基于视觉加强注意力模型的植物病虫害检测

doi:10.12302/j.issn.1000-2006.202210022

国家林草科技领军期刊
中国精品科技期刊
中国高校百佳科技期刊
江苏省新闻出版政府奖期刊奖
RCCSE林学权威期刊（A+）
CSCD核心期刊
Scopus数据库收录期刊
中文核心期刊
SCD核心期刊

作者加群：102861116

微信公众号：南京林业大学学报

高级检索

南京林业大学学报（自然科学版） ›› 2023, Vol. 47 ›› Issue (3): 11-18.doi: 10.12302/j.issn.1000-2006.202210022

所属专题：第三届中国林草计算机应用大会论文精选

• 专题报道：第三届中国林草计算机应用大会论文精选（执行主编李凤日） • 上一篇下一篇

基于视觉加强注意力模型的植物病虫害检测

杨堃(), 范习健(), 薄维昊, 刘婕, 王俊玲

南京林业大学信息科学技术学院,江苏南京 210037

收稿日期:2022-10-16 修回日期:2023-02-20 出版日期:2023-05-30 发布日期:2023-05-25
通讯作者: 范习健
基金资助:
国家自然科学基金项目(61902187);辽宁省科技厅和机器人国家重点实验室联合基金项目(2020-KF-22-04);南京市留学归国人员科技创新择优资助项目

Plant disease and pest detection based on visiual attention enhancement

YANG Kun(), FAN Xijian(), BO Weihao, LIU Jie, WANG Junling

School of Information Science and Technology, Nanjing Forestry University, Nanjing 210037, China

Received:2022-10-16 Revised:2023-02-20 Online:2023-05-30 Published:2023-05-25
Contact: FAN Xijian

摘要/Abstract

摘要：

【目的】植物病虫害准确检测是病虫害精准化防治的关键,笔者构建准确高效的植物病虫害监测模型,为病虫害的早期诊断与预警提供重要依据。【方法】针对现有植物病虫害检测模型泛化能力弱、小目标漏检率高等问题,提出一种基于视觉加强注意力改进的植物病虫害检测模型--YOLOv 5-VE(vision enhancement)。为方便检测实验样本中的小目标采用Mosaic 9数据增强方法;设计出基于视觉注意力的特征加强模块CBAM(convolutional block attention module);为确定不同目标重叠在一起和被遮挡的定位损失引入边界框定位损失函数DIoU。【结果】YOLOv 5-VE模型在实验数据集上的识别精度和检测平均准确率达到65.87%和73.49%,比原模型提高了1.07%和8.25%,在型号为1 080 Ti的GPU上检测速度可达35帧/s。【结论】该方法可以在背景复杂的野外场景快速有效地检测和识别种类多样的病害和虫害,可以提高检测的鲁棒性能,提升模型对病虫害目标的特征提取能力,降低野外复杂场景对检测带来的干扰,表现出良好的应用潜力,可广泛运用于大规模的植物病虫害检测。

关键词: 植物病虫害检测, 目标检测, 注意力机制, 数据增强

Abstract:

【Objective】 Accurate detection is the key to precise control of plant diseases and pests. Building an accurate and efficient monitoring model of plant diseases and pests provides an important basis for the early diagnosis and warning of plant diseases and pests.【Method】In view of the weak generalization ability of the existing plant diseases and pests detection models and the high rate of missed detection of small targets, a plant disease and pest detection model based on visual attention enhancement improvement YOLOv 5-VE (vision enhancement) was proposed. The Mosaic 9 data enhancement method was used to facilitate the detection of small targets in experimental samples; a feature enhancement module based on a visual attention convolutional block attention module (CBAM) was designed; to determine the location loss of overlapping and occluded targets, the DIoU bounding box location loss function was introduced.【Result】The recognition and average detection accuracies of the YOLOv 5-VE model on the experimental dataset reached 65.87% and 73.49%, respectively, which were 1.07% and 8.25% higher, respectively, than those of the original model. The detection speed on the GPU with the model 1 080 Ti reached 35 fps. 【Conclusion】This method can quickly and effectively detect and identify a variety of diseases and pests in field scenes with complex backgrounds, improve robustness of detection, improve the feature extraction ability of the model for pests and diseases, reduce the interference of complex field scenes in detection, and shows good application potential. It can be widely used for the large-scale detection of plant diseases and pests.

Key words: detection of plant diseases and pests, object detection, attention mechanism, data augmentation

中图分类号:

TP3
S763

杨堃,范习健,薄维昊,等. 基于视觉加强注意力模型的植物病虫害检测[J]. 南京林业大学学报（自然科学版）, 2023, 47(3): 11-18.

YANG Kun, FAN Xijian, BO Weihao, LIU Jie, WANG Junling. Plant disease and pest detection based on visiual attention enhancement[J].Journal of Nanjing Forestry University (Natural Science Edition), 2023, 47(3): 11-18.DOI: 10.12302/j.issn.1000-2006.202210022.

图/表 10

图1

图2

图3

图4

图5

表1

图6

图7

图8

图9

参考文献 23

[1]	LIANG Q K, XIANG S, HU Y C, et al. PD2SE-Net: computer-assisted plant disease diagnosis and severity estimation network[J]. Comput Electron Agric, 2019, 157(C):518-529.DOI: 10.1016/j.compag.2019.01.034.
[2]	MARTINELLI F, SCALENGHE R, DAVINO S, et al. Advanced methods of plant disease detection: a review[J]. Comput Electron Agric, 2018, 145:311-318.DOI: 10.1016/j.compag.2018.01.009.
[3]	FERENTINOS K P. Deep learning models for plant disease detection and diagnosis[J]. Comput Electron Agric, 2018, 145:311-318.DOI: 10.1016/j.compag.2018.01.009.
[4]	周宏威, 姜钦啸, 刘洋, 等. 树木病虫害缺陷精准检测方法的综述和展望[J]. 昆虫学报, 2020, 63(9): 1145-1152.
	ZHOU H W, JIANG Q X, LIU Y, et al. Review and prospects of accurate detection methods for wood defects caused by diseases and pests[J]. Acta Entomol Sin, 2020, 63(9):1145-1152.DOI: 10.16380/j.kcxb.2020.09.012.
[5]	肖志云, 刘洪. 小波域马铃薯典型虫害图像特征选择与识别[J]. 农业机械学报, 2017, 48(9): 24-31.
	XIAO Z Y, LIU H. Features selection and recognition of potato typical insect pest images in wavelet domain[J]. Trans Chin Soc Agric Mach, 2017, 48(9):24-31.DOI: 10.6041/j.issn.1000-1298.2017.09.003.
[6]	SCHOR N, BECHAR A, IGNAT T, et al. Robotic disease detection in greenhouses: combined detection of powdery mildew and tomato spotted wilt virus[J]. IEEE Robotics Autom Lett, 2016, 1(1):354-360.DOI: 10.1109/LRA.2016.2518214.
[7]	党满意, 孟庆魁, 谷芳, 等. 基于机器视觉的马铃薯晚疫病快速识别[J]. 农业工程学报, 2020, 36(2): 193-200.
	DANG M Y, MENG Q K, GU F, et al. Rapid recognition of potato late blight based on machine vision[J]. Trans Chin Soc Agric Eng, 2020, 36(2):193-200.DOI: 10.11975/j.issn.1002-6819.2020.02.023.
[8]	HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[G]// IEEE Comference on Computer Vision and Pattern Recognition, 2016: 770-778.DOI: 10.1109/CVPR.2016.90.
[9]	SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition[EB/OL]. [2022-02-01].http:arxiv.org/abs/1409.1556. DOI: 10.48550/arXiv.1409.1556.
[10]	IOANNIDOU A, CHATZILARI E, NIKOLOPOULOS S, et al. Deep learning advances in computer vision with 3D data: a survey[J]. ACM Comput Surv, 2018, 50(2):1-38.DOI: 10.1145/3042064.
[11]	FAN X J, LUO P, MU Y E, et al. Leaf image based plant disease identification using transfer learning and feature fusion[J]. Comput Electron Agric, 2022, 196:106892.DOI: 10.1016/j.compag.2022.106892.
[12]	MOHANTY S P, HUGHES D P, SALATHÉ M. Using deep learning for image-based plant disease detection[J]. Front Plant Sci, 2016, 7:1419.DOI: 10.3389/fpls.2016.01419.
[13]	李鑫然, 李书琴, 刘斌. 基于改进Faster R_CNN的苹果叶片病害检测模型[J]. 计算机工程, 2021, 47(11): 298-304.
	LI X R, LI S Q, LIU B. Apple leaf diseases detection model based on improved faster R_CNN[J]. Comput Eng, 2021, 47(11):298-304.DOI: 10.19678/j.issn.1000-3428.0059290.
[14]	刘延鑫, 王俊峰, 杜传印, 等. 基于 YOLOv3 的多类烟草叶部病害检测研究[J]. 中国烟草科学, 2022, 43(2): 94-100.
	LIU Y X, WANG J F, DU C Y, et al. Study on detection of various tobacco leaf diseases based on YOLOv 3[J]. China Tobacco Science, 2022, 43(2):94-100. DOI:10.13496/j.issn.1007-5119.2022.02.015.
[15]	WANG J, YU L Y, YANG J, et al. DBA_SSD: a novel end-to-end object detection algorithm applied to plant disease detection[J]. Information, 2021, 12(11):474.DOI: 10.3390/info12110474.
[16]	WANG H Q, SHANG S Q, WANG D W, et al. Plant disease detection and classification method based on the optimized lightweight YOLOv5 model[J]. Agriculture, 2022, 12(7):931.DOI: 10.3390/agriculture12070931.
[17]	SINGH D, JAIN N, JAIN P, et al. PlantDoc: a dataset for visual plant disease detection[R]// Association for Computing Machinery. Proceedings of the 7th ACM IKDD CoDS and 25th COMAD. New York, USA, IEEE, 2019.DOI: 10.1145/3371158.3371196.
[18]	WU X P, ZHAN C, LAI Y K, et al. Ip102: a large-scale benchmark dataset for insect pest recognition[R]. 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Long Beach,CA, USA: IEEE 2019.DOI: 10.1109/CVPR.2019.00899.
[19]	ZHANG Q, LIU Y Q, GONG C Y, et al. Applications of deep learning for dense scenes analysis in agriculture: a review[J]. Sensors (Basel), 2020, 20(5):1520.DOI: 10.3390/s20051520.
[20]	WOO S, PARK J, LEE J Y, et al. Cbam: convolutional block attention module[M]// Computer Vision-ECCV 2018.Cham: Springer International Publishing, 2018:3-19.DOI: 10.1007/978-3-030-01234-2_1.
[21]	REDMON J, FARHADI A. Yolov3: an incremental improvement[EB/OL]. [2022-02-01].http:arxiv.org/abs/1804.02767. 2018. DOI: 10.48550/arXiv.180402767.
[22]	BOCHKOVSKIY A, WANG C Y, LIAO H Y M. Yolov4: optimal speed and accuracy of object detection[EB/OL]. [2022-02-01]http:arxiv.org/abs/2004.10934. DOI: 10.48550/arXiv.2004.10934.
[23]	GE Z, LIU S T, WANG F, et al. Yolox: exceeding yolo series in 2021[EB/OL]. [2022-02-01].http:arxiv.org/abs/2107.08430..DOI: 10.48550/arXiv.2107.08430.

类型 type	模型 model	参数大小/ MB parameter size	输入大小/pixel input size	查准率/% precision	查全率/% recall	阈值大于0.5 的平均准确度/% m_{AP_0.5}	F₁/%	单张检测耗时/ms detection time
原模型 original model	YOLOv 3	237.73	640	52.58	59.12	62.17	55.66	22.75
	YOLOv 4	136.43	640	21.29	26.42	39.45	23.58	14.91
	YOLOv 5	174.77	640	64.80	59.39	65.24	61.98	36.90
	YOLO X	378.82	640	33.61	42.15	42.78	37.40	7.52
	YOLOv 5-VE	194.77	640	65.87	68.39	73.49	67.11	35.46
改进模型 improved model	YOLOv 5	174.77	640	64.80	59.39	65.24	61.98	36.90
	+CBAM	194.77	640	62.72	68.90	69.61	65.66	34.36
	+CBAM+Mosaic 9	194.77	640	66.27	67.30	70.69	66.78	35.34
	YOLOv 5-VE	194.77	640	65.87	68.39	73.49	67.11	35.46

基于视觉加强注意力模型的植物病虫害检测

Plant disease and pest detection based on visiual attention enhancement

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献 23

相关文章 1

编辑推荐

Metrics

本文评价

作者

读者

审稿