In response to the problems of subtle and diverse disease symptoms in tomato disease detection and the limited computing power of edge devices,which make it difficult to balance detection accuracy and real-time performance,this paper proposes a collaborative tomato leaf disease detection framework MIG-YOLO11 based on YOLO11,integrating a mixed local channel attention（MLCA）module,an iterative attention feature fusion（iAFF）module,and a lightweight convolution（GhostConv）module.In the MIG-YOLO11 architecture,the MLCA module can enhance the model＇s ability to capture global features of large targets and semantic expression in complex scenarios;the iAFF module can improve the iterative fusion ability of shallow high-resolution features and deep semantic features,thereby enhancing the detection accuracy of small targets;the GhostConv module can reduce the computational load and model parameters while maintaining feature expression capabilities,enabling the network to maintain high accuracy while having good real-time inference performance.Experimental results show that the improved MIG-YOLO11 performs excellently in tomato disease detection tasks,with an average detection accuracy of 94.5％,an increase of approximately 2.1 percentage points compared to the baseline model;at the same time,the model parameters are reduced by about 8.3% ,and the GFLOPs are decreased by approximately 1.6% .MIG-YOLO11 effectively breaks through the performance bottleneck of lightweight model deployment while balancing detection accuracy and computational efficiency,demonstrating broad application potential in resource-constrained scenarios.
Key words
light weight /
object detection /
attention mechanism /
disease recognition /
edge device computing
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
References
[1] 王沛,司智霞.基于高品质需求谈我国番茄产业现状与发辰J].中国蔬菜,2025(11):1-7.
[2] 樊佩知,王志敏.番茄抗病基因育种研究进展[J].中国农业文摘-农业工程,2024,36(2):3-9
[3] 段育辰. 基于智能优化深度学习的农作物病虫害识别研究[D].南京:南京信息工程大学,2024
[4] 吴世琛,毛玉明,胡慧中,等.基于改进YOLO11n的葡萄果叶病害检测方法[J].农业工程学报,2025,41(14):140-147
[5] Girshick R.Fast R-CNN[C]//2015 IEEE International Conference on Computer Vision (ICCV) IEEE,2016:1440-1448
[6] Liu W,Anguelov D,Erhan D,ct al.SSD:Single Shot MultiBox Detector[C]//Computer Vision-ECCV 2016.Cham:Springer,2016:21-37.
[7] Liu Y,He G,Wang Z,et alNRT-YOLO:Improved YOLOv5 Based on Nested Residual Transformer for Tiny Remote Sensing Object Detection[J].Sensors,2022.22(13):4953
[8] REIS D,KUPEC J,HONG J,et al. Real-time flying object detection with YOLOv8. arXiv preprint arXiv2305.09972,2023.
[9] Chen H,Chen K,Ding G,et al.YOLOv10:Real-Time End-to-End Object Detection[C]//Advances in Ncural Information Processing Systems 37.Ncural Information Processing Systems Foundation,Ine.(NeuIPS)2024:107984-108011
[10] 彭林聪,王克瑞,周浩,等.基于改进SSD的目标检测算法[J]. 激光杂志,2024,45(11):71-76
[11] 王伟辉,信泽阳,车清论,等.基于改进Faster RCNN模型的冬枣缺陷检测方法[J].农业工程学报,2024,40(22):283-289.
[12] Sujatha R,Krishnan S,Chatterjee J M,et al.Advancing Plant Leaf Disease Detection Integrating Machine Learning and Deep Learning[J].Scientific Reports,2025,15:11552.
[13] 李江川,韩彦岭,董传胜,等.YOLO-U:基于结构重参数化和双重注意力机制的水下目标检测算法[J].上海海洋大学学报,2025,34(3):696-706.
[14] 李茂,肖洋轶,宗望远.基于改进YOLO 11的多尺度板栗果实识别方法[J].农业机械学报,2025,56(5):443-454.
[15] Khanam R,Hussain M.YOLOv11:An Overview of the Key Architectural Enhancements[EB/OL].2024: arXiv:2410.17725.https://arxiv.org/abs/2410.17725.
[16] Wan D,Lu R,Shen S,et al.Mixed Local Channel Attention for Object Detection[J].Engineering Applications of Artificial Intelligence,2023,123:106442.
[17] Dai Y,Gieseke F,Oehmcke S,et al.Attentional Feature Fusion[C]//2021 IEEE Winter Conference on Applications of Computer Vision(WACV).IEEE,2021:3559-3568.
[18] Han K,Wang Y,Tian Q,et al.GhostNet:More Features from Cheap Operations[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).IEEE,2020:1577-1586.
[19] 孙佩月,黄娟,顾寄南,等.基于YOLO11n的轻量化芯片表面缺陷检测方法[J].半导体技术,2025,50(7):707-713.
[20] Hussain M.YOLOv5,YOLOv8 and YOLOv10:The Go-to Detectors for Real-Time Vision[EB/OL].2024: arXiv:2407.02988.https://arxiv.org/abs/2407.02988
