基于改进Cascade R-CNN算法的黄粘板图像上十字花科蔬菜害虫田间智能监测系统

doi:10.1016/j.jia.2024.06.017

摘要/Abstract

摘要：

十字花科蔬菜是重要的食用蔬菜，但在生长过程中极易受到害虫的影响。为了实现虫情测报和科学防治，需要对这些害虫进行实时准确的监测。目前，悬挂黄粘板是黄色趋向性蔬菜害虫监测和诱捕的常用方法。为了实现对粘虫板上蔬菜害虫的实时、低成本、智能化监测，我们建立了一个由智能摄像头、Web平台和部署在服务器上的害虫检测算法组成的智能监测系统。工作人员在系统平台上设置好摄像头的拍摄预置点和拍摄时间后，布置在田间的摄像头每天定点、定时自动采集多张黄粘板的图像。黄粘板上诱捕的害虫包括小菜蛾（Plutella xylostella)、黄曲条跳甲（Phyllotreta striolata）、蝇类（Muscidae）三种，它们的虫体小且易破损，增加了模型检测的难度。针对害虫体型小、易破损导致的识别效果差的问题，我们提出了基于改进Cascade R-CNN模型的蔬菜害虫智能检测算法来识别这三种目标害虫。该算法引入了重叠滑动窗口方法，使用改进的Res2Net网络作为骨干网络，同时采用递归特征金字塔网络作为特征融合网络。测试结果表明，该算法对黄粘板图像上的三种目标害虫取得了良好的检测效果，精确率分别为96.5%、92.2%和75.0%，召回率分别为96.6%、93.1%和74.7%，F₁为0.880。与其他算法相比，我们的算法在检测小目标害虫的能力上具有明显优势。为了准确获取图像上每天新增的害虫，我们提出了一种基于双阶段匹配的蔬菜害虫匹配计数算法，测试结果显示，该算法取得了与手工统计结果高度一致的害虫增长情况，平均误差仅2.2%。蔬菜害虫智能监测系统实现了蔬菜害虫监测的精准化、可视化和智能化，为农户防控害虫提供依据，具有重要意义。

Abstract:

Cruciferous vegetables are important edible vegetable crops. However, they are susceptible to various pests during their growth process, which requires real-time and accurate monitoring of these pests for pest forecasting and scientific control. Hanging yellow sticky boards is a common way to monitor and trap those pests which are attracted to the yellow color. To achieve real-time, low-cost, intelligent monitoring of these vegetable pests on the boards, we established an intelligent monitoring system consisting of a smart camera, a web platform and a pest detection algorithm deployed on a server. After the operator sets the monitoring preset points and shooting time of the camera on the system platform, the camera in the field can automatically collect images of multiple yellow sticky boards at fixed places and times every day. The pests trapped on the yellow sticky boards in vegetable fields, Plutella xylostella, Phyllotreta striolata and flies, are very small and susceptible to deterioration and breakage, which increases the difficulty of model detection. To solve the problem of poor recognition due to the small size and breaking of the pest bodies, we propose an intelligent pest detection algorithm based on an improved Cascade R-CNN model for three important cruciferous crop pests. The algorithm uses an overlapping sliding window method, an improved Res2Net network as the backbone network, and a recursive feature pyramid network as the neck network. The results of field tests show that the algorithm achieves good detection results for the three target pests on the yellow sticky board images, with precision levels of 96.5, 92.2 and 75.0%, and recall levels of 96.6, 93.1 and 74.7%, respectively, and an F₁ value of 0.880. Compared with other algorithms, our algorithm has a significant advantage in its ability to detect small target pests. To accurately obtain the data for the newly added pests each day, a two-stage pest matching algorithm was proposed. The algorithm performed well and achieved results that were highly consistent with manual counting, with a mean error of only 2.2%. This intelligent monitoring system realizes precision, good visualization, and intelligent vegetable pest monitoring, which is of great significance as it provides an effective pest prevention and control option for farmers.

Key words: vegetable pests , yellow sticky boards , intelligent monitoring system , deep learning , pest detection

Yufan Gao, Fei Yin, Chen Hong, Xiangfu Chen, Hang Deng, Yongjian Liu, Zhenyu Li, Qing Yao. 基于改进Cascade R-CNN算法的黄粘板图像上十字花科蔬菜害虫田间智能监测系统[J]. Journal of Integrative Agriculture, 2025, 24(1): 220-234.

Yufan Gao, Fei Yin, Chen Hong, Xiangfu Chen, Hang Deng, Yongjian Liu, Zhenyu Li, Qing Yao. Intelligent field monitoring system for cruciferous vegetable pests using yellow sticky board images and an improved Cascade R-CNN[J]. Journal of Integrative Agriculture, 2025, 24(1): 220-234.

参考文献

Arantza B, Laura G, Artzai P, Navarra R, Klukas C. 2022. Insect counting through deep learning-based density maps estimation. Computers and Electronics in Agriculture, 3, 197.

Böhme M H, Dimitrov-Skatov A G. 2018. Investigations to biological and organic treatments against pests in vegetable cultivation. IOP Conference Series (Earth and Environmental Science), 215, 012012.

Cai L, Wang S T, Liu J H, Zhu Y Y. 2020. A review of data labeling studies. Journal of Software, 31, 302–320. (in Chinese)

Cai Z, Vasconcelos N. 2019. Cascade R-CNN: High quality object detection and instance segmentation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 43, 1483–1498.

Cen G J, Fang S M, Yang D G, Zhang S X, Zhang L K. 2021. Comprehensive experimental design for image recognition of the yellow-striped flea beetle. Experimental Technology and Management, 38, 198–203. (in Chinese)

Dai F, Wang F C, Yang D Z, Lin S M, Chen X, Lan Y B, Deng X L. 2022. Detection method of citrus psyllids with field high-definition camera based on improved cascade region-based convolution neural networks. Frontiers in Plant Science, 12, 816272.

Ding W G, Graham T. 2016. Automatic moth detection from trap images for pest management. Computers and Electronics in Agriculture, 123, 17–28.

Ebrahimi M A, Khoshtaghaza M H, Minaei S, Jamshidi B. 2017. Vision-based pest detection based on SVM classification method. Computers and Electronics in Agriculture, 137, 52–58.

Gao S H, Cheng M M, Zhao K, Zhang X Y, Yang M S, Torr P. 2021. Res2Net: A new multi-scale backbone architecture. IEEE Transactions on Pattern Analysis and Machine Intelligence, 43, 652–662.

Girshick R. 2015. Fast R-CNN. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.Institute of Electrical and Electronic Engineers Computer Society, USA. pp. 1440–1448.

Girshick R, Donahue Y, Darrell T. 2014. Rich feature hierarchies for accurate object detection and semantic segmentation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Institute of Electrical and Electronic Engineers Computer Society, USA. pp. 580–587.

Guo Q W, Wang C X, Xiao D Q, Huang Q. 2021. Building saliency map based attention driven deep network to detect vegetable pests of sticky trap images. Transactions of the Chinese Society of Agricultural Engineering, 37, 211–219. (in Chinese)

He K M, Gkioxari G, Piotr D. 2017. Mask R-CNN. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Institute of Electrical and Electronic Engineers Computer Society, USA. pp. 2961–2969.

He K M, Zhang X Y, Ren S Q, Sun J. 2016. Deep residual learning for image recognition. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Institute of Electrical and Electronic Engineers Computer Society, USA. pp. 770–778.

Hong S, Nam I, Kim S, Kim E, Lee C, Ahn S, Park I, Kim G. 2021. Automatic pest counting from pheromone trap images using deep learning object detectors for matsucoccus thunbergianae monitoring. Insects, 12, 342.

Huang J, Kwong M, Leung W. 2021. Species of pests and its integrated control technologies in Brassicaceae vegetable. Vegetables, 9, 58–61. (in Chinese)

Lei Z R. 2016. Biological control of diseases and insect pests is valid method to ensure vegetable safe producing. Scientia Agricultura Sinica, 49, 2932–2934. (in Chinese)

Li W Y, Wang D J, Li M, Gao Y L, Wu J W, Yang X T. 2021. Field detection of tiny pests from sticky trap images using deep learning in agricultural greenhouse. Computers and Electronics in Agriculture, 183, 106048.

Liu Z, Lin Y T, Cao Y. 2021. Swin transformer hierarchical vision transformer using shifted windows. Computer Vision and Pattern Recognition, 9, 10012–10022.

Lu X, Li B Y, Yue Y X, Li Q Q, Yan J J. 2019. Grid R-CNN plus: Faster and better. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Institute of Electrical and Electronic Engineers Computer Society, USA. pp. 7363–7372.

Rong M X, Wang Z Z, Ban B, Guo X L. 2022. Pest recognition and counting of yellow plate in field based on improved mask R-CNN. Discrete Dynamics in Nature, 22, 1–9.

Rustia D J A, Chao J J, Chiu L Y, Wu Y F, Chung J Y, Hsu J C, Lin T T. 2020. Automatic greenhouse insect pest detection and recognition based on a cascaded deep learning classification method. Journal of Applied Entomology, 145, 206–222.

Sun Y R, Cheng H, Cheng Q, Zhou H Y, Li M H, Fan Y, Shan G L, Damerow L, Lammers P S, Jones S B. 2017. A smart-vision algorithm for counting whiteflies and thrips on sticky boards using two-dimensional Fourier transform spectrum. Biosystems Engineering, 153, 82–88.

Wang X J, Mi L C. 2022. The trapping effect of yellow-blue sticky boards on insects in the main agricultural areas of Qinghai Province. China Agricultural Technology Extension, 38, 92–95. (in Chinese)

Xiao D Q, Huang Y G, Zhang Y Q, Liu Y F, Lin S, Yang W T. 2021. Field yellow sticky trap pest detection algorithm based on improved faster R-CNN. Journal of Agricultural Machinery, 52, 242–251.

Xie S N, Girshick R, Dollar P, Tu Z W, He K M. 2017. Aggregated residual transformations for deep neural networks. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Institute of Electrical and Electronic Engineers Computer Society, USA. pp. 5987–5995.

Yao Q, Wu S Z, Kuai N Y, Yang B J, Tang J, Feng J, Zhu X. 2021. Automatic detection of rice planthoppers through light-trap insect images using improved CornerNet. Transactions of the Chinese Society of Agricultural Engineering, 37, 183–189. (in Chinese)

Zhang H K, Chang H, Ma B P, Wang N Y, Chen X L. 2020. Dynamic R-CNN: Towards high quality object detection via dynamic training. Computer Vision, 6, 260–275.

Zhang J, Zhang S. 2024. Lightweight crop pest detection method based on attention mechanism and YOLOv5s. Journal of Science of Teachers′ College and University, 44, 36–42, 50. (in Chinese)

Zhang L K, Zhang C, Cen C P, Gao Y. 2020. Identifying and counting phyllotreta striolata fabriciuson based on yellow sticky trap. Journal of Jiangsu University (Natural Science Edition), 41, 339–345. (in Chinese)

Zhong Y H, Gao J Y, Lei Q L, Zhou Y. 2018. A vision-based counting and recognition system for flying insects in intelligent agriculture. Sensors, 18, 1489.

Zhu L X, Wu Y W, Zheng L H, Gao P, Zhang G C, Li F, Wang W, Xv W. 2024. Research progress on prevention and control of Phyllotreta striolata. Heilongjiang Agricultural Science, 2, 105–112. (in Chinese)