Version 1
: Received: 19 August 2024 / Approved: 19 August 2024 / Online: 27 August 2024 (11:41:45 CEST)
How to cite:
Yan, S.; Lu, Y.; Li, T.; Rong, Y.; Huang, Z.; Hu, Z. Research on Evolutionary Algorithm for Soybean Seedling and Weed Identification Driven by Computational Complexity. Preprints2024, 2024081891. https://doi.org/10.20944/preprints202408.1891.v1
Yan, S.; Lu, Y.; Li, T.; Rong, Y.; Huang, Z.; Hu, Z. Research on Evolutionary Algorithm for Soybean Seedling and Weed Identification Driven by Computational Complexity. Preprints 2024, 2024081891. https://doi.org/10.20944/preprints202408.1891.v1
Yan, S.; Lu, Y.; Li, T.; Rong, Y.; Huang, Z.; Hu, Z. Research on Evolutionary Algorithm for Soybean Seedling and Weed Identification Driven by Computational Complexity. Preprints2024, 2024081891. https://doi.org/10.20944/preprints202408.1891.v1
APA Style
Yan, S., Lu, Y., Li, T., Rong, Y., Huang, Z., & Hu, Z. (2024). Research on Evolutionary Algorithm for Soybean Seedling and Weed Identification Driven by Computational Complexity. Preprints. https://doi.org/10.20944/preprints202408.1891.v1
Chicago/Turabian Style
Yan, S., Zhen Huang and Zhigang Hu. 2024 "Research on Evolutionary Algorithm for Soybean Seedling and Weed Identification Driven by Computational Complexity" Preprints. https://doi.org/10.20944/preprints202408.1891.v1
Abstract
To achieve real-time identification of soybean seedlings and weeds in soybean seedling fields on agricultural intelligent terminal devices, this article constructs a real-time evolutionary identification system ESWIM driven by computational complexity under few-shot training. Firstly, based on the identification task, high-quality image samples are collected and annotated. By analyzing the terrain, the weather during the growth period, image acquisition methods, and the status of equipment fieldwork, we obtained a suitable enhancement algorithm. This algorithm not only effectively simulates the diversity of actual images, but also solves the problem of significant imbalance in the number of targets. Then, using the computational complexity driven approach and YOLOv5n as the foundation model, a soybean seedling and weed recognition model SWIM is constructed. The model's computational space and time complexity are reduced by 45.19% and 42.86%, respectively, while mAP (0.5) increased by 0.1%. Finally, by utilizing human-computer interaction to enable real-time and effective evolution of the dataset, an evolutionary strategy is designed to rapidly iterate and update model parameters, resulting in the development of the evolutionary system ESWIM. Laboratory experiments have shown that the ESWIM system can quickly adjust the recognition system during device operation to achieve target identification tasks in different environments.
Biology and Life Sciences, Agricultural Science and Agronomy
Copyright:
This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
