改进YOLOv5s的蝗虫识别系统

doi:10.13733/j.jcam.issn.2095‑5553.2024.11.029

摘要/Abstract

摘要： 为准确有效识别宁夏荒漠草原地区蝗虫种类，基于YOLOv5s网络模型提出一种复杂背景下蝗虫目标检测模型YOLOv5s-CG，在主干网络使用CoTNet保留复杂背景下的蝗虫特征信息，同时在颈部网络中融入GAM全局注意力机制提高检测模型的特征融合能力。结果表明，在对宁夏荒漠草原蝗虫进行识别时，模型YOLOv5s-CG 精确率为92.5%，平均精度均值为93.2%，相比原始模型分别提高4.8个百分点和5.3个百分点，与 Fast R-CNN、YOLOv3、YOLOv4、YOLOv5s、YOLOv6s、YOLOv7模型相比，YOLOv5s-CG网络模型对宁夏荒漠草原蝗虫具有更好的检测性能。基于YOLOv5s-CG，开发宁夏荒漠草原蝗虫识别APP系统，实现宁夏荒漠草原地区复杂背景下的蝗虫在线识别检测，为宁夏荒漠草原地区蝗虫的监管防控和综合治理提供数据支持。

关键词: 蝗虫识别, 深度学习, 自注意力机制, 全局注意力机制

Abstract: In order to accurately and effectively identify locust species in the Ningxia desert grassland region, a locust target detection model YOLOv5s-CG was proposed based on the YOLOv5s network model in a complex context, by using Contextual Transformer Networks (CoTNet) in the backbone network to retain locust feature information in the complex context, while incorporating the global attention mechanism in the neck network to improve the feature fusion capability of the detection model. The experimental results showed that the model YOLOv5s-CG had a precision of 92.5% and a mean average precision of 93.2% in locust identification in Ningxia desert grassland, which was 4.8 percentage points and 5.3 percentage points higher than the original model, respectively. Compared with the Fast R-CNN, YOLOv3, YOLOv4,YOLOv5s, YOLOv6s and YOLOv7 models, the YOLOv5s-CG network model had better detection performance for locusts in Ningxia desert grassland. Finally, based on the YOLOv5s-CG, a locust identification APP system was developed to achieve online locust identification and detection in the complex background of the Ningxia desert grassland region, providing data support for the supervision, prevention, control and comprehensive management of locusts in the Ningxia desert grassland region.

Key words: locust identification, deep learning, self?attention mechanisms, global attention mechanism

中图分类号:

S435

马宏兴, 董凯兵, 丁雨恒, 盛铁雷, 张薇. 改进YOLOv5s的蝗虫识别系统[J]. 中国农机化学报, 2024, 45(11): 189-195.

Ma Hongxing, Dong Kaibing, Ding Yuheng, Sheng Tielei, Zhang Wei. Locust identification system based on improved YOLOv5s[J]. Journal of Chinese Agricultural Mechanization, 2024, 45(11): 189-195.

参考文献

[ 1 ] 马志宁, 俞鸿千, 王颖, 等. 宁夏草原蝗虫多样性及其对环境因子的响应[J]. 中国生物防治学报, 2022, 38(6): 1459-1472.
Ma Zhining, Yu Hongqian, Wang Ying, et al. Diversity of grassland grasshoppers and responses to environmental factors in Ningxia [J]. Chinese Journal of Biological Control, 2022, 38(6): 1459 -1472.
[ 2 ] 王广君, 李贝贝, 田野. 主要蝗虫生态治理策略在我国的实践应用[J]. 植物保护学报, 2021, 48(1): 84-89.
[ 3 ] 黄文江, 董莹莹, 赵龙龙, 等. 蝗虫遥感监测预警研究现状与展望[J]. 遥感学报, 2020, 24(10): 1270-1279.
[ 4 ] Kasinathan T, Uyyala S R. Machine learning ensemble with image processing for pest identification and classification in field crops [J]. Neural Computing & Applications, 2021, 33: 7491-7504.
[ 5 ] Lima M C F, Leandor M E D D A, Valero C, et al. Automatic detection and monitoring of insect pests-A review [J]. Agriculture, 2020, 10(5): 161.
[ 6 ] Lu S H, Ye S J. Using an image segmentation and support vector machine method for identifying two locust species and instars [J]. Journal of Integrative Agriculture, 2020, 19(5): 1301-1313.
[ 7 ] Ye S J, Lu S H, Bai X S. ResNet-Locust-BN network‑based automatic identification of east asian migratory locust species and instars from RGB images [J]. Insects, 2020, 11(8): 458.
[ 8 ] Liu L M, Liu M, Meng K X, et al. Camouflaged locust segmentation based on PraNet [J]. Computers and Electronics in Agriculture, 2022, 198: 107061.
[ 9 ] Cao X, Wei Z Y, Gao Y J, et al. Recognition of common insect in field based on deep learning [J]. Journal of Physics: Conference Series, 2020, 1634: 012034.
[10] 武英洁, 房世波, Piotr Chudzik, 等. 基于 Faster R-CNN 的野外环境中蝗虫快速识别[J]. 气象与环境学报, 2020, 36(6): 137-143.
[11] 李林, 柏召, 刁磊, 等. 基于 K-SSD-F的东亚飞蝗视频检测与计数方法[J]. 农业机械学报, 2021, 52(S1): 261-267.
Li Lin, Bai Zhao, Diao Lei, et al. Video detection and counting method of east Asian migratory locusts based on K-SSD-F [J]. Transactions of the Chinese Society for Agricultural Machinery, 2021, 52(S1): 261-267.
[12] 马宏兴, 张淼, 董凯兵, 等. 基于改进YOLO v5的宁夏草原蝗虫识别模型研究[J]. 农业机械学报, 2022, 53(11): 270-279.
Ma Hongxing, Zhang Miao, Dong Kaibing, et al. Research of locust recognition in Ningxia grassland based on improved YOLOv5 [J]. Transactions of the Chinese Society for Agricultural Machinery, 2022, 53(11): 270-279.
[13] Li Y H, Yao T, Pan Y W, et al. Contextual transformer networks for visual recognition [J]. arXiv preprint arXiv: 2107. 12292, 2021.
[14] Liu Y C, Shao Z, Hoffmann N. Global attention mechanism: Retain information to enhance channel‑spatial interactions [J]. arXiv preprint arXiv: 2112. 05561, 2021.
[15] 朱猛蒙, 黄文广, 张蓉, 等. 宁夏草原蝗虫适生区和分布区划分[J]. 植物保护学报, 2021, 48(1): 237-238.
[16] Liu Z, Lin Y T, Cao Y, et al. Swin transformer: Hierarchical vision transformer using shifted windows [C]. Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021: 10012-10022.
[17] Girshick R. Fast R-CNN [J]. arXiv preprint arXiv: 1804. 02767, 2018.
[18] Redmon J, Farhadi A. YOLOv3: An incremental improvement [J]. arXiv preprint arXiv: 1504. 08083, 2015.
[19] Bochkovskiy A, Wang C Y, Liao H Y M. YOLO v4: Optimal speed and accuracy of object detection [J]. arXiv preprint arXiv: 1504. 08083, 2015.
[20] 马宏兴, 董凯兵, 王英菲, 等. 基于改进YOLOv5s的轻量化植物识别模型研究[J]. 农业机械学报, 2023, 54(8): 267-276.
Ma Hongxing, Dong Kaibing, Wang Yingfei, et al. Research on lightweight plant recognition model based on improved YOLOv5s [J]. Transactions of the Chinese Society for Agricultural Machinery, 2023, 54(8): 267-276.
[21] 罗小权, 潘善亮. 改进YOLOV3的火灾检测方法[J]. 计算机工程与应用, 2020, 56(17): 187-196.
Luo Xiaoquan, Pan Shanliang. Improved YOLOV3 fire detection method [J]. Computer Engineering and Applications, 2020, 56(17): 187-196.
[22] 高菊玲. 基于多视图特征融合的植物病害识别算法[J]. 中国农机化学报, 2020, 41(12): 147-152.
Gao Juling. Recognition algorithm of plant disease based on multi‑view feature fusion [J]. Journal of Chinese Agricultural Mechanization, 2020, 41(12): 147-152.

[1]	韩鹏飞, 宋其江, 贾梦实. 基于改进轻量化 EfficientNet-V2模型的小麦种子分类[J]. 中国农机化学报, 2024, 45(9): 111-117.
[2]	李志臣, 凌秀军, 李鸿秋. 基于声音深度学习的发动机失火故障诊断[J]. 中国农机化学报, 2024, 45(9): 134-140.
[3]	桑一男 , 徐增莱 , 汪琼 , 葛海涛 , 王殿广 . 一种轮式采花机器人设计——以黄蜀葵为例[J]. 中国农机化学报, 2024, 45(9): 202-208.
[4]	李邦国, 王辉, 宋杨, 任志伟, 刘跃华, 徐乐程. 基于无人驾驶小麦收割机立体视觉感知系统[J]. 中国农机化学报, 2024, 45(9): 244-249.
[5]	许贞辉, 李晓娟. 果园树干检测与导航线拟合算法研究[J]. 中国农机化学报, 2024, 45(8): 217-222.
[6]	胡国玉, 刘广, 周星光, 董娅兰, 周建平, . 基于Swin-TDL算法的果园环境下葡萄病害检测方法[J]. 中国农机化学报, 2024, 45(8): 234-239.
[7]	吴利刚, 吕媛媛, 周倩, 陈乐, 张梁, 史建华. 基于改进YOLOv5的轻量级黄花成熟检测方法[J]. 中国农机化学报, 2024, 45(7): 235-242.
[8]	张继成, 黄向党. 基于改进YOLOv3的玉米病害识别方法[J]. 中国农机化学报, 2024, 45(7): 269-275.
[9]	雷浩, 苑迎春, 何振学. 基于多尺度卷积和注意力机制的枣品种识别[J]. 中国农机化学报, 2024, 45(6): 135-141.
[10]	崔海朋, 秦朝旭, 马志宇. 基于深度学习的鱼类特征点检测与体征识别方法[J]. 中国农机化学报, 2024, 45(6): 201-207.
[11]	黄尧, 何敬, 付饶, 刘刚, , 林远杨. YOLOv5s-CBAM算法在福寿螺虫卵识别中的应用分析[J]. 中国农机化学报, 2024, 45(6): 223-228.
[12]	陶旭, , 余富强, , 蔡金金, 么炜, , 刘博, . 面向激光雷达点云数据的多结构树种识别[J]. 中国农机化学报, 2024, 45(5): 168-175.
[13]	漆海霞, , , 李承杰, 黄桂珍. 基于轻量化YOLOv4的死淘鸡目标检测算法[J]. 中国农机化学报, 2024, 45(5): 195-201.
[14]	罗鑫, 李加强, 何超. 基于改进YOLOv4的澳洲坚果视觉监测方法[J]. 中国农机化学报, 2024, 45(5): 217-222.
[15]	廖舒怀, , 王凯, 宋健, 解福祥, 王名声, , 龚中良. 基于移动机器人的树木识别与冠层信息测量[J]. 中国农机化学报, 2024, 45(5): 232-238.