback from their actions to learn optimal behaviors [17]. Deep learning, a subset of ML, uses multi-layered neural networks to model complex patterns in data. Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) are key architectures within deep learning, used for image and sequence data respectively [18].
Computer Vision: This field enables machines to interpret and make decisions based on visual data. CNNs have been pivotal in advancing image recognition capabilities. Techniques in computer vision include object detection, image segmentation, and facial recognition, which are widely used in applications ranging from autonomous vehicles to security systems [19]. Advances in hardware, such as Graphics Processing Units (GPUs), have significantly enhanced the processing power available for computer vision tasks, allowing for real-time analysis and decision-making [20].
Natural Language Processing (NLP): NLP allows machines to process and generate human language. Techniques in NLP include syntactic parsing, sentiment analysis, and machine translation. The development of transformers and attention mechanisms has significantly improved the performance of NLP models, enabling them to handle tasks such as text generation, question answering, and language translation with high accuracy [21]. NLP applications are found in virtual assistants, chatbots, and automated customer service, transforming how humans interact with machines [22].
Expert Systems: These represent human knowledge in structured domains. They use a set of rules to analyze information and make decisions or solve problems in specific areas. Expert systems are particularly useful in fields requiring specialized knowledge, such as medical diagnosis and financial forecasting. Recent advancements have demonstrated their efficacy in these applications, particularly in enhancing decision-making processes and improving diagnostic accuracy [23].
Robotics: Integrating AI with robotics has led to the development of autonomous systems capable of performing complex tasks in dynamic environments. Robotics combines mechanical engineering, electrical engineering, and computer science to design and build robots. AI enhances robotics through the implementation of advanced algorithms for path planning, object manipulation, and human-robot interaction. Examples of AI-driven robotics applications that improve efficiency and safety include autonomous drones, robotic arms in manufacturing, and service robots in healthcare [24].
2.2. Evolution of AI
The evolution of AI began with early conceptualizations of pioneering figures such as Ada Lovelace and Alan Turing, who established the groundwork for the notion of machines capable of intelligent behavior. During the 1950s and 1960s, AI was formally recognized as a discipline, accompanied by early attempts at creating intelligent machines. This led to the development of the first AI programs and the establishment of AI research labs [16]. Despite these initial successes, the field encountered significant challenges, including limited computational power and insufficient data, which led to periods of stagnation known as "AI winters" during the 1970s and 1980s [15,25]
The resurgence of AI in the late 1990s and early 2000s was driven by advances in computational power, the availability of large datasets, and improvements in algorithms. Breakthroughs in machine learning, particularly the development of support vector machines and neural networks, led to a resurgence of interest in AI [26]. The advent of deep learning in the 2010s, with neural networks capable of learning from vast amounts of data, marked a significant milestone in the field. This era witnessed remarkable achievements in image recognition, speech processing, and natural language understanding [17]. Recent advancements in AI have been characterized by the development of sophisticated models, such as transformers, which have revolutionized NLP and extended their applications to other domains like computer vision [27]. The integration of AI with other technologies, such as the Internet of Things (IoT) and blockchain, is paving the way for new innovations and applications across various industries [28].
2.3. Different AI Application in AgroSciences
Plant Disease Detection
The advent of AI has revolutionized plant disease detection by enhancing diagnostic accuracy and efficiency. Advanced models, such as those integrating Transfer Learning with Vision Transformers (TLMViT), leverage the architecture of transformers to focus on relevant image parts, achieving high accuracy in disease identification [29]. Studies using large datasets, like PlantVillage with over 54,000 images of diverse crop species, have demonstrated that AI models can achieve identification accuracies above 98%, highlighting the potential of hybrid models in agricultural applications [30]. These advancements are pivotal for precision agriculture and sustainable crop management, providing farmers with powerful tools to detect diseases early and manage crops more effectively.
AI in Precision Agriculture
Precision agriculture is a farming management concept based on the observation, measurement , and response to inter- and intra-field variability in crops. AI plays a crucial role in this field by processing data from various sources, such as satellite imagery, drones, and sensors, to provide actionable insights [31]. Machine learning algorithms analyze this data to predict pest outbreaks, optimize irrigation schedules, and improve yield forecasts [30]. AI-driven systems can monitor crop health in real-time, allowing for timely interventions that enhance crop quality and yield. For example, deep learning models can analyze drone-captured images to identify pest or disease-affected areas , enabling targeted treatments that reduce chemical use and environmental impact [32].
AI in Food Safety and Traceability
AI plays a crucial role in enhancing food safety and traceability within the agro-food industry. Using advanced algorithms and machine learning techniques, AI enables precise, real-time monitoring of agricultural products throughout the supply chain. For instance, AI systems can identify and prevent contaminations in food products, thereby improving consumer safety [33].
AI-based technologies in agriculture, such as sensors and drones, allow for detailed and real-time traceability of crops from planting to harvesting and distribution, ensuring product integrity throughout the supply chain [34].
Moreover, the implementation of AI systems for big data analysis in the food supply chain enables early anomaly detection and efficient food safety management. These systems can analyse vast amounts of data to foresee potential risks and optimize management practices, ensuring that the food reaching our tables is safe and of high quality [35].
3. Applications of AI in Viticulture
The application of AI has significantly transformed viticulture by implementation precision agriculture techniques that enhance vineyard management and optimize grape production. Various AI-powered applications in viticulture, including disease prediction models, pest management using drone technology, automated grape picking systems, and data-driven approaches for optimizing water and nutrient management. These technologies are collectively driving the transition towards more efficient and sustainable practices in the viticultural sector.
3.1. Disease Prediction Models
Disease prediction models are among the most impactful applications of AI in viticulture. These models utilize machine learning algorithms to analyze historical data, weather conditions, and environmental factors to predict the likelihood of disease outbreaks. Effective disease management is critical for maintaining vineyard health and productivity, and AI provides advanced tools to enhance this process.
Data collection is the foundational step in developing effective disease prediction models. Various sources, including weather stations, soil sensors, and satellite imagery, provide crucial information about environmental conditions that affect vine health. Sensors placed throughout the vineyard monitor factors such as humidity, temperature, soil moisture, and leaf wetness, which are essential for the development of diseases like powdery mildew and downy mildew[36].
Machine learning models such as Support Vector Machines (SVM), Neural Networks, and Random Forests are employed to identify patterns and correlations that are not immediately apparent [15]. These models are trained on historical data to recognize conditions that typically precede disease occurrence, enabling early warnings and proactive management [37]. Recent studies have shown that using data from inside the grape canopy, as opposed to external weather stations, significantly improves the precision of forecasting models by capturing microclimatic variations critical for disease prediction [38]. Moreover, the integration of IoT technology with AI has further enhanced the data collection process. IoT sensors deployed in vineyards continuously collect and transmit real-time data on environmental parameters to cloud-based platforms for analysis. This approach not only improves the accuracy of disease prediction but also enables timely and precise interventions [39].
3.1.1. Predictive Modeling and Implementation
Predictive modeling involves creating algorithms that can forecast disease outbreaks based on real-time data inputs. For instance, a model might predict the likelihood of a powdery mildew outbreak by analyzing recent weather patterns, current humidity levels, and historical outbreak data. This predictive capability allows vineyard managers to implement targeted interventions before a disease becomes widespread [40].
Implementing these models involves integrating them with decision support systems that provide actionable insights to vineyard managers. For example, a decision support system might send alerts when conditions are favorable for disease development, recommending specific actions such as adjusting irrigation schedules or applying fungicides. This proactive approach reduces the reliance on routine chemical applications, promoting more sustainable farming practices [41]. Recent advancements in integrating IoT devices and machine learning models have further enhanced these systems, allowing for real-time monitoring and decision-making [39].
The efficacy of AI-powered disease prediction models in viticulture has been demonstrated by several case studies. For instance, research conducted in Tuscany (Italy), demonstrated how the Etat Potentiel d’Infection (EPI) current model accurately predicted the risk of downy and powdery mildew infections, enabling vineyard managers to reduce fungicide applications by up to 50-77% for powdery mildew and 42-50% for downy mildew, without compromising crop protection [40]. Another study conducted by Chen et al. developed and tested predictive models for grape downy mildew (GDM) in Bordeaux vineyards. The research utilized a dataset spanning nine years and incorporated both statistical and machine learning algorithms, including generalized linear models, LASSO, random forests, and gradient boosting, to predict the probability of high incidence and severity of GDM. These models used real-time data on disease onset dates and average monthly temperatures and precipitation. Notably, the integration of these models into decision support systems could significantly reduce fungicide applications by over 50%, contributing to more sustainable viticultural practices [39].
A recent study by Miguel Madeira et al. presents an innovative approach using satellite imagery, real-time weather data, and deep learning (DL) and a custom convolutional neural network (CNN) architecture tailored for plant disease classification using vineyard imagery. This model was tested against other state-of-the-art models, such as ResNet50 and MobileNetV2, and showed comparable performance, with MobileNetV2 achieving a remarkable accuracy of 99.375%. In addition, the front-end uses Mapbox for satellite data visualization, providing an intuitive user interface for vineyard managers. This system not only facilitates accurate disease classification and treatment, but also incorporates geo-tagging to map disease hotspots, improving targeted intervention strategies [42].
Moreover, integrating these models with mobile applications and user-friendly interfaces allows vineyard managers to access real-time data and recommendations, facilitating rapid decision-making. These tools empower even small-scale vineyard operations to leverage advanced AI technologies for disease management, enhancing overall productivity and sustainability [41]. Recent advancements have also shown that combining AI with IoT sensors and machine learning models improves the precision of forecasting models by capturing microclimatic variations critical for disease prediction [39].
3.2. Pest Management Using Drone Technology
Unmanned Aerial Vehicles (UAVs) equipped with AI-powered imaging systems has brought about a revolutionary change in the field of pest management in vineyards. These drones can capture high-resolution images of the vineyard, which are then analyzed using computer vision algorithms to detect signs of pest infestations. AI algorithms identify patterns and anomalies in the foliage that indicate the presence of pests like grapevine moths or phylloxera [43].
The use of drones offers several advantages over traditional pest management methods. Firstly, drones can cover large areas quickly, providing comprehensive surveillance of the vineyard in a fraction of the time required for manual inspections. This rapid coverage allows for the timely detection of pest issues, crucial for preventing widespread infestations. Additionally, drones can access areas that are difficult to reach on foot, such as steep slopes or densely planted sections of the vineyard, ensuring complete monitoring of the vineyard [44].
AI-powered imaging systems on drones analyze the captured high-resolution images to detect visual signs of pest damage, such as discoloration, holes in leaves, or unusual growth patterns. By training on extensive datasets of labeled images, AI models can learn to recognize subtle differences between healthy and infested plants, improving the accuracy of pest detection [45]. For example, AI models can identify grapevine moth larvae by detecting specific patterns of leaf damage, allowing precise identification and localization of pest infestations. This level of detail enables targeted interventions that minimize pesticide use and reduce the environmental impact of pest control measures [46].
Once pests are detected, drones can also be used to implement pest control measures directly to the affected areas. For instance, they can release biological control agents, such as beneficial insects, to target specific pest populations. This method, known as integrated pest management (IPM), reduces the reliance on chemical pesticides and promotes a more sustainable approach to vineyard management [47]. Additionally, drones equipped with precision sprayers can apply pesticides only where needed, reducing the overall amount of chemicals used and preventing unnecessary exposure to non-target plants and animals. This targeted application protects the environment and reduce costs for vineyard managers by optimizing the utilization of pest control resources [44].
Several vineyards have successfully implemented drone technology for pest management. A study by Albetis et al. explored the potential of Unmanned Aerial Vehicle (UAV) multispectral imagery combined with AI to detect Flavescence dorée (FD) and Grapevine Trunk Diseases (GTD) in vineyards. The research was conducted in southern France, covering seven vineyards with five different red grape cultivars. UAVs equipped with MicaSense RedEdge® sensors captured multispectral images, which were processed to create high-resolution reflectance mosaics. The study tested 24 variables, including spectral bands, vegetation indices, and biophysical parameters, to distinguish symptomatic vines from healthy ones. This integration of UAVs, AI, and real-time data demonstrated a promising approach for accurate and efficient disease detection, facilitating targeted interventions and reducing the need for extensive manual monitoring [48,49].
For instance, a vineyard in Spain reported a 40% reduction in pesticide use subsequent to the integration of drones into their pest management strategy, while maintaining or even improving crop health and yield [50]. Gennaro et al. and Junges et al. utilized UAV-mounted sensors to identify esca infections in vine leaves and crops, employing vegetation indices to distinguish between healthy and infected plants, which enabled accurate early detection [51,52]. Similarly, MacDonald et al. applied hyperspectral imaging to detect leafroll virus in Cabernet Sauvignon vineyards. They used specific spectral bands and classification models to achieve precise infection identification, facilitating efficient vineyard management through early detection [53]. Furthermore, a study by Vélez et al., demonstrated the use of UAV multispectral imagery combined with AI to map the spatial variability of Botrytis bunch rot (BBR) risk in vineyards. Conducted in Galicia, Spain, UAVs equipped with multispectral cameras captured high-resolution images to create Digital Terrain Models (DTM), Normalized Difference Vegetation Index (NDVI), Canopy Height Models (CHM), and Leaf Area Index (LAI) maps. These variables, combined with a Random Forest algorithm, generated accurate BBR risk heatmaps (R² > 0.7), enhancing disease management and reducing unnecessary treatments [54].
Looking forward, advancements in drone technology and AI are expected to further enhance the capabilities of pest management systems. Improvements in drone autonomy, battery life, and imaging resolution, coupled with more sophisticated AI models, will enable even more precise and efficient pest detection and control. The integration of real-time data from IoT sensors and the use of predictive analytics will allow for proactive pest management strategies, further reducing the reliance on reactive chemical treatments and supporting sustainable viticulture practices [55].
3.3. Automated Grape Picking Systems
AI-driven automated grape picking systems represent a significant advancement in viticulture. These systems integrate several advanced technologies: robotics for picking and handling grapes, computer vision for assessing grape clusters, and machine learning for processing visual data to make harvesting decisions. This combination allows grapes to be selected at their optimal ripeness [56].
Similar to UAVs, robotic harvesters are equipped with cameras and sensors that capture detailed images of the grape clusters. These high-resolution images are analyzed in real time by AI algorithms, which assess parameters such as color, size, shape, and texture to distinguish between ripe and unripe grapes [56]. This selective harvesting improves the overall quality of the wine by reducing the number of unripe grapes picked [57]. Advanced AI models, including those based on deep learning, have shown high accuracy in identifying ripe grapes even under challenging conditions, such as varying light levels and leaf occlusion [45].
Additionally, GPS technology and mapping systems ensure efficient navigation of the vineyard, covering all areas without redundancy [58]. Recent advancements in sensor technologies, such as the GA-YOLO model, have significantly enhanced the ability of robotic systems to detect and harvest grapes in dense and occluded environments [56]. The integration of advanced kinematic models and real-time data processing ensures high accuracy and efficiency, even in complex vineyard layouts [59].
The impact of automated grape picking systems on vineyard operations is profound. By increasing the speed and efficiency of harvesting, these systems allow vineyard managers to maximize output during the harvest season. The precision of AI-driven harvesters ensures that grapes are picked at their peak, enhancing wine quality and reducing waste from overripe or damaged grapes [57]. Automated systems also contribute to more sustainable practices by reducing the need for chemical treatments and minimizing the environmental impact of harvesting operations. Precise targeting of ripe grapes reduces waste and ensures more efficient use of resources [60].
These automated systems offer several advantages over traditional manual harvesting. They can operate continuously, day and night, increasing efficiency and reducing labor costs. Automation ensures that grapes are picked at their precise moment of optimal ripeness, which is crucial for producing high-quality wine [61]. Studies have shown that these systems can reduce labor costs by up to 40% while maintaining or improving grape quality [62]. Additionally, they enhance safety by reducing the need for human workers to perform repetitive, physically demanding tasks in potentially hazardous conditions, such as steep slopes or extreme weather [58].
Several vineyards have successfully implemented these systems. A vineyard in France reported a 25% increase in harvesting efficiency and significant improvements in grape quality after adopting robotic harvesters [62]. Many studies have focused on the application and improvement of the GA-YOLO model. Chen et al. introduced a lightweight YOLO model to detect dense and occluded grape targets using drones in Bagui Garden, Guangxi, China. This model achieved an average accuracy (mAP) of 96.87% and a detection speed of 55,867 FPS, significantly reducing model parameters by 82.79% [63]. In addition, Zhao et al. developed a lightweight-improved YOLOv5s model for grape fruit and stem recognition to enhance mechanized harvesting, achieving a precision of 96.8%, recall of 97.7%, and mAP of 98.6% [49,63]
In another work, Coll-Ribes et al, developed a robust accurate method for the detection and localization of the peduncle of table grapes based on a combination of instance segmentation and monocular depth estimation using Convolutional Neural Networks (CNN), with direct implementation in automatic grape harvesting with robots [64].
The results of these studies indicate that he future of automated grape picking systems looks promising. Ongoing research and development are focused on further enhancing their capabilities. Innovations such as advanced AI algorithms, improved robotic dexterity, and enhanced sensor technology are expected to make these systems even more efficient and reliable. Future automated pickers may also incorporate AI-driven predictive analytics to forecast optimal harvest times based on real-time environmental data and historical trends [59]. Integrating these systems with other vineyard technologies, such as IoT sensors and blockchain for traceability, will create a more interconnected and intelligent vineyard management ecosystem. This holistic approach will enable vineyard managers to make data-driven decisions across all aspects of viticulture, from planting to harvesting to marketing the final product [60].
3.4. Data-Driven Approaches for Optimizing Water and Nutrient Management
Efficient water and nutrient management are crucial for optimizing grape yield and quality. AI-driven data analysis and precision agriculture techniques provide vineyard managers with the tools needed to manage these resources effectively.
3.4.1. Water Management
The management of water resources is a fundamental aspect of viticulture, exerting a significant influence on the yield and quality of grapes. AI technologies enable precise irrigation management by analyzing data from soil moisture sensors, weather forecasts, and plant stress indicators. ML models can predict the optimal irrigation schedules, ensuring vines receive the right amount of water at the right time, thus preventing both under- and over-irrigation [65]. By integrating data from remote sensing, ground-based sensors, and weather stations, AI systems create a comprehensive picture of the vineyard's water needs, allowing for more efficient water application and conservation of this valuable resource [66].
For example, a study in New Zealand demonstrated the effectiveness of a multi-agent system for intelligent irrigation management. This system maximized water sharing within a community by accurately estimating water needs based on crop type, farm size, and other factors, leading to more efficient water use [67]. Additionally, a vineyard in Italy implemented an AI-driven irrigation management system that reduced water usage by 20% while maintaining or improving grape quality [65]. In Portugal, the SIMDualKc model was used to evaluate irrigation water management under scarcity in Mediterranean vineyards. This study demonstrated improved water use efficiency and better soil water balance, contributing to sustainable viticulture practices [68]. These examples highlight the ability of AI-driven water management systems to conserve water resources while maintaining high standards of grape quality [69].
3.4.2. Nutrient Management
Nutrient management is equally important for the health and productivity of grapevines. AI-driven approaches analyze soil data, plant tissue analyses, and environmental conditions to optimize the application of fertilizers. By understanding the specific nutrient requirements of different vineyard zones, AI models can recommend precise fertilizer application rates and timings, ensuring that vines receive the nutrients they need without waste [70]. Machine learning algorithms can also predict nutrient deficiencies before they become visible, allowing for timely interventions that prevent yield losses. For example, AI models can analyze leaf color and texture changes captured by drone imagery to detect early signs of nutrient stress, guiding corrective actions [71].
In a study carried out in China, demonstrated the efficacy of combining unmanned aerial vehicle (UAV) multispectral remote sensing and machine learning was demonstrated to be an effective method for predicting the nitrogen, phosphorus, and potassium contents in grape leaves [71].
Hongyi Lyu et al., investigated the use of hyperspectral reflectance combined with machine learning models to assess the nutrient status of Pinot Noir grapevine leaves in New Zealand. This research utilized a handheld spectroradiometer to collect spectral data from grapevine leaves. The study applied partial least squares regression (PLSR), random forest regression (RFR), and support vector regression (SVR) to predict concentrations of key nutrients (N, P, K, Ca, and Mg) [72]. This approach enhances precision in nutrient management, optimizing grape yield and maintaining vine health. Other studies have developed a super learner ensemble model to map potassium (K) fixation and availability in vineyard soils within the Lodi American Viticulture Area (AVA) in California. This model integrates machine learning algorithms, including random forest (RF), extreme gradient boosting (XGB), and cubist, into a super learner ensemble. The research underscores the importance of combining machine learning with soil mapping to enhance nutrient management in vineyards, facilitating more precise K fertilization strategies and improving vine health [73].
In this context, the integration of water and nutrient management strategies through the use of AI-driven systems allows for the implementation of a comprehensive approach to vineyard management. By examining the interrelationships between water accessibility and nutrient absorption, artificial intelligence (AI) models can simultaneously optimize the utilization of these resources, thereby enhancing grape yield and quality while minimizing environmental impact [74].
To illustrate, an integrated system that combines water and nutrient management can utilize real-time data from IoT sensors and AI algorithms to provide precise recommendations for both irrigation and fertilization schedules. This integrated approach guarantees that vines receive the optimal combination of water and nutrients, which results in enhanced vine health and productivity [70].
4. AI in the Production, Fermentation Process and Bottling Quality Control
The application of AI in the production and fermentation process of winemaking is transformative, enhancing the precision and control of various critical parameters. AI systems enable real-time monitoring and predictive control of fermentation variables, ensuring consistency and quality in the final product.
4.1. Monitoring Fermentation Variables
AI technologies facilitate the continuous monitoring of key fermentation variables such as temperature, sugar levels, and yeast activity. These variables are crucial for the successful fermentation of wine, affecting both the process efficiency and the quality of the final product.
Temperature Control
Temperature is one of the most critical factors in the fermentation process. Maintaining an ideal temperature range is essential for yeast metabolism, preventing the formation of unwanted by-products, and ensuring consistent flavor profiles in the wine [75]. AI systems equipped with sensors and machine learning algorithms can monitor and adjust the temperature in real-time to maintain optimal conditions.
Advanced AI systems integrate data from multiple sources, including ambient temperature, fermentation vessel temperatures, and even heat generated during the fermentation process. The comprehensive data analysis by AI, can predict temperature fluctuations and make proactive adjustments to prevent deviations that could affect the fermentation process. This, enables precise control over the fermentation environment, thereby reducing the risk of thermal shock or prolonged exposure to suboptimal temperatures [76].
A study by Rachmadi et al. demonstrated the effectiveness of AI system that predicted temperature fluctuations and automatically adjust cooling mechanisms to maintain the optimal fermentation temperature, resulting in improved wine quality and consistency [75]
Malavasi et al. also highlighted how integrated AI systems that incorporate data from multiple sources provide precise control of the fermentation environment, reducing the risk of thermal shocks and optimizing outcomes [76].
Moreover, AI systems can also learn from historical fermentation data, enabling them to identify patterns and trends that may affect temperature stability. This learning capability allows the AI to continuously improve its predictive accuracy and control strategies, adapting to the specific needs of different grape varieties and fermentation conditions.
Sugar Level Monitoring
Monitoring sugar levels during fermentation is essential for determining the progress and completion of the process. AI systems utilize spectroscopy, refractometry, and other sensor technologies to continuously measure sugar concentrations. Machine learning algorithms process this data to provide accurate predictions of sugar level changes, enabling timely interventions to maintain desired fermentation trajectories. Accurate sugar level monitoring ensures that the fermentation process is proceeding as expected and helps winemakers decide when to make necessary adjustments, such as adding nutrients or adjusting temperature. This continuous monitoring also helps in detecting and correcting issues such as stuck or sluggish fermentations, which can significantly impact the quality of the wine [77].
A study by Samphao et al. demonstrated the application of a bienzymatic biosensor for monitoring glucose and ethanol content during wine fermentation. The biosensor, based on glucose oxidase (GOx) and alcohol dehydrogenase (ADH), was effective in real-time monitoring and provided results consistent with conventional methods, highlighting its potential utility in winemaking [78].
Yang et al. developed an online glucose analysis system with high precision and a wide detection range for fermentation monitoring. This system, based on a homemade screen-printed enzymatic biosensor chip, showed excellent detection performance during the fermentation of sophorolipid and ethanol, making it a valuable tool for precise process control [79].
Furthermore, Fuller et al. successfully used Raman spectroscopy combined with machine learning algorithms to predict total sugar concentrations, ethanol levels, and pH during wine fermentation. This approach proved to be a viable tool for rapid and accurate fermentation monitoring [80].
Yeast Activity
Yeast activity is another vital parameter that influences the rate and quality of fermentation. In this context, AI-driven models can analyze patterns in yeast metabolism, identifying any anomalies that may indicate issues such as stuck fermentation or contamination. This predictive capability allows winemakers to address potential problems before they escalate, ensuring a smooth and efficient fermentation process [81].
In their study, Torello Pianale, Rugbjerg, and Olsson demonstrated the use of a toolbox of biosensors for real-time monitoring of the yeast intracellular state during bioprocesses. This approach involves the integration of advanced biosensors with AI systems to monitor yeast viability, population dynamics, and metabolic by-products in real-time. The biosensors provide detailed data on the physiological state of the yeast cells, which AI algorithms can analyze to detect and predict deviations from optimal fermentation conditions [81]
The real-time monitoring facilitated by these biosensors is crucial for managing the fermentation process, as yeast health directly impacts fermentation kinetics and the development of flavor compounds. By maintaining optimal yeast activity, winemakers can ensure that the fermentation process produces wine with the desired characteristics. The AI systems, leveraging the data from the biosensors, enable precise control and timely interventions, thus enhancing the overall efficiency and reliability of the fermentation process.
Integration of Multiple Variables
The integration of multiple fermentation variables into a single AI system allows for a more holistic approach to fermentation management. AI platforms can simultaneously monitor and analyze temperature, sugar levels, and yeast activity, providing comprehensive insights into the fermentation process. This integrated approach enables more precise control and optimization, leading to higher quality and more consistent wine production.
For instance, a study by Flores-Hernández et al. discusses the implementation of AI tools for modeling, predicting, and managing the fermentation process in white wine production. The authors used a multi-layer perceptron neural network integrated with genetic algorithms to predict the concentration of alcohol and the amount of substrate at different stages of fermentation. This approach enabled the researchers to automate the process, reducing the need for manual measurements and increasing the accuracy of predictions. The integration of variables such as temperature, time, initial substrate concentration, and biomass concentration into a single model significantly improved the predictive accuracy and control of the fermentation process [82].
Another example is provided by Urtubia et al., who used Support Vector Machines (SVM) to detect abnormal fermentations in wine production. By analyzing a large dataset of fermentation control variables and chemical markers, the SVM model was able to predict deviations early in the process. The integration of multiple variables, including density, YAN, brix, and acidity, allowed the system to identify potential issues and suggest corrective actions, thereby ensuring a more stable and controlled fermentation process [83].
Real-Time Data Analytics
The ability to analyze fermentation data in real-time is a significant advantage of AI systems. Real-time data analytics allow for immediate detection of deviations from desired fermentation conditions, enabling prompt corrective actions. This proactive approach minimizes the risk of fermentation issues and enhances the overall efficiency and reliability of the winemaking process.
Schwinn et al. explored the use of real-time data analytics for monitoring and controlling the fermentation process in bioreactors. By employing advanced sensor technology to measure parameters such as dissolved oxygen, carbon dioxide production, and volatile compound concentrations, the researchers developed a robust AI system capable of real-time monitoring and control. The system's ability to process and analyze data instantaneously allowed for immediate adjustments to fermentation parameters, ensuring optimal conditions throughout the process [84].
A study by Xu et al. demonstrated the integration of virtual data augmentation and deep neural networks (DNNs) to predict volatile fatty acid (VFA) production in anaerobic fermentation processes. By utilizing real-time data from multiple sensors, the AI system could accurately predict VFA levels, allowing for dynamic adjustments to the fermentation environment. This approach highlighted the potential of real-time data analytics in enhancing the predictive accuracy and control capabilities of AI systems in fermentation processes [85].
These examples illustrate how the integration of advanced sensors and real-time data analytics can significantly improve the management and optimization of the fermentation process, leading to higher quality and more consistent wine production.
4.2. Predictive Control of Fermentation
4.2.1. Predictive Modeling
In addition to monitoring, AI systems are utilized for the purposes of predicting and controlling the fermentation process. Predictive models utilize historical data and real-time inputs to forecast the behavior of the fermentation process, thereby enabling preemptive adjustments that ensure consistency and quality. These models, which employ machine learning algorithms, can predict the optimal time to intervene in the fermentation process, such as adjusting temperatures or adding nutrients, in order to achieve the desired wine characteristics [86].
For instance, neural networks can analyze fermentation data from past batches to identify patterns indicating successful fermentations. By recognizing these patterns in real-time data, the AI system can forecast potential issues and recommend adjustments before problems arise. This foresight helps maintain the consistency and quality of the wine. Florea et al. highlight the use of advanced predictive models in winemaking, incorporating techniques like regression analysis and decision trees to predict fermentation outcomes accurately. Integrating multiple data sources, such as temperature, pH levels, and sugar content, allows for a comprehensive understanding of fermentation dynamics. The study also discusses Transfer Learning, enabling models to adapt to new datasets with minimal retraining, enhancing generalization across different fermentation batches. [86]. This study presents the development of a web-based software application to model, predict, and manage the white wine fermentation process using a neural network pre-trained with genetic algorithms. This research aimed to digitalize the fermentation process to optimize white wine production by using a multi-layer perceptron neural network to predict alcohol and substrate concentrations during fermentation, based on variables such as temperature, initial substrate concentration, and biomass concentration. By optimizing the neural network configuration with genetic algorithms, the prediction accuracy was significantly enhanced. This AI-driven approach minimized the need for manual measurements and extended the lifecycle of sensors, demonstrating the potential of integrating AI and real-time data to improve fermentation efficiency and product quality [86].
Integrating IoT devices with predictive modeling further refines the process, with real-time data from sensors informing dynamic adjustments to maintain optimal fermentation conditions. Meanwhile, the incorporating feedback loops, where outcomes of previous adjustments refine future predictions, ensures continuous improvement in model precision and effectiveness. By leveraging predictive modeling, winemakers achieve greater control, reducing variability and improving wine quality [86].
4.2.2. Dynamic Control Systems
AI-driven dynamic control systems go beyond static monitoring by actively managing the fermentation process. These systems use predictive models to adjust fermentation parameters continuously, based on real-time data inputs. For example, if the AI predicts an undesirable rise in temperature, it can automatically trigger cooling mechanisms to maintain optimal conditions [87].
Juuso developed intelligent dynamic simulation models for fed-batch fermentation processes, which predict critical variables such as dissolved oxygen concentration and carbon dioxide levels. These models facilitate real-time adjustments to maintain optimal fermentation conditions, enhancing the overall efficiency and consistency of the process. This comprehensive control ensures a smooth and efficient fermentation process, leading to consistent high-quality wine production [87].
Dynamic control systems can also manage nutrient additions and oxygenation schedules. By continuously analyzing the fermentation kinetics, AI systems can determine the precise timing and quantity of nutrient additions needed to support yeast activity, ensuring a robust fermentation process. This level of control minimizes the risk of stuck fermentations and other common issues [88].
Pantazi et al. demonstrated that machine learning algorithms, when integrated into dynamic control systems, could optimize nutrient management by predicting the precise needs of the fermentation process. These systems analyzed real-time data to ensure timely nutrient additions, thus preventing nutrient deficiencies or excesses that could disrupt yeast metabolism. This proactive management of nutrient levels contributes significantly to the efficiency and success of the fermentation process [88].
4.2.3. Quality Assurance
AI systems enhance quality assurance by providing consistent monitoring and control throughout the fermentation process. By predicting potential deviations and automatically adjusting parameters, AI ensures that each batch of wine meets the established quality standards. This reduces variability and enhances the overall reliability of the winemaking process. AI-driven quality assurance systems can also detect, and correct issues related to off-flavors or other sensory defects. For example, AI can identify conditions that could lead to the production of unwanted compounds such as hydrogen sulfide and take preemptive actions to mitigate these risks. This capability helps maintain the desired flavor profile and quality of the wine.
Several studies have demonstrated that the integration of advanced sensor technology with AI has the potential to enhance the ability to monitor and control fermentation processes in real-time that can affect wine quality. For example, the study by Cavaglia et al. utilized ATR-MIR spectroscopy combined with multivariate analysis to control the alcoholic fermentation process and detect wine fermentation problems early on. This approach allows winemakers to maintain the desired sensory attributes by monitoring the volatile compounds that contribute to the wine's aroma and flavor profile. When this data is fed into AI systems, the algorithms can predict potential issues and recommend adjustments to maintain optimal conditions, ensuring consistent quality [89].
Similarly, the research by Samphao et al. demonstrated the use of a bienzymatic biosensor to monitor glucose and ethanol levels during wine fermentation. The real-time data provided by such sensors are critical for AI systems to make immediate adjustments to the fermentation parameters. By continuously analyzing fermentation kinetics, AI systems can determine the precise timing and quantity of nutrient additions needed to support yeast activity, ensuring a robust fermentation process. This level of control minimizes the risk of stuck fermentations and other common issues, thereby enhancing the quality and consistency of the final product [78].
Furthermore, AI-driven quality assurance systems can use data from advanced sensors to develop more accurate predictive models. These models can forecast potential deviations from the desired fermentation trajectory, allowing for preemptive adjustments. For example, the study by Berbegal et al. utilized PTR-ToF-MS for online monitoring of volatile organic compounds during alcoholic fermentation, providing insights into the impact of different yeast strains on wine quality. AI systems can leverage this detailed chemical analysis to predict and control the fermentation process more precisely, ensuring that each batch meets high-quality standards [90].
4.2.4. Advanced Sensor Integration
The integration of advanced sensors with AI systems further enhances predictive control capabilities. Sensors that measure a wide range of parameters, including dissolved oxygen, carbon dioxide production, and volatile compound concentrations, provide comprehensive data for AI analysis. This data allows the AI to develop more accurate and nuanced predictive models, improving its ability to control the fermentation process effectively [91].
By utilizing advanced sensor technology, AI systems can detect subtle changes in the fermentation environment that may not be apparent through traditional monitoring methods. This granular level of detail supports more precise control and optimization of the fermentation process, leading to higher quality and more consistent wine production.
For instance, the integration of optical fiber sensors and neural network models has been shown to significantly enhance the monitoring of fermentation parameters. Although Bakosová et al. conducted their study in the context of biocombustible production, the principles are applicable to winemaking. These sensors can dynamically monitor critical fermentation variables such as temperature and pH in real-time. The data collected is then analyzed by neural network models to predict fermentation trends and optimize process parameters, ensuring stable and efficient fermentation [92]. Integrating such advanced sensor technology with AI can similarly improve the efficiency and consistency of wine fermentation.
Recent advancements in sensor technology include the development of optical fiber sensors that can dynamically monitor alcoholic and oxidative fermentations. Soares et al. highlighted the use of these sensors for providing real-time data on key parameters such as temperature, pH, and biomass concentration. This real-time monitoring capability is critical for maintaining optimal fermentation conditions and improving the overall efficiency of the fermentation process [93]. When integrated with AI, these sensors enable the development of predictive models that can optimize fermentation conditions, ensuring consistent high-quality wine production.
Furthermore, AI-based forecasting models using yeast morphological data have demonstrated high accuracy in predicting ethanol yields. Itto-Nakama et al. utilized data from yeast cell morphology to develop predictive models that could forecast ethanol production with a high degree of accuracy. Although their study focused on ethanol fermentation for biocommodities, the approach is highly relevant for winemaking. This predictive capability allows for better management and stable production of wine, showcasing the potential of integrating AI with advanced sensor data in fermentation processes [94].
The use of advanced process analytical technology (PAT) tools, such as redox potential and capacitance sensors, further enhances the predictive control capabilities of AI systems. Rivera et al. demonstrated that these sensors provide real-time insights into the fermentation process, allowing for more precise adjustments to maintain desired fermentation conditions. Although their study was conducted in the context of sugarcane fermentation, the technology is directly applicable to winemaking. By integrating these advanced sensors with AI-driven systems, winemakers can achieve a higher level of control and optimization, ensuring the production of high-quality wine with consistent characteristics [95].
These case studies underscore the practical advantages of AI in fermentation control, illustrating how predictive modeling and dynamic control systems can enhance both the efficiency and quality of winemaking. As more wineries adopt these technologies, the overall standard of wine production is likely to improve, benefiting producers and consumers alike.
4.3. Machine Vision Systems for Bottling Quality Inspection
The application of AI in the bottling and quality control stages of winemaking is pivotal in ensuring product consistency, safety, and efficiency. By implementing AI-driven technologies, winemakers can maintain high standards and reduce variability across batches, thereby enhancing the overall quality and marketability of their products.
Machine vision systems, powered by AI, are revolutionizing the bottling process by providing advanced quality inspection capabilities. These systems utilize high-resolution cameras and sophisticated image processing algorithms to detect imperfections in bottles, labels, and packaging materials in real-time. Recent advancements in machine vision technology have enabled the development of highly accurate and efficient inspection systems, which are crucial for maintaining the high standards expected in the winemaking industry.
4.3.1. Detecting Imperfections
AI-driven machine vision systems can identify a wide range of defects during the bottling process. These defects can include cracks or chips in bottles, incorrect or misaligned labels, and inconsistencies in fill levels. By using machine learning algorithms, these systems can learn from historical data to improve their accuracy over time. This capability ensures that only products meeting stringent quality criteria proceed to the market, thereby reducing the risk of customer dissatisfaction and recalls.
For example, a study by Silva et al. demonstrates the implementation of a machine vision system for industrial quality control inspections, highlighting the significant improvements in defect detection accuracy achieved through the integration of AI technologies [96].
Additionally, a more recent study by Mahor and Yadav discusses an in-process intelligent inspection system that uses machine vision and deep learning techniques to detect flaws in the dimensions of bottles on a production line. This system achieved a mean average precision of 99.5% for object detection, demonstrating the efficacy of AI in maintaining high-quality standards in manufacturing processes [97].
These systems can also detect micro-defects that might be missed by human inspectors, ensuring a higher level of quality control. For instance, Ren et al. highlight the role of machine vision in detecting minute defects and ensuring the quality of visual inspections, emphasizing the importance of deep learning in enhancing image analysis and defect detection [98]. Machine vision systems, powered by AI, are revolutionizing the bottling process by providing advanced quality inspection capabilities. These systems utilize high-resolution cameras and sophisticated image processing algorithms to detect imperfections in bottles, labels, and packaging materials in real-time. Recent advancements in machine vision technology have enabled the development of highly accurate and efficient inspection systems, which are crucial for maintaining the high standards expected in the winemaking industry.
4.3.2. Ensuring Label Accuracy
Accurate labeling is crucial for compliance with industry regulations and for providing consumers with essential information. AI-powered vision systems can verify the accuracy and placement of labels, ensuring that each bottle displays the correct information regarding its contents, origin, and other mandatory details. This automated verification process significantly reduces the likelihood of human error and enhances operational efficiency.
Advanced machine vision systems use high-resolution cameras and sophisticated algorithms to inspect labels in real-time. These systems can identify defects such as smudges, misprints, and improper adhesion. For instance, Gong et al. (2020) developed a machine vision system for automatic online inspection of transparent label defects on curved glass bottles. This system employs an area-array camera and a custom-made blue dome illumination device to capture high-quality images, effectively eliminating reflection and refraction issues. By integrating advanced image processing techniques, the system can accurately detect even small defects, ensuring that all labels are applied correctly and legibly, which is essential for maintaining brand integrity and consumer trust [99].
Additionally, AI-driven vision systems can be trained using large datasets to improve their accuracy over time. The work by Wu demonstrated the effectiveness of using convolutional neural networks (CNNs) for wine label image recognition. By applying data augmentation techniques, the system achieved high accuracy in identifying wine labels, showcasing the potential of AI in enhancing label verification processes [100]
Another study by Li and Ma focused on distributed search and fusion strategies for wine label image retrieval. This approach addresses the challenge of imbalanced sample images among different wine brands, significantly improving the training of retrieval systems based on deep learning. Such advancements ensure that AI systems remain robust and reliable, even when dealing with a vast array of label designs and quality standards [101].
4.3.3. Integration with Automated Systems
AI-powered machine vision systems can be seamlessly integrated with other automated processes in the bottling line. For instance, when a defect is detected, the AI system can automatically remove the faulty product from the production line, ensuring that only flawless bottles are packed and shipped. This integration enhances the overall productivity and reliability of the bottling process.
A study by Liu et al. demonstrated the integration of multiple industrial cameras with edge-computing boxes and a programmable logic controller (PLC) for inline crack inspection in industrial manufacturing. This collaborative system allowed for real-time detection and correction of defects, optimizing the performance of the production process. Although this study was conducted in a different industry, its principles can be effectively applied to the wine bottling process to enhance defect detection and correction [102].
Moreover, integrating machine vision systems with robotic arms and conveyors allows for real-time adjustments and corrections. For example, if a label is detected as misaligned, a robotic arm can reposition it correctly before the bottle moves to the next stage. This level of automation reduces downtime and increases the throughput of the bottling line. Mikhail and Abed developed an integrated inspection system using a PLC to control the conveyor belt and a robot to remove defective bottles from the line in an advanced manufacturing setting. This approach can be adapted for the wine industry to improve production efficiency and final product quality [103].
Another relevant example is the work by Vagaš et al., who proposed a vision system for the automated MPS 500 line from FESTO, used primarily in educational settings for automation and machine control. This system not only monitors and evaluates data in real-time but also supports the practical training of automation technologies. Applying this approach to the wine industry could enhance both the training of new personnel and the operational efficiency of automated bottling lines [104].
Lastly, Tung et al. developed an automatic loading and unloading production system using robotic arms, automated guided vehicles (AGVs), and computer numerical control (CNC) machines for manufacturing. The integration of these components significantly improved production stability and reduced the need for human intervention. Adapting this level of automation to the wine industry could optimize the bottling process, ensuring consistent quality and efficiency [105].
Another example involves the integration of machine vision systems in a European winery's bottling line, which led to a 25% increase in throughput and a 20% reduction in labor costs. The system's ability to automate defect detection and corrective actions improved the overall productivity and reliability of the bottling process [106].
5.1.5. Future Directions and Innovations
The future of machine vision systems in the bottling process is promising, with ongoing advancements aimed at further enhancing their capabilities. Innovations such as 3D vision systems, hyperspectral imaging, and advanced machine learning algorithms are expected to improve defect detection accuracy and expand the range of detectable defects.
For instance, 3D vision systems can provide more detailed inspections of bottle surfaces, identifying defects that may be invisible in 2D images. This technology enables more precise quality control, ensuring that only the highest quality products reach consumers [107].
Hyperspectral imaging can detect chemical inconsistencies in labels and packaging materials, ensuring that all products meet stringent quality standards. This innovation enhances the ability of machine vision systems to detect a broader range of defects and maintain high-quality standards in the bottling process [98].
Additionally, the integration of machine vision systems with AI-driven predictive maintenance tools can help identify potential equipment failures before they occur, reducing downtime and improving the overall efficiency of the bottling process. This proactive approach to maintenance ensures that bottling lines operate smoothly and consistently, further enhancing the reliability and quality of wine production [108].
5. Food Safety and Traceability in the Wine Industry
Ensuring food safety and traceability in the wine industry is critical for maintaining consumer trust, complying with regulatory standards, and preserving the integrity of wine products. AI plays a pivotal role in enhancing these aspects by integrating data from IoT devices throughout the supply chain and providing sophisticated tools for detecting and managing microbiological risks.
5.1. Enhancing Traceability from Vineyard to Consumer
AI significantly enhances traceability in the wine industry by leveraging data from IoT devices deployed throughout the supply chain. These devices include sensors, RFID tags, and smart labels, which collect and transmit real-time data on various aspects of wine production, from vineyard conditions to final distribution.
5.1.1. Integration of IoT Devices
The integration of IoT devices enables comprehensive monitoring of the entire wine production process. In the vineyard, sensors monitor environmental conditions such as soil moisture, temperature, and humidity, providing valuable data that AI algorithms use to optimize grape cultivation practices. These sensors can also monitor variables like light intensity, wind speed, and nutrient levels, ensuring that vines receive optimal conditions for growth [109]. Misra et al. highlighted how the integration of AI and sensor data from IoT devices can enhance efficiency and precision in agriculture and the food industry, providing a valuable model for its application in viticulture [109].
Advanced IoT systems can also include drone-based sensors and satellite imagery. Drones equipped with multispectral cameras can capture detailed images of the vineyard, which are then analyzed by AI to assess vine health, identify areas affected by disease or pests, and monitor canopy density [110]. Taneja et al. discussed the implications of artificial intelligence in the agrifood sector, emphasizing the use of drones and multispectral cameras to monitor vine health and proactively manage pest and disease issues [110].
During the harvesting phase, RFID tags attached to grape bins track the location and condition of the harvest, ensuring traceability from the field to the winery. These tags provide data on harvest time, transport conditions, and grape variety, which is crucial for maintaining the integrity and quality of the grapes [111]. Expósito et al. presented RFID-based solutions for traceability in wine production, underlining the importance of these systems for maintaining product quality and integrity from harvest to winery [111].
In the winery, IoT devices continue to play a crucial role. Sensors monitor fermentation conditions, storage environments, and bottling processes, generating continuous streams of data. This data is analyzed by AI systems to ensure that each stage of production adheres to predefined quality standards. For instance, temperature and humidity sensors in storage facilities provide data that AI uses to maintain optimal aging conditions, preventing spoilage and ensuring product consistency [112]. Verzeletti et al. discussed the integration of information technology for traceability and quality control, demonstrating how temperature and humidity sensors can be used by AI systems to maintain optimal conditions during wine aging [112].
5.1.2. Real-Time Data Analysis and Decision Making
The continuous flow of data from IoT devices allows for real-time analysis and decision making. AI systems process this data to provide actionable insights to vineyard managers and winemakers. For example, if sensors detect a deviation from optimal growing conditions in the vineyard, the AI system can recommend immediate corrective actions, such as adjusting irrigation or applying fertilizers.
Real-time data analysis in the winery ensures that fermentation and storage conditions are continuously optimized. If the AI system detects that the temperature in a fermentation tank is rising beyond the ideal range, it can automatically adjust cooling systems to maintain the desired conditions. This proactive management reduces the risk of spoilage and ensures consistent quality across batches.
Advanced AI algorithms can also predict potential issues before they occur. By analyzing trends and patterns in the data, AI systems can forecast problems such as equipment failures or contamination risks, allowing for preemptive measures to be taken. This predictive capability enhances the efficiency and reliability of the winemaking process.
The potential of real-time data analysis using machine learning-based models is highlighted by ChandraPrabha and Lakshmi, who demonstrated its application in industrial settings. Their study utilized linear regression and decision tree algorithms to predict production outcomes based on real-time data points such as temperature and load. The emphasis on continuous monitoring and innovative prediction techniques aligns well with the needs of the viticulture and winemaking industries, where maintaining optimal conditions is crucial for high-quality production [113].
Furthermore, the robustness of AI models in handling real-time data perturbations is crucial for their effective deployment. Benedick et al. systematically evaluated the performance of various machine learning and deep learning algorithms under conditions of data quality degradation. Their findings underscore the importance of ensuring data integrity in real-time applications, which is directly relevant to maintaining consistent and reliable winemaking processes [114].
Real-time decision-making systems are also applicable beyond traditional industries. For instance, Ramanujam et al. proposed a real-time decision-making system for advertising networks using the STORM framework. Although their focus was on online advertising, the principles of real-time data aggregation and analysis can be effectively transferred to the wine industry. Implementing similar systems can help winemakers quickly respond to changes in production conditions, thereby maintaining high-quality standards [115].
The application of AI and machine learning in decision-making processes is further explored by Choubey and Karmakar in the oil and gas industry. They reviewed the use of predictive and inferential analytics based on historical and real-time data, which significantly enhanced operational efficiency and decision-making accuracy. These insights are valuable for optimizing various stages of the winemaking process, from vineyard management to bottling [116]
Additionally, Yan and Yang a decision support system based on deep learning for real-time data analysis and decision-making in marketing. Their system's ability to process large quantities of data quickly and provide efficient decision support demonstrates the potential for similar applications in winemaking. By adopting such advanced AI models, winemakers can ensure timely and precise adjustments to production processes, enhancing overall product quality [117].
5.1.3. Blockchain Technology for Traceability
AI systems can be integrated with blockchain technology to further enhance traceability. Blockchain provides a decentralized and immutable ledger of all transactions and data points throughout the supply chain. By combining AI with blockchain, wineries can create a transparent and tamper-proof record of each bottle’s journey from vineyard to consumer.
This integration ensures that all data collected by IoT devices is securely stored and easily accessible for verification by stakeholders, including regulators and consumers. Consumers can scan a QR code on the bottle to access detailed information about the wine’s production history, including vineyard practices, fermentation details, and storage conditions. This level of transparency builds consumer trust and adds value to the product.
Moreover, blockchain technology enhances supply chain security by preventing fraud and ensuring product authenticity. Each transaction and data point recorded on the blockchain is cryptographically secured, making it nearly impossible to alter or falsify information. This provides an additional layer of assurance that the wine has been produced and handled according to the highest standards.
The implementation of blockchain technology in the wine supply chain has demonstrated significant benefits. A study by Adamashvili et al. explored the impact of blockchain technology in the wine supply chain, emphasizing its role in enhancing traceability and protecting against fraud and contamination. Their research utilized Agent Based Models (ABMs) and simulations to compare traditional supply chains with blockchain-based ones. The findings highlighted that blockchain not only ensures a robust traceability system but also safeguards the production process from various types of fraud and contamination. This study underscores the comparative advantages of blockchain-based supply chains, particularly in terms of security and transparency, making a strong case for its adoption in the wine industry [118]
Further insights are provided by Prencipe et al., who analyzed the adoption of blockchain technology in the winery industry through an in-depth case study of Cantina Volpone, the first winery in Italy to implement blockchain for traceability. This study examined the operational and financial challenges faced during the implementation process, and highlighted how blockchain can unlock significant value and create competitive advantages despite these hurdles. The authors emphasized that blockchain enables comprehensive monitoring and recording of every step in the wine production process, from grape cultivation to distribution. This detailed traceability ensures that each bottle of wine can be tracked back to its origin, enhancing transparency and trust among consumers and stakeholders [119].
Malisic et al. conducted a systematic literature review on the adoption of blockchain technology in the wine supply chain, providing a comprehensive overview of current research and practices. Their review identified key benefits of blockchain, including improved traceability, authenticity, and consumer trust. The study also highlighted significant challenges, such as high implementation costs, the need for standardized protocols, and the lack of expertise in blockchain technology. By addressing these challenges, the authors suggested that the wine industry could significantly enhance its traceability and transparency, ultimately leading to greater consumer confidence and reduced fraud [120].
In addition to these studies, Gayialis et al. proposed a framework for developing Ethereum-based blockchain applications specifically for wine traceability. Their research presented a practical use case demonstrating how such applications can enable comprehensive supervision of production and distribution processes. The study showcased how blockchain technology could be leveraged to ensure timely and accurate traceability, which is crucial for combating counterfeiting and ensuring the authenticity of wine products. The authors also discussed the technical aspects of implementing blockchain in the wine industry, providing valuable insights for practitioners looking to adopt this technology [121].
Lastly, Kang et al. used a Stackelberg game-theoretical analysis to explore the economic implications of blockchain adoption in wine supply chains. Their study compared scenarios with and without blockchain implementation, uncovering the conditions under which blockchain can enhance traceability and authenticity. The authors found that blockchain adoption could lead to higher supply chain prices, driven by increased consumer willingness to pay for verifiable and authentic products. They also identified the threshold for third-party blockchain service fees, offering insights into the economic feasibility of blockchain technology for wine traceability. This analysis provides a strategic perspective on the adoption of blockchain, highlighting the potential economic benefits for the wine industry [122].
5.1.4. Enhancing Supply Chain Efficiency
Integrating IoT data with AI significantly enhances the efficiency of winery supply chains. AI algorithms can predict and manage disruptions by analyzing historical data and real-time inputs. For example, AI can forecast potential delays in grape deliveries due to weather conditions and recommend alternative routes or schedules to ensure timely processing. This proactive approach ensures that production schedules remain intact, reducing the risk of bottlenecks.
Moreover, AI-driven logistics management optimizes inventory levels, ensuring that essential resources such as bottles, corks, and labels are available when needed. This reduces downtime and improves the overall efficiency of the bottling process. The resulting enhancement in supply chain efficiency leads to cost savings and a more reliable production cycle, benefiting both wineries and consumers.
Several studies highlight the impact of AI on supply chain management. Toorajipour et al. conducted a systematic review, emphasizing AI's role in optimizing resource allocation, improving logistics, marketing, and production, and reducing operational inefficiencies. They underscored the importance of predictive analytics in anticipating disruptions and managing risks, crucial for a smooth and efficient wine industry supply chain [123].
Rajagopal et al. explored AI's role in supply chain finance, demonstrating its ability to streamline processes, predict risks, detect fraud, and optimize working capital. This is particularly relevant for wineries, where managing financial risks is crucial to maintaining smooth production and distribution processes [124].
In a complementary study, Dash et al. highlighted the benefits of AI in automating supply chain management. including forecast customer demand, optimize research and development, and enhance manufacturing processes. Their findings show how AI-driven predictive analytics leads to higher quality products and reduced costs, providing a competitive advantage for wineries looking to optimize their supply chains [125].
Further demonstrating the practical applications of these technologies, Masetti et al. introduced an IoT-based measurement system designed to monitor wine fermentation during the vinification process. This system, which consists of sensorized buoys and access points, provides real-time data on wine pH, liquid level, and temperature. Integrating such IoT systems with AI can further enhance supply chain efficiency by using real-time data for informed production decisions [126].
AI also plays a critical role in coordinating different production stages. For instance, AI systems can align harvesting schedules with processing capacity, ensuring grapes are processed at their peak freshness. This synchronization helps maintain product quality and reduces waste, improving overall supply chain efficiency.
In consequence, the integration of AI, IoT, and blockchain technologies has demonstrated substantial benefits. The combination of real-time monitoring, data analysis, and secure information sharing has enhanced traceability, transparency, and efficiency, ultimately leading to higher consumer trust and satisfaction. These practical applications underscore the transformative potential of advanced technologies in the wine industry, paving the way for more sustainable and innovative production practices.
5.2. Detecting and Managing Microbiological Risks
Microbiological risks are a significant concern in the wine industry, as they can affect both the safety and quality of the wine. AI plays a crucial role in detecting and managing these risks, focusing on the identification of mycotoxins, bacteria, wild yeasts, and other contaminants that can compromise the integrity of wine products. By leveraging advanced technologies such as machine learning, spectroscopy, and biosensors, AI systems provide winemakers with powerful tools to ensure the safety and quality of their wines.
5.2.1. Applications in Food Safety and Risk Management
AI applications extend beyond operational efficiencies to play a crucial role in enhancing food safety and managing microbiological risks. For instance, AI systems can detect mycotoxins and other contaminants, ensuring that the wine produced meets stringent safety standards. The use of machine learning models for predicting contamination events allows for proactive management and mitigation of risks, thereby safeguarding consumer health [127]. Liu et al. conducted a bibliometric review on artificial intelligence in food safety, highlighting its capability to detect contaminants like mycotoxins and predict contamination events [127].
Furthermore, the study by Qian et al. (2022) emphasized the broad applications of AI in food safety, including risk prediction, monitoring, and optimization throughout the supply chain [128]. The authors noted that AI could significantly enhance public health systems by providing early warnings of outbreaks and identifying sources of contamination.
Moreover, AI technologies can integrate data from various points in the supply chain to provide comprehensive traceability from vineyard to consumer. This capability is vital for maintaining consumer trust and meeting regulatory requirements. AI-driven traceability systems ensure that every bottle of wine can be traced back to its origin, improving transparency and accountability in the production process [129]. Ling et al. (2021) investigated intelligent supervision systems for food traceability, emphasizing how integrating data from multiple points in the supply chain can enhance transparency and accountability [129].
5.2.2. Identifying and Monitoring Mycotoxins
Mycotoxins are toxic compounds produced by certain molds that can contaminate grapes and affect wine safety. The presence of mycotoxins such as ochratoxin A (OTA) is a major concern for winemakers, as these compounds are harmful to human health and impact the sensory qualities of wine. AI systems utilize spectroscopic techniques, such as near-infrared (NIR) and ultraviolet-visible (UV-Vis) spectroscopy, to detect mycotoxins in grapes and wine. Spectroscopy provides a rapid, non-destructive method for analyzing samples. When combined with machine learning algorithms, it can identify specific mycotoxins based on their unique spectral signatures.
Research has shown that surface-enhanced Raman spectroscopy (SERS) is effective for detecting OTA in wine, providing a rapid and cost-effective alternative to traditional methods [130]. Additionally, non-invasive spectroscopic methods like two-photon-induced fluorescence (TPIF) have been employed to detect mycotoxins, offering advantages in terms of sensitivity and the ability to reduce background interference [131].
Machine learning models, such as support vector machines (SVM) and neural networks, are trained on spectral data to recognize patterns associated with mycotoxin contamination. These models can detect even low concentrations of mycotoxins, enabling early intervention to prevent contaminated grapes from entering the production process. For example, a novel enzyme-free biosensor using a toehold-mediated catalytic hairpin assembly demonstrated high sensitivity in detecting multiple mycotoxins, including OTA, aflatoxin B1, and zearalenone in wine samples [132].
Real-time monitoring of mycotoxin levels is also crucial for maintaining wine safety. AI-driven systems equipped with biosensors can continuously monitor mycotoxin levels during the winemaking process. These biosensors, often based on immunoassays or molecular recognition elements, provide specific and sensitive detection of mycotoxins. For instance, the development of a multiplex mycotoxin SERS immunoassay using functional gold nanotags has shown significant potential in providing rapid and accurate mycotoxin detection in wine [133].
The integration of biosensors with AI allows for real-time data analysis and decision-making. When mycotoxin levels exceed safe thresholds, the AI system can alert winemakers and suggest corrective actions, such as removing contaminated batches or adjusting processing parameters to mitigate risks. This proactive approach enhances the overall safety and quality of the wine. Techniques such as liquid chromatography-tandem mass spectrometry (LC-MS/MS) have also been optimized to efficiently determine multiple mycotoxins in wine, further supporting these AI-integrated systems [134].
5.2.2. Managing Microbiological Contaminants
In addition to mycotoxins, other microbiological contaminants such as bacteria and wild yeasts can pose significant risks during wine production. These microorganisms can cause spoilage, off-flavors, and health hazards if not properly managed. AI-powered microbial analysis tools use DNA sequencing and biosensors to monitor the microbial composition of wine at various stages of production. Metagenomic sequencing provides comprehensive insights into the microbial communities present in grape must, fermentation tanks, and aging barrels.
Machine learning algorithms analyze genetic sequences to identify the presence of harmful bacteria, such as acetic acid bacteria and lactic acid bacteria, which can cause spoilage and undesirable flavors. By understanding the dynamics of microbial populations, winemakers can take targeted actions to control bacterial growth. For example, AI systems can recommend adjustments to sulfur dioxide levels, temperature, and pH to inhibit the growth of spoilage bacteria while promoting the activity of beneficial microbes.
In a study by Renouf et al. (2007), the use of PCR-DGGE was employed to follow wine microbial consortia at each stage of the process, from grapes to bottles, and on cellar equipment surfaces. While this method allowed for a more exhaustive vision of the diversity of bacteria and yeast species and their evolution during winemaking, the integration of AI could enhance this approach by automating the analysis of complex microbial data. AI algorithms could process the DGGE data more efficiently, identifying trends and predicting potential contamination issues, thereby providing actionable insights faster [135].
Wild yeasts, such as Brettanomyces, can produce off-flavors and compromise wine quality. AI systems can monitor yeast populations through biosensors and genomic analysis, identifying potential contamination early in the fermentation process. By analyzing patterns in fermentation data, AI can predict Brettanomyces outbreaks and suggest preventative measures, such as adjusting sulfite levels or modifying storage conditions. Research has demonstrated that the use of real-time PCR can effectively detect Brettanomyces at early stages, allowing for timely intervention to prevent spoilage. Integrating AI with real-time PCR could automate the detection process, providing real-time alerts and recommended actions based on predictive models [136]. Additionally, impedance detection methods have shown potential in reducing the time required to identify Brettanomyces contamination, further enhancing proactive management strategies when coupled with AI-driven data analysis [137].
AI also assists in optimizing the fermentation process by managing the activity of Saccharomyces cerevisiae, the primary yeast used in winemaking. Real-time monitoring of yeast metabolism and population dynamics ensures that fermentation proceeds smoothly, reducing the risk of stuck or sluggish fermentations and ensuring consistent wine quality. An innovative oligonucleotide microarray developed by Cimaglia et al. was designed to detect various spoilage microorganisms in wine. Incorporating AI could enhance the microarray's functionality by enabling real-time data processing and predictive analytics, which would allow winemakers to respond to microbial threats more quickly and effectively [138].
Furthermore, Puértolas et al. explored the application of pulsed electric fields (PEF) technology for microbiological control in wineries, demonstrating its effectiveness in inactivating spoilage yeasts and bacteria such as Dekkera bruxellensis and Lactobacillus species. The integration of AI with PEF technology could optimize the application parameters, such as field strength and pulse duration, by continuously analyzing real-time microbial data and adjusting the process to ensure maximum effectiveness while preserving the sensory qualities of the wine [139].
5.2.3. Advanced Detection Techniques and Predictive Analytics
Beyond traditional methods, several advanced detection techniques are being integrated with AI to enhance microbiological risk management. These technologies offer precise, real-time insights into the microbial ecosystem of wine, ensuring that winemakers can proactively manage and mitigate potential risks.
Hyperspectral imaging captures a wide spectrum of light beyond the visible range, providing detailed information about the chemical composition of samples. When combined with AI, hyperspectral imaging can detect microbial contaminants that might be missed by conventional methods. Machine learning algorithms analyze hyperspectral data to identify unique signatures of various microbes, enabling early and accurate detection of contamination. A study by Cimaglia et al. demonstrated the potential of oligonucleotide microarrays in detecting spoilage microorganisms in wine, highlighting the importance of integrating advanced detection methods with AI to improve accuracy and speed of identification [138].
Lab-on-a-chip (LOC) devices miniaturize laboratory processes onto a single chip, allowing for rapid and on-site analysis of microbial contaminants. AI enhances the functionality of LOC devices by analyzing the data generated from these miniaturized assays in real-time. This integration allows for quick detection and quantification of bacteria, yeasts, and mycotoxins, providing winemakers with immediate insights to manage microbiological risks effectively. Dittrich and Manz discussed the advancements in LOC technologies, emphasizing their potential in various applications, including microbiological analysis in the wine industry [140].
Predictive analytics leverage historical data and real-time monitoring to assess the risk of microbial contamination. By analyzing trends and correlations, AI systems can predict potential contamination events and provide risk assessments. This allows winemakers to implement preventive measures proactively, reducing the likelihood of contamination and ensuring the safety of the final product. Bolinger et al. explored the use of microbiota and machine learning algorithms to assess the risk of Salmonella contamination in poultry rinsate. Their study demonstrated how combining microbiome data with AI can effectively predict contamination risks. The researchers analyzed 299 poultry rinsate samples using 16S sequencing data to identify microbial patterns indicative of Salmonella presence. They developed a Random Forest-based model that achieved high accuracy (88%), sensitivity (85%), and specificity (90%), showcasing the potential of AI in predictive analytics for food safety [141].
Additionally, next-generation sequencing (NGS) has revolutionized the way we study microbial communities. By providing high-throughput, detailed analyses of microbial DNA, NGS allows for comprehensive profiling of the wine microbiome. Morgan et al. highlighted the contributions of high-throughput sequencing and metagenomics to vineyard microbial ecology, showing how these advanced techniques can reveal microbial diversity and dynamics in winemaking [142]. These insights can be further enhanced by AI to predict and manage microbial risks more effectively.
AI can also enhance the use of flow cytometry for rapid detection of viable yeasts and bacteria in wine. Flow cytometry, combined with fluorescent dyes, can quickly estimate microbial counts and assess their viability. The integration of AI allows for real-time data analysis, improving the accuracy and speed of microbial detection. Longin et al. reviewed the application of flow cytometry in the wine industry, demonstrating its potential to rapidly detect and quantify microbial populations, including spoilage microorganisms like Brettanomyces. This study provided a foundation that could be significantly enhanced by integrating AI, which could further optimize data processing and predictive analytics for microbiological quality control [143].
6. Challenges and Future Opportunities
The integration of artificial intelligence (AI) in enology holds transformative potential, promising to revolutionize traditional winemaking practices. Despite the numerous advancements, the implementation of AI presents several challenges that must be addressed to fully exploit its capabilities. Concurrently, there are immense opportunities for future development that could reshape the industry. This section provides a comprehensive reflection on the challenges and future prospects of AI in winemaking, synthesizing insights from recent literature and studies.
One of the primary challenges in implementing AI in enology is the collection and management of large datasets required for training robust models. Winemaking involves numerous variables, from vineyard conditions to fermentation processes, each significantly influencing the final product's quality. Accurately capturing and standardizing this data across different regions and vineyards is a complex task. Variability in climate, soil types, and grape varieties necessitates AI models that are adaptable to diverse conditions, requiring extensive and high-quality data. This challenge is further compounded by the need for continuous data updates to ensure models remain relevant and effective.
In addition to data collection, the integration of AI systems into existing winemaking processes presents logistical and technical hurdles. Traditional winemaking practices have been refined over centuries, and introducing new technology requires significant changes in both mindset and infrastructure. Vineyard managers and winemakers may be resistant to adopting AI-driven methods due to a lack of understanding or trust in these technologies. Education and training are crucial to overcoming this barrier, ensuring that stakeholders are equipped to utilize AI tools effectively. Moreover, the integration process itself can be costly and time-consuming, requiring investment in new hardware, software, and personnel skilled in AI.
Another significant challenge is the interpretability of AI models. While machine learning algorithms, particularly deep learning models, have shown great promise in predicting outcomes and optimizing processes, their "black box" nature can be a drawback. Winemakers need to understand the rationale behind AI-driven recommendations to trust and adopt them fully. Developing more transparent and explainable AI models is essential to bridge this gap, allowing winemakers to see not just the results but the reasoning behind them.
Despite these challenges, the future opportunities presented by AI in enology are vast and varied. One of the most significant advantages of AI is its ability to enhance precision viticulture. AI-driven models can analyze data from drones, satellite imagery, and ground sensors to provide real-time insights into vineyard conditions. This capability enables proactive management of disease and pest outbreaks, optimized water and nutrient usage, and overall improved vineyard health. Such precision can lead to higher yields and better-quality grapes, directly impacting the quality of wine produced.
AI also offers substantial benefits in the fermentation and bottling stages of winemaking. Predictive models can monitor and control fermentation variables such as temperature, sugar levels, and yeast activity with high accuracy, ensuring consistent quality across batches. Automated systems powered by AI can handle the tedious task of bottling and labeling, reducing human error and increasing efficiency. These advancements not only streamline production but also maintain the high standards expected in the industry.
Moreover, AI's role in enhancing food safety and traceability is increasingly crucial. In an era where consumers are more concerned about the origins and safety of their food and beverages, AI can ensure that every step of the winemaking process is monitored and recorded. This capability enhances transparency and accountability, providing consumers with the assurance that their wine meets stringent safety standards.
Looking forward, the potential for AI to drive sustainable practices in winemaking is a particularly exciting opportunity. By optimizing resource usage, AI can help reduce the environmental impact of viticulture. Precision irrigation and nutrient management reduce waste and promote sustainable farming practices. Additionally, AI can assist in developing new grape varieties that are more resistant to climate change and diseases, ensuring the long-term viability of vineyards.
Integrating AI into winemaking also opens avenues for innovation in product development and marketing. By analyzing consumer preferences and market trends, AI can help winemakers tailor their products to meet evolving demands. This data-driven approach can lead to the creation of unique wine profiles that cater to niche markets, enhancing both profitability and consumer satisfaction.
Furthermore, the collaborative potential of AI in research and development should not be overlooked. AI can facilitate the sharing of data and insights across different stakeholders in the wine industry, fostering a collaborative environment that drives innovation. This collective approach can lead to the development of more comprehensive and effective AI solutions that benefit the entire industry.
However, as AI becomes more integrated into enology, ethical issues and regulatory frameworks must also be considered to ensure responsible development and deployment. Ethical considerations include data privacy, accountability, labor impacts, and environmental sustainability. According to González-Rodríguez et al. [15], AI-enabled crop management regimes could reinforce unsustainable industrial farming at the cost of rural livelihoods, localized knowledge, and the food sovereignty of smallholder farmers. Therefore, it is imperative that AI systems in enology are designed through inclusive stakeholder participation, centering human needs and values.
Data privacy is a significant concern, as the collection of large agricultural datasets for training AI models could include proprietary information from farmers alongside field images or soil data. Researchers must implement ethical data management practices that protect farmer interests and anonymity. Furthermore, the proprietary nature of some commercial AI technologies, often referred to as "black-box" systems, obscures model biases and prevents oversight into decision-making rationales. Such opacity becomes ethically problematic for AI systems deployed in social realms including agriculture.
Additionally, there is apprehension surrounding the potential marginalization of rural communities as data-driven farming becomes more prevalent. To avoid this, human-centered design considerations must shape AI integration in enology, ensuring that it serves to augment rather than replace agricultural expertise and intuition. Ongoing farmer education and upskilling initiatives are essential to democratize AI access, allowing rural communities to reap the benefits equitably and partake in co-developing solutions attuned to local needs.
The establishment of regulatory frameworks and standards is also crucial for the successful implementation of AI in enology. Currently, there is a lack of governance surrounding the creation, sales, and monitoring of AI technologies in agriculture. Policy interventions are required at national and global levels to regulate the quality control, risk assessments, and liability attribution of agricultural AI systems. Such oversight can mitigate the dangers of hastily implemented tools with unreliable real-world performances or unexamined biases causing harm. Global agreements on technological approaches, architecture choices, data formats, curation protocols, and performance benchmarks will support collaboration, open data sharing, and interoperability, accelerating innovation. Additionally, voluntary professional codes of ethics around topics such as model transparency, auditability, and farmer privacy could guide institutional research and industry product design.
In summary, while the integration of AI in enology presents several challenges, including data management, integration into traditional practices, model interpretability, ethical issues, and the need for regulatory frameworks, the opportunities it offers far outweigh these obstacles. AI has the potential to revolutionize viticulture and winemaking by enhancing precision, efficiency, quality, and sustainability. As the technology continues to evolve, its adoption in the winemaking industry is likely to grow, driven by the significant benefits it can provide. The future of winemaking, bolstered by AI, promises to be more innovative, efficient, and sustainable, paving the way for a new era in enology.
7. Conclusion
Artificial intelligence (AI) has brought significant advancements to the winemaking industry, enhancing product consistency, operational efficiency, and safety. By leveraging AI-driven technologies, winemakers can achieve precise control over various aspects of production, from vineyard management to bottling. AI systems enable real-time monitoring of critical parameters, predictive modeling for fermentation, and advanced quality control during bottling, ensuring that each stage of the process meets high standards of quality and efficiency.
AI's ability to integrate data from IoT devices and employ machine learning algorithms for predictive analytics has revolutionized traceability and food safety in the wine industry. By detecting and managing microbiological risks, such as mycotoxins and bacterial contamination, AI ensures the production of safe and high-quality wines. The use of AI in enhancing traceability from vineyard to consumer also builds consumer trust and meets regulatory requirements.
Despite the numerous benefits, the adoption of AI in winemaking faces challenges such as data integration issues, high implementation costs, and the need for robust AI models that can generalize across diverse conditions. Addressing these challenges requires collaboration between wineries, technology providers, and research institutions to develop standardized protocols and cost-effective AI solutions.
Data management remains a fundamental challenge due to the vast variability in climate, soil types, and grape varieties. The collection, standardization, and continuous updating of high-quality data are essential for training robust AI models that can adapt to diverse conditions. The technical and logistical hurdles in integrating AI systems into existing winemaking processes require significant investments in infrastructure, education, and training to ensure effective adoption and utilization.
Another critical issue is the interpretability of AI models. To foster trust and adoption among winemakers, AI models must be transparent and explainable, providing clear insights into the rationale behind their recommendations. This need for transparency underscores the importance of developing AI models that are not only accurate but also understandable and actionable.
Looking ahead, the future of AI in winemaking is promising. Prospective advancements include enhanced sensory analysis tools, deeper integration with global supply chain systems, and personalized wine recommendations. AI-driven sustainable winemaking practices, such as optimized resource usage and integrated pest management, will further promote environmental stewardship. By analyzing consumer preferences and market trends, AI can help winemakers tailor their products to meet evolving demands, leading to the creation of unique wine profiles that cater to niche markets.
Ethical considerations and the establishment of regulatory frameworks are crucial for the responsible development and deployment of AI in enology. Addressing issues of data privacy, accountability, labor impacts, and environmental sustainability is essential to ensure that AI integration benefits all stakeholders. Regulatory frameworks and standards will provide the necessary oversight to mitigate risks and ensure the ethical and effective use of AI technologies.
Furthermore, the collaborative potential of AI in research and development should not be overlooked. AI can facilitate the sharing of data and insights across different stakeholders in the wine industry, fostering a collaborative environment that drives innovation. This collective approach can lead to the development of more comprehensive and effective AI solutions that benefit the entire industry.
To conclude, AI technologies have the potential to revolutionize traditional winemaking practices, improving efficiency, quality, and sustainability. By embracing AI, the wine industry can address broader challenges and continue to innovate, ensuring a bright future for winemaking. The integration of AI in enology, though complex and multifaceted, promises to lead to more innovative, efficient, and sustainable practices, heralding a new era in winemaking that aligns with the evolving demands of the global market and the pressing need for sustainability in agricultural practices.
Funding
This research was supported by a grant from PID2021-1228990-B-C22 funded by MCIN/AEI/ 10.13039/501100011033 and by and by ERDF/EU.
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Conflicts of Interest
The authors declare no conflicts of interest
References
- Russell, S.; Norvig, P. Artificial Intelligence: A Modern Approach (4th Edition); Pearson, 2020; ISBN 9780134610993.
- Esteva, A.; Robicquet, A.; Ramsundar, B.; Kuleshov, V.; DePristo, M.; Chou, K.; Cui, C.; Corrado, G.; Thrun, S.; Dean, J. A Guide to Deep Learning in Healthcare. Nature Medicine 2019 25:1 2019, 25, 24–29. [CrossRef]
- Liakos, K.G.; Busato, P.; Moshou, D.; Pearson, S.; Bochtis, D. Machine Learning in Agriculture: A Review. Sensors 2018, Vol. 18, Page 2674 2018, 18, 2674. [CrossRef]
- Liakos, K.G.; Busato, P.; Moshou, D.; Pearson, S.; Bochtis, D. Machine Learning in Agriculture: A Review. Sensors 2018, Vol. 18, Page 2674 2018, 18, 2674. [CrossRef]
- Kamilaris, A.; Prenafeta-Boldú, F.X. Deep Learning in Agriculture: A Survey. Comput Electron Agric 2018, 147, 70–90. [CrossRef]
- Chouhan, M.; Banerjee, P.S.; Kumar, A. Identifying the Suitability of Artificial Intelligence Technology for Modern Farming. ACM International Conference Proceeding Series 2023, 207–215. [CrossRef]
- Chlingaryan, A.; Sukkarieh, S.; Whelan, B. Machine Learning Approaches for Crop Yield Prediction and Nitrogen Status Estimation in Precision Agriculture: A Review. Comput Electron Agric 2018, 151, 61–69. [CrossRef]
- Morales, A.; Villalobos, F.J. Using Machine Learning for Crop Yield Prediction in the Past or the Future. Front Plant Sci 2023, 14, 1128388. [CrossRef]
- Bonissone, P.P.; Liani, A. Industrial AI [Industrial and Governmental Activities]. IEEE Comput Intell Mag 2023, 18, 10–12. [CrossRef]
- Xue, B.; Green, R.; Zhang, M. Artificial Intelligence in New Zealand: Applications and Innovation. J R Soc N Z 2023, 53, 1–5. [CrossRef]
- Nelson, J.; Coleman, R.; Chacón-Rodríguez, L.; Runnebaum, R.; Boulton, R.; Knoesen, A. Advanced Monitoring and Control of Redox Potential in Wine Fermentation across Scales. Fermentation 2023, Vol. 9, Page 7 2022, 9, 7. [CrossRef]
- Lijffijt, J.; Gkorou, D.; Van Hertum, P.; Ypma, A.; Pechenizkiy, M.; Vanschoren, J. Introduction to the Special Section on AI in Manufacturing. ACM SIGKDD Explorations Newsletter 2022, 24, 81–85. [CrossRef]
- Qian, C.; Murphy, S.I.; Orsi, R.H.; Wiedmann, M. How Can AI Help Improve Food Safety? Annu Rev Food Sci Technol 2023, 14, 517–538. [CrossRef]
- Liu, Z.; Wang, S.; Zhang, Y.; Feng, Y.; Liu, J.; Zhu, H. Artificial Intelligence in Food Safety: A Decade Review and Bibliometric Analysis. Foods 2023, Vol. 12, Page 1242 2023, 12, 1242. [CrossRef]
- González-Rodríguez, V.E.; Izquierdo-Bueno, I.; Cantoral, J.M.; Carbú, M.; Garrido, C. Artificial Intelligence: A Promising Tool for Application in Phytopathology. Horticulturae 2024, Vol. 10, Page 197 2024, 10, 197. [CrossRef]
- Sanin, C. Artificial Intelligence: Current Perspectives and Alternative Paths. TecnoLógicas 2023, 26, e2731. [CrossRef]
- Xu, D.; Liu, B.; Wang, J.; Zhang, Z. Bibliometric Analysis of Artificial Intelligence for Biotechnology and Applied Microbiology: Exploring Research Hotspots and Frontiers. Front Bioeng Biotechnol 2022, 10, 998298. [CrossRef]
- Cardon, P. Searching for the Right Metaphors to Understand and Interrogate the AI Age. Business Communication Research and Practice 2023, 6, 65–69. [CrossRef]
- Bonissone, P.P.; Liani, A. Industrial AI [Industrial and Governmental Activities]. IEEE Comput Intell Mag 2023, 18, 10–12. [CrossRef]
- Pang, G.; Pang, G. The AI Chip Race. IEEE Intell Syst 2022, 37, 111–112. [CrossRef]
- Huh, J.; Nelson, M.R.; Russell, C.A. ChatGPT, AI Advertising, and Advertising Research and Education. J Advert 2023, 52, 477–482. [CrossRef]
- Colmenarejo, A.B.; Nannini, L.; Rieger, A.; Scott, K.M.; Zhao, X.; Patro, G.K.; Kasneci, G.; Kinder-Kurlanda, K. State of AI Ethics Report (Volume 6, February 2022). ArXiv 2022, abs/2202.07435, 107–118. [CrossRef]
- Gardashova, L.; Ismayilova, A.; Tarverdiyeva, G.; Expert System for Dental Diseases. In Proceedings of the PAHTEI-Procedings of Azerbaijan High Technical Educational Institutions; 2023; pp. 23–30.
- Menekse, M. Envisioning the Future of Learning and Teaching Engineering in the Artificial Intelligence Era: Opportunities and Challenges. Journal of Engineering Education 2023, 112, 578–582. [CrossRef]
- Sona Vikas, Mandeep Kaur, R.A.G. Tracing the Advancements in Artificial Intelligence: Drawing Lessons for Tomorrow. Towards Excellence 2022, 14.
- Fitsilis, F.; von Lucke, J.; Frank, S. A Comprehensive Research Workshop on Artificial Intelligence in Parliaments. International Journal of Parliamentary Studies 2023, 379. [CrossRef]
- Schlagwein, D.; Willcocks, L. ‘ChatGPT et al.’: The Ethics of Using (Generative) Artificial Intelligence in Research and Science. 10.1177/02683962231200411 2023, 38, 232–238. [CrossRef]
- Kovalenko, I.; Barton, K.; Moyne, J.; Tilbury, D.M. Opportunities and Challenges to Integrate Artificial Intelligence Into Manufacturing Systems: Thoughts From a Panel Discussion [Opinion]. IEEE Robot Autom Mag 2023, 30, 109–112. [CrossRef]
- Prajapati, S.; Qureshi, S.; Rao, Y.; Nadkarni, S.; Retharekar, M.; Avhad, A. Plant Disease Identification Using Deep Learning. In Proceedings of the 2023 4th International Conference for Emerging Technology, INCET 2023; Institute of Electrical and Electronics Engineers Inc., May 26 2023; pp. 1–5.
- Elbasi, E.; Mostafa, N.; Alarnaout, Z.; Zreikat, A.I.; Cina, E.; Varghese, G.; Shdefat, A.; Topcu, A.E.; Abdelbaki, W.; Mathew, S.; et al. Artificial Intelligence Technology in the Agricultural Sector: A Systematic Literature Review. IEEE Access 2023, 11, 171–202. [CrossRef]
- Xue, B.; Green, R.; Zhang, M. Artificial Intelligence in New Zealand: Applications and Innovation. J R Soc N Z 2023, 53, 1–5. [CrossRef]
- González-Rodríguez, V.E.; Izquierdo-Bueno, I.; Cantoral, J.M.; Carbú, M.; Garrido, C. Artificial Intelligence: A Promising Tool for Application in Phytopathology. Horticulturae 2024, Vol. 10, Page 197 2024, 10, 197. [CrossRef]
- Bechar, A.; Vigneault, C. Agricultural Robots for Field Operations: Concepts and Components. Biosyst Eng 2016, 149, 94–111. [CrossRef]
- Kamilaris, A.; Kartakoullis, A.; Prenafeta-Boldú, F.X. A Review on the Practice of Big Data Analysis in Agriculture. Comput Electron Agric 2017, 143, 23–37. [CrossRef]
- Verdouw, C.N.; Wolfert, J.; Beulens, A.J.M.; Rialland, A. Virtualization of Food Supply Chains with the Internet of Things. J Food Eng 2016, 176, 128–136. [CrossRef]
- Marcu, I.; Drăgulinescu, A.M.; Oprea, C.; Suciu, G.; Bălăceanu, C. Predictive Analysis and Wine-Grapes Disease Risk Assessment Based on Atmospheric Parameters and Precision Agriculture Platform. Sustainability 2022, Vol. 14, Page 11487 2022, 14, 11487. [CrossRef]
- Pantazi, X.E.; Moshou, D.; Alexandridis, T.; Whetton, R.L.; Mouazen, A.M. Wheat Yield Prediction Using Machine Learning and Advanced Sensing Techniques. Comput Electron Agric 2016, 121, 57–65. [CrossRef]
- Kleb, M.; Merkt, N.; Zörb, C. New Aspects of In Situ Measurements for Downy Mildew Forecasting. Plants 2022, Vol. 11, Page 1807 2022, 11, 1807. [CrossRef]
- Jovanovska, E.M.; Chorbev, I.; Davcev, D.; Mitreski, K. Integrated IoT System for Prediction of Diseases in the Vineyards. International Conference on Electrical, Computer, Communications and Mechatronics Engineering, ICECCME 2022 2022. [CrossRef]
- Maddalena, G.; Marone Fassolo, E.; Bianco, P.A.; Toffolatti, S.L. Disease Forecasting for the Rational Management of Grapevine Mildews in the Chianti Bio-District (Tuscany). Plants 2023, Vol. 12, Page 285 2023, 12, 285. [CrossRef]
- Rossi, V.; Giosuè, S.; Caffi, T. Modelling Plant Diseases for Decision Making in Crop Protection. Precision Crop Protection - The Challenge and Use of Heterogeneity 2010, 241–258. [CrossRef]
- Madeira, M.; Porfírio, R.P.; Santos, P.A.; Madeira, R.N. AI-Powered Solution for Plant Disease Detection in Viticulture. Procedia Comput Sci 2024, 238, 468–475. [CrossRef]
- Anushi; Jain, S.; Bhujel, S.; Shrivastava, U.; Rishabh; Mohapatra, A.; Rimpika; Mishra, G. Advancements in Drone Technology for Fruit Crop Management: A Comprehensive Review. International Journal of Environment and Climate Change 2023, 13, 4367–4378. [CrossRef]
- Dubuis, P.H.; Jaquerod, A. Evaluation of the Performance of Drone Treatments to Control Downy and Powdery Mildew in Grapevines. BIO Web Conf 2022, 50, 01006. [CrossRef]
- Slimani, H.; Mhamdi, J. El; Jilbab, A. Drone-Assisted Plant Disease Identification Using Artificial Intelligence: A Critical Review. International Journal of Computing and Digital Systems 2023, 14, 10433–10446. [CrossRef]
- Rosado, L.; Faria, P.; Gonçalves, J.; Silva, E.; Vasconcelos, A.; Braga, C.; Oliveira, J.; Gomes, R.; Barbosa, T.; Ribeiro, D.; et al. EyesOnTraps: AI-Powered Mobile-Based Solution for Pest Monitoring in Viticulture. Sustainability 2022, Vol. 14, Page 9729 2022, 14, 9729. [CrossRef]
- Nasim, S.; Rashid, M.; Syed, S.A.; Brohi, I. Intelligent Agricultural Pest Manager Drone in Pakistan. Pakistan Journal of Biotechnology 2023, 20, 238–242. [CrossRef]
- Žibrat, U.; Knapič, M. Detection of Grapevine Yellows Using Multispectral Imaging. Remote Sensing in Precision Agriculture: Transforming Scientific Advancement into Innovation 2024, 367–386. [CrossRef]
- Albetis, J.; Jacquin, A.; Goulard, M.; Poilvé, H.; Rousseau, J.; Clenet, H.; Dedieu, G.; Duthoit, S. On the Potentiality of UAV Multispectral Imagery to Detect Flavescence Dorée and Grapevine Trunk Diseases. Remote Sensing 2019, Vol. 11, Page 23 2018, 11, 23. [CrossRef]
- González-Jorge, H.; Miguel González-deSantos, L.; Fariñas-Álvarez, N.; Martínez-Sánchez, J. Operational Study of Drone Spraying Application of Phytosanitary Products in Vineyards. Dyna Ingenieria e Industria 2023, 97, 23–26. [CrossRef]
- di Gennaro, S.F.; Battiston, E.; di Marco, S.; Facini, O.; Matese, A.; Nocentini, M.; Palliotti, A.; Mugnai, L. Unmanned Aerial Vehicle (UAV)-Based Remote Sensing to Monitor Grapevine Leaf Stripe Disease within a Vineyard Affected by Esca Complex. Phytopathol Mediterr 2016, 55, 262–275. [CrossRef]
- Junges, A.H.; Ducati, J.R.; Scalvi Lampugnani, C.; Almança, M.A.K. Detection of Grapevine Leaf Stripe Disease Symptoms by Hyperspectral Sensor. Phytopathol Mediterr 2018, 57, 399–406. [CrossRef]
- MacDonald, S.L.; Staid, M.; Staid, M.; Cooper, M.L. Remote Hyperspectral Imaging of Grapevine Leafroll-Associated Virus 3 in Cabernet Sauvignon Vineyards. Comput Electron Agric 2016, 130, 109–117. [CrossRef]
- Vélez, S.; Ariza-Sentís, M.; Valente, J. Mapping the Spatial Variability of Botrytis Bunch Rot Risk in Vineyards Using UAV Multispectral Imagery. European Journal of Agronomy 2023, 142, 126691. [CrossRef]
- Ateya, A.A.; Algarni, A.D.; Koucheryavy, A.; Soliman, N.F. Drone-Based AI/IoT Framework for Monitoring, Tracking and Fighting Pandemics. Computers, Materials & Continua 2022, 71, 4677–4699. [CrossRef]
- Chen, J.; Ma, A.; Huang, L.; Su, Y.; Li, W.; Zhang, H.; Wang, Z. GA-YOLO: A Lightweight YOLO Model for Dense and Occluded Grape Target Detection. Horticulturae 2023, Vol. 9, Page 443 2023, 9, 443. [CrossRef]
- Xu, Z.; Liu, J.; Wang, J.; Cai, L.; Jin, Y.; Zhao, S.; Xie, B. Realtime Picking Point Decision Algorithm of Trellis Grape for High-Speed Robotic Cut-and-Catch Harvesting. Agronomy 2023, Vol. 13, Page 1618 2023, 13, 1618. [CrossRef]
- Wu, N.; Huang, H.; Meng, X.; Xiong, Y.; Li, S. Design and Kinematic Modeling of Grape Picking Robot Arm. ACM International Conference Proceeding Series 2023. [CrossRef]
- Liu, J.; Liang, J.; Zhao, S.; Jiang, Y.; Wang, J.; Jin, Y.; Zhang, B.; Qiao, Y.; Liu, J.; Liang, J.; et al. Design of a Virtual Multi-Interaction Operation System for Hand–Eye Coordination of Grape Harvesting Robots. Agronomy 2023, Vol. 13, Page 829 2023, 13, 829. [CrossRef]
- Milburn, L.; Gamba, J.; Fernandes, M.; Semini, C. Computer-Vision Based Real Time Waypoint Generation for Autonomous Vineyard Navigation with Quadruped Robots. 2023 IEEE International Conference on Autonomous Robot Systems and Competitions, ICARSC 2023 2023, 239–244. [CrossRef]
- Zhang, T.; Wu, F.; Wang, M.; Chen, Z.; Li, L.; Zou, X. Grape-Bunch Identification and Location of Picking Points on Occluded Fruit Axis Based on YOLOv5-GAP. Horticulturae 2023, Vol. 9, Page 498 2023, 9, 498. [CrossRef]
- Jiang, Y.; Liu, J.; Wang, J.; Li, W.; Peng, Y.; Shan, H. Development of a Dual-Arm Rapid Grape-Harvesting Robot for Horizontal Trellis Cultivation. Front Plant Sci 2022, 13, 881904. [CrossRef]
- Chen, J.; Ma, A.; Huang, L.; Su, Y.; Li, W.; Zhang, H.; Wang, Z. GA-YOLO: A Lightweight YOLO Model for Dense and Occluded Grape Target Detection. Horticulturae 2023, Vol. 9, Page 443 2023, 9, 443. [CrossRef]
- Coll-Ribes, G.; Torres-Rodríguez, I.J.; Grau, A.; Guerra, E.; Sanfeliu, A. Accurate Detection and Depth Estimation of Table Grapes and Peduncles for Robot Harvesting, Combining Monocular Depth Estimation and CNN Methods. Comput Electron Agric 2023, 215, 108362. [CrossRef]
- Lowe, M.; Qin, R.; Mao, X. A Review on Machine Learning, Artificial Intelligence, and Smart Technology in Water Treatment and Monitoring. Water 2022, Vol. 14, Page 1384 2022, 14, 1384. [CrossRef]
- Mahardhika, S.P.; Putriani, O. Deployment and Use of Artificial Intelligence (AI) in Water Resources and Water Management. IOP Conf Ser Earth Environ Sci 2023, 1195, 012056. [CrossRef]
- Xue, B.; Green, R.; Zhang, M. Artificial Intelligence in New Zealand: Applications and Innovation. J R Soc N Z 2023, 53, 1–5. [CrossRef]
- Darouich, H.; Ramos, T.B.; Pereira, L.S.; Rabino, D.; Bagagiolo, G.; Capello, G.; Simionesei, L.; Cavallo, E.; Biddoccu, M. Water Use and Soil Water Balance of Mediterranean Vineyards under Rainfed and Drip Irrigation Management: Evapotranspiration Partition and Soil Management Modelling for Resource Conservation. Water 2022, Vol. 14, Page 554 2022, 14, 554. [CrossRef]
- Alibabaei, K.; Gaspar, P.D.; Campos, R.M.; Rodrigues, G.C.; Lopes, C.M. Evaluation of a Deep Learning Approach for Predicting the Fraction of Transpirable Soil Water in Vineyards. Applied Sciences 2023, Vol. 13, Page 2815 2023, 13, 2815. [CrossRef]
- D’amore, G.; Di Vaio, A.; Balsalobre-Lorente, D.; Boccia, F. Artificial Intelligence in the Water–Energy–Food Model: A Holistic Approach towards Sustainable Development Goals. Sustainability 2022, Vol. 14, Page 867 2022, 14, 867. [CrossRef]
- Peng, X.; Chen, D.; Zhou, Z.; Zhang, Z.; Xu, C.; Zha, Q.; Wang, F.; Hu, X. Prediction of the Nitrogen, Phosphorus and Potassium Contents in Grape Leaves at Different Growth Stages Based on UAV Multispectral Remote Sensing. Remote Sensing 2022, Vol. 14, Page 2659 2022, 14, 2659. [CrossRef]
- Lyu, H.; Grafton, M.; Ramilan, T.; Irwin, M.; Sandoval, E. Assessing the Leaf Blade Nutrient Status of Pinot Noir Using Hyperspectral Reflectance and Machine Learning Models. Remote Sensing 2023, Vol. 15, Page 1497 2023, 15, 1497. [CrossRef]
- Wilson, S.G.; Rees, G.L.; O’Geen, A.T. A Super Learner Ensemble to Map Potassium Fixation in California Vineyard Soils. Geoderma 2024, 445, 116824. [CrossRef]
- Kumar, J.; Chawla, R.; Katiyar, D.; Chouriya, A.; Nath, D.; Sahoo, S.; Ali, A.; Singh, B. veer Optimizing Irrigation and Nutrient Management in Agriculture through Artificial Intelligence Implementation. International Journal of Environment and Climate Change 2023, 13, 4016–4022. [CrossRef]
- Rachmadi, M.A.; Nugroho, T.H.; Taufiqurrohman, H.; Cahya, Z.; Cahya, D.E.; Prihantoro, G. Precise Temperature Control Strategy for Laboratory-Scale Fermentation Process. Proceeding - 2023 International Conference on Radar, Antenna, Microwave, Electronics, and Telecommunications: Empowering Global Progress: Innovative Electronic and Telecommunication Solutions for a Sustainable Future, ICRAMET 2023 2023, 36–40. [CrossRef]
- Malavasi, M.; Cattani, L.; Bozzoli, F.; Rainieri, S. Development of an Innovative Temperature Control System in a Fermenter: Application to the Case of the Wine Industry. J Phys Conf Ser 2022, 2177, 012036. [CrossRef]
- Chiou, C.W.; Huang, S.-Z.; Sun, Y.-S. A Mobile Control System of Fermentation Temperature on Winemaking. International Journal of Food Science and Agriculture 2022, 6, 372–378. [CrossRef]
- Samphao, A.; Butmee, P.; Saejueng, P.; Pukahuta, C.; Švorc, Ľ.; Kalcher, K. Monitoring of Glucose and Ethanol during Wine Fermentation by Bienzymatic Biosensor. Journal of Electroanalytical Chemistry 2018, 816, 179–188. [CrossRef]
- Yang, S.; Bo, C.; Li, J.; Tan, Q.; Wang, Y.; Wu, H.; Chu, Z.; Jin, W. Online Glucose Analysis System With High Precision and Wide Detection Range for Fermentation Monitoring. IEEE Trans Instrum Meas 2022, 71. [CrossRef]
- Fuller, H.; Beaver, C.; Harbertson, J.; Jordão, M.; Michelle Wine Estates, S. Alcoholic Fermentation Monitoring and PH Prediction in Red and White Wine by Combining Spontaneous Raman Spectroscopy and Machine Learning Algorithms. Beverages 2021, Vol. 7, Page 78 2021, 7, 78. [CrossRef]
- Torello Pianale, L.; Rugbjerg, P.; Olsson, L. Real-Time Monitoring of the Yeast Intracellular State During Bioprocesses With a Toolbox of Biosensors. Front Microbiol 2022, 12, 802169. [CrossRef]
- Flores-Hernández, A.A.; Reyes-Reyes, J.; Astorga-Zaragoza, C.M.; Osorio-Gordillo, G.L.; García-Beltrán, C.D. Temperature Control of an Alcoholic Fermentation Process through the Takagi–Sugeno Modeling. Chemical Engineering Research and Design 2018, 140, 320–330. [CrossRef]
- Urtubia, A.; León, R.; Vargas, M. Identification of Chemical Markers to Detect Abnormal Wine Fermentation Using Support Vector Machines. Comput Chem Eng 2021, 145, 107158. [CrossRef]
- Schwinn, M.; Durner, D.; Wacker, M.; Delgado, A.; Fischer, U. Impact of Fermentation Temperature on Required Heat Dissipation, Growth and Viability of Yeast, on Sensory Characteristics and on the Formation of Volatiles in Riesling. Aust J Grape Wine Res 2019, 25, 173–184. [CrossRef]
- Xu, R.Z.; Cao, J.S.; Wu, Y.; Wang, S.N.; Luo, J.Y.; Chen, X.; Fang, F. An Integrated Approach Based on Virtual Data Augmentation and Deep Neural Networks Modeling for VFA Production Prediction in Anaerobic Fermentation Process. Water Res 2020, 184, 116103. [CrossRef]
- Florea, A.; Sipos, A.; Stoisor, M.C. Applying AI Tools for Modeling, Predicting and Managing the White Wine Fermentation Process. Fermentation 2022, Vol. 8, Page 137 2022, 8, 137. [CrossRef]
- Juuso, E.K. Intelligent Dynamic Simulation of Fed-Batch Fermentation Processes. Proceedings of The 60th SIMS Conference on Simulation and Modelling SIMS 2019, August 12-16, Västerås, Sweden 2020, 170, 132–138. [CrossRef]
- Pantazi, X.E.; Moshou, D.; Alexandridis, T.; Whetton, R.L.; Mouazen, A.M. Wheat Yield Prediction Using Machine Learning and Advanced Sensing Techniques. Comput Electron Agric 2016, 121, 57–65. [CrossRef]
- Cavaglia, J.; Schorn-García, D.; Giussani, B.; Ferré, J.; Busto, O.; Aceña, L.; Mestres, M.; Boqué, R. ATR-MIR Spectroscopy and Multivariate Analysis in Alcoholic Fermentation Monitoring and Lactic Acid Bacteria Spoilage Detection. Food Control 2020, 109, 106947. [CrossRef]
- Berbegal, C.; Khomenko, I.; Russo, P.; Spano, G.; Fragasso, M.; Biasioli, F.; Capozzi, V. PTR-ToF-MS for the Online Monitoring of Alcoholic Fermentation in Wine: Assessment of VOCs Variability Associated with Different Combinations of Saccharomyces/Non-Saccharomyces as a Case-Study. Fermentation 2020, Vol. 6, Page 55 2020, 6, 55. [CrossRef]
- Schenk, C.; Schulz, V.H. Existence, Uniqueness, and Numerical Modeling of Wine Fermentation Based on Integro-Differential Equations. 2022, 82, 1220–1245. [CrossRef]
- Bakošová, M.; Oravec, J.; Vasičkaninová, A.; Mészáros, A.; Artzová, P. Advanced Control of a Biochemical Reactor for Yeast Fermentation. Chem Eng Trans 2019, 76, 769–774. [CrossRef]
- Soares, M.C.P.; Fujiwara, E.; Gomes, M.K.; Mendes, B.F.; La Torre, L.G.; Suzuki, C.K. Application of Optical Fiber Sensor on Fermentation Monitoring. 26th International Conference on Optical Fiber Sensors (2018), paper ThE39 2018, Part F124-OFS 2018, ThE39. [CrossRef]
- Itto-Nakama, K.; Watanabe, S.; Kondo, N.; Ohnuki, S.; Kikuchi, R.; Nakamura, T.; Ogasawara, W.; Kasahara, K.; Ohya, Y. AI-Based Forecasting of Ethanol Fermentation Using Yeast Morphological Data. Biosci Biotechnol Biochem 2021, 86, 125–134. [CrossRef]
- Rivera, E.C.; Yamakawa, C.K.; Rossell, C.E.V.; Nolasco, J.; Kwon, H.J. Prediction of Intensified Ethanol Fermentation of Sugarcane Using a Deep Learning Soft Sensor and Process Analytical Technology. Journal of Chemical Technology & Biotechnology 2024, 99, 207–216. [CrossRef]
- Silva, R.L.; Rudek, M.; Szejka, A.L.; Canciglieri Junior, O. Machine Vision Systems for Industrial Quality Control Inspections. IFIP Adv Inf Commun Technol 2018, 540, 631–641. [CrossRef]
- Mahor, A.; Yadav, R.S. In-Process Intelligent Inspection of the Specimen Using Machine Vision. ASME International Mechanical Engineering Congress and Exposition, Proceedings (IMECE) 2023, 2-B. [CrossRef]
- Ren, Z.; Fang, F.; Yan, N.; Wu, Y. State of the Art in Defect Detection Based on Machine Vision. International Journal of Precision Engineering and Manufacturing-Green Technology 2021 9:2 2021, 9, 661–691. [CrossRef]
- Gong, W.; Zhang, K.; Yang, C.; Yi, M.; Wu, J. Adaptive Visual Inspection Method for Transparent Label Defect Detection of Curved Glass Bottle. Proceedings - 2020 International Conference on Computer Vision, Image and Deep Learning, CVIDL 2020 2020, 90–95. [CrossRef]
- Wu, M.Y. Wine Label Image Recognition Using Convolutional Neural Network with Augmented Data. International Conferences Interfaces and Human Computer Interaction 2019; Game and Entertainment Technologies 2019; and Computer Graphics, Visualization, Computer Vision and Image Processing 2019 2019, 479–481. [CrossRef]
- Li, X.; Ma, J. Distributed Search and Fusion for Wine Label Image Retrieval. PeerJ Comput Sci 2022, 8, e1116. [CrossRef]
- Liu, F.; Tang, J.; Yang, J.; Wang, H. Automated Industrial Crack Inspection System Based on Edge-Edge Collaboration of Multiple Cameras and Programmable Logic Controller. IEEE International Symposium on Broadband Multimedia Systems and Broadcasting, BMSB 2023, 2023-June. [CrossRef]
- Mikhail, M.; Abad, K. Robotic Inspection and Automated Analysis System for Advanced Manufacturing. IAES International Journal of Robotics and Automation (IJRA) 2023, 12, 352–364. [CrossRef]
- Vagas, M.; Galajdová, A.; Džongov, M. Proposal of a Vision System for Automated Line MPS 500. TECHNICAL SCIENCES AND TECHNOLOGIES 2019, 108–113. [CrossRef]
- Tung, L.C.; Chen, S.H.; Ye, Y.C.; Liu, G.Z.; Chen, L.Y.; Tu, Y.R.; Lai, Y.W.; Lee, J.D. Implementation of Multi-Process Automatic Loading and Unloading Production System. 2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023 - Proceedings 2023, 121–122. [CrossRef]
- Bahaghighat, M.; Mirfattahi, M.; Akbari, L.; Babaie, M. Designing Quality Control System Based on Vision Inspection in Pharmaceutical Product Lines. 2018 International Conference on Computing, Mathematics and Engineering Technologies: Invent, Innovate and Integrate for Socioeconomic Development, iCoMET 2018 - Proceedings 2018, 2018-January, 1–4. [CrossRef]
- Sacco, C.; Baz Radwan, A.; Anderson, A.; Harik, R.; Gregory, E. Machine Learning in Composites Manufacturing: A Case Study of Automated Fiber Placement Inspection. Compos Struct 2020, 250, 112514. [CrossRef]
- Frustaci, F.; Spagnolo, F.; Perri, S.; Cocorullo, G.; Corsonello, P. Robust and High-Performance Machine Vision System for Automatic Quality Inspection in Assembly Processes. Sensors 2022, Vol. 22, Page 2839 2022, 22, 2839. [CrossRef]
- Misra, N.N.; Dixit, Y.; Al-Mallahi, A.; Bhullar, M.S.; Upadhyay, R.; Martynenko, A. IoT, Big Data, and Artificial Intelligence in Agriculture and Food Industry. IEEE Internet Things J 2022, 9, 6305–6324. [CrossRef]
- Taneja, A.; Nair, G.; Joshi, M.; Sharma, S.; Sharma, S.; Jambrak, A.R.; Roselló-Soto, E.; Barba, F.J.; Castagnini, J.M.; Leksawasdi, N.; et al. Artificial Intelligence: Implications for the Agri-Food Sector. Agronomy 2023, Vol. 13, Page 1397 2023, 13, 1397. [CrossRef]
- Expósito, I.; Cuiñas, I.; Gay-Fernández, J.A. Efficient Traceability Solutions in the Wine Production by RFID and WSN. European Conference on Antennas and Propagation 2013.
- Verzeletti, T.N.; Priamo, W.L.; Flores, S.S. Use of Information Technology for Traceability and Quality Control: Case Study in a Brazilian Winery. BIO Web Conf 2023, 68. [CrossRef]
- Chandra Prabha, S.; Kantha Lakshmi, S.; Prabha, S.C. Data Analysis and Machine Learning-Based Modeling for Real-Time Production. The Scientific Temper 2023, 14, 383–386. [CrossRef]
- Benedick, P. Lou; Robert, J.; Traon, Y. Le A Systematic Approach for Evaluating Artificial Intelligence Models in Industrial Settings. Sensors 2021, Vol. 21, Page 6195 2021, 21, 6195. [CrossRef]
- Ramanujam, S.; Caliamourthy, S.; Arunprasath, B.; Hemaprakash, N. Real Time Decision Making in Advertising Networks Using STORM. International journal of engineering research and technology 2014.
- Choubey, S.; Karmakar, G.P. Artificial Intelligence Techniques and Their Application in Oil and Gas Industry. Artif Intell Rev 2021, 54, 3665–3683. [CrossRef]
- Yan, Y.; Yang, H. Big Data Analysis and Decision Support System Based on Deep Learning. Comput Aided Des Appl 2024, 21, 62–74. [CrossRef]
- Adamashvili, N.; State, R.; Tricase, C.; Fiore, M. Blockchain-Based Wine Supply Chain for the Industry Advancement. Sustainability 2021, Vol. 13, Page 13070 2021, 13, 13070. [CrossRef]
- Prencipe, M.P.; Maleki Vishkaei, B.; De Giovanni, P. Blockchain Adoption in the Winery Industry. 2021, 43–72. [CrossRef]
- Malisic, B.; Misic, N.; Krco, S.; Martinovic, A.; Tinaj, S.; Popovic, T. Blockchain Adoption in the Wine Supply Chain: A Systematic Literature Review. Sustainability (Switzerland) 2023, 15, 14408. [CrossRef]
- Gayialis, S.P.; Kechagias, E.P.; Papadopoulos, G.A. Implementing a Wine Traceability System as an Ethereum Distributed Application. 2023. [CrossRef]
- Kang, Y.; Shi, X.; Yue, X.; Zhang, W.; Liu, S.S. Enhancing Traceability in Wine Supply Chains through Blockchain: A Stackelberg Game-Theoretical Analysis. Journal of Theoretical and Applied Electronic Commerce Research 2023, Vol. 18, Pages 2142-2162 2023, 18, 2142–2162. [CrossRef]
- Toorajipour, R.; Sohrabpour, V.; Nazarpour, A.; Oghazi, P.; Fischl, M. Artificial Intelligence in Supply Chain Management: A Systematic Literature Review. J Bus Res 2021, 122, 502–517. [CrossRef]
- Rajagopal, M.; Nayak, K.M.; Balasubramanian, K.; Abdul Karim Shaikh, I.; Adhav, S.; Gupta, M. Application of Artificial Intelligence in the Supply Chain Finance. Proceedings of 8th IEEE International Conference on Science, Technology, Engineering and Mathematics, ICONSTEM 2023 2023. [CrossRef]
- Dash, R.; McMurtrey, M.; Rebman, C.; Kar, U.K. Application of Artificial Intelligence in Automation of Supply Chain Management. Journal of Strategic Innovation and Sustainability 2019, 14, 43. [CrossRef]
- Masetti, G.; Marazzi, F.; Di Cecilia, L.; Rovati, L. IOT-Based Measurement System for Wine Industry. 2018 Workshop on Metrology for Industry 4.0 and IoT, MetroInd 4.0 and IoT 2018 - Proceedings 2018, 163–168. [CrossRef]
- Liu, Z.; Wang, S.; Zhang, Y.; Feng, Y.; Liu, J.; Zhu, H. Artificial Intelligence in Food Safety: A Decade Review and Bibliometric Analysis. Foods 2023, Vol. 12, Page 1242 2023, 12, 1242. [CrossRef]
- Qian, C.; Murphy, S.I.; Orsi, R.H.; Wiedmann, M. How Can AI Help Improve Food Safety? Annu Rev Food Sci Technol 2023, 14, 517–538. [CrossRef]
- Ling, C.; Zeng, T.; Su, Y. Research on Intelligent Supervision and Application System of Food Traceability Based on Blockchain and Artificial Intelligence. Proceedings of 2021 IEEE 2nd International Conference on Information Technology, Big Data and Artificial Intelligence, ICIBA 2021 2021, 370–375. [CrossRef]
- Rostami, S.; Zór, K.; Zhai, D.S.; Viehrig, M.; Morelli, L.; Mehdinia, A.; Smedsgaard, J.; Rindzevicius, T.; Boisen, A. High-Throughput Label-Free Detection of Ochratoxin A in Wine Using Supported Liquid Membrane Extraction and Ag-Capped Silicon Nanopillar SERS Substrates. Food Control 2020, 113, 107183. [CrossRef]
- Rasch, C.; Kumke, M.; Löhmannsröben, H.G. Sensing of Mycotoxin Producing Fungi in the Processing of Grains. Food Bioproc Tech 2010, 3, 908–916. [CrossRef]
- Pan, J.; Deng, F.; Liu, Z.; Shi, G.; Chen, J. Toehold-Mediated Cascade Catalytic Assembly for Mycotoxin Detection and Its Logic Applications. Anal Chem 2022, 94, 3693–3700. [CrossRef]
- Sun, J.; Li, W.; Zhu, X.; Jiao, S.; Chang, Y.; Wang, S.; Dai, S.; Xu, R.; Dou, M.; Li, Q.; et al. A Novel Multiplex Mycotoxin Surface-Enhanced Raman Spectroscopy Immunoassay Using Functional Gold Nanotags on a Silica Photonic Crystal Microsphere Biochip. J Agric Food Chem 2021, 69, 11494–11501. [CrossRef]
- Pizzutti, I.R.; De Kok, A.; Scholten, J.; Righi, L.W.; Cardoso, C.D.; Necchi Rohers, G.; Da Silva, R.C. Development, Optimization and Validation of a Multimethod for the Determination of 36 Mycotoxins in Wines by Liquid Chromatography–Tandem Mass Spectrometry. Talanta 2014, 129, 352–363. [CrossRef]
- Renouf, V.; Strehaiano, P.; Lonvaud-Funel, A. Yeast and Bacteria Analysis of Grape, Wine and Cellar Equipments by PCR-DGGE. OENO One 2007, 41, 51–61. [CrossRef]
- Tofalo, R.; Schirone, M.; Corsetti, A.; Suzzi, G. Detection of Brettanomyces Spp. in Red Wines Using Real-Time PCR. J Food Sci 2012, 77, M545–M549. [CrossRef]
- van Wyk, S.; Silva, F.V.M. Enumeration of Brettanomyces in Wine Using Impedance. Applied Microbiology 2021, Vol. 1, Pages 352-360 2021, 1, 352–360. [CrossRef]
- Cimaglia, F.; Tristezza, M.; Saccomanno, A.; Rampino, P.; Perrotta, C.; Capozzi, V.; Spano, G.; Chiesa, M.; Mita, G.; Grieco, F. An Innovative Oligonucleotide Microarray to Detect Spoilage Microorganisms in Wine. Food Control 2018, 87, 169–179. [CrossRef]
- Puértolas, E.; López, N.; Condón, S.; Raso, J.; Álvarez, I. Pulsed Electric Fields Inactivation of Wine Spoilage Yeast and Bacteria. Int J Food Microbiol 2009, 130, 49–55. [CrossRef]
- Dittrich, P.S.; Manz, A. Lab-on-a-Chip: Microfluidics in Drug Discovery. Nature Reviews Drug Discovery 2006 5:3 2006, 5, 210–218. [CrossRef]
- Bolinger, H.; Tran, D.; Harary, K.; Paoli, G.C.; Guron, G.K.P.; Namazi, H.; Khaksar, R. Utilizing the Microbiota and Machine Learning Algorithms To Assess Risk of Salmonella Contamination in Poultry Rinsate. J Food Prot 2021, 84, 1648–1657. [CrossRef]
- Morgan, H.H.; du Toit, M.; Setati, M.E. The Grapevine and Wine Microbiome: Insights from High-Throughput Amplicon Sequencing. Front Microbiol 2017, 8, 249460. [CrossRef]
- Longin, C.; Petitgonnet, C.; Guilloux-Benatier, M.; Rousseaux, S.; Alexandre, H. Application of Flow Cytometry to Wine Microorganisms. Food Microbiol 2017, 62, 221–231. [CrossRef]
Figure 1.
Key AI technologies
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Copyright: This open access article is published under a Creative Commons CC BY 4.0 license, which permit the free download, distribution, and reuse, provided that the author and preprint are cited in any reuse.
MDPI Initiatives
Important Links
© 2024 MDPI (Basel, Switzerland) unless otherwise stated