基于深度学习的三七病害识别方法适应性研究

doi:10.14088/j.cnki.issn0439-8114.2024.08.010

摘要/Abstract

摘要： 基于深度学习技术,构建SSD、YOLOv5以及同基础网络的Fast RCNN模型（vgg16-Fast R-CNN,darknet53-Fast R-CNN）,对不同条件的三七（Panax notoginseng）圆斑病、灰霉病、白粉病和病毒病进行检测。结果表明,YOLOv5的m权重模型在YOLO各权重模型中表现最佳,准确率为88.62%,召回率为89.59%,F1精度为89.10%,平均精度为83.55%,单幅图像检测时间仅为0.031 s。对比两阶段模型中表现较优的vgg16-Fast R-CNN,准确率、召回率、F1精度、平均精度仅分别降低了1.69个百分点、3.92个百分点、2.78个百分点、3.47个百分点,但单幅图像的检测速度提高了451.4%;对比SSD模型,YOLOv5m的准确率、召回率、F1精度、平均精度分别提高了1.06个百分点、1.32个百分点、1.19个百分点、0.61个百分点,单幅图像的检测速度提高了83.52%。此外通过置信度与鲁棒性试验分析可得,YOLOv5m对于小区域病害检测能力以及复杂环境下的抗干扰能力更强,且更利于在嵌入式设备中部署,符合实时检测三七病害的要求。

关键词: 三七（Panax notoginseng）, 病害检测, YOLOv5, 深度学习

Abstract: Based on deep learning technology, SSD, YOLOv5 and Fast RCNN models with the same basic network （vgg16-Fast R-CNN, darknet53 -Fast R-CNN） were built to detect round spot, gray mold, powdery mildew and viral diseases of Panax notoginseng under different conditions. The results showed that the m-weight model of YOLOv5 performed the best among all weight models of YOLO, with accuracy rate of 88.62%, recall rate of 89.59%, F1 precision of 89.10%, and average precision of 83.55%. The detection time of a single image was only 0.031 s. Compared with vgg16-Fast R-CNN, which performed better in the two-stage model, the accuracy rate, recall rate, F1 precision, and average precision were only reduced by 1.69 percentage points, 3.92 percentage points, 2.78 percentage points, and 3.47 percentage points respectively, but the detection speed of a single image was increased by 451.4%. Compared with the SSD model, the accuracy rate, recall rate, F1 precision, and average precision of YOLOv5m were improved by 1.06 percentage points, 1.32 percentage points, 1.19 percentage points, and 0.61 percentage points respectively, and the detection speed of a single image was improved by 83.52%. In addition, through the analysis of the confidence and robustness test, it could be seen that YOLOv5m had better disease detection ability in small areas and stronger anti-interference ability in complex environment, and was more conducive to deployment in embedded devices, which met the requirements of real-time detection of Panax notoginseng disease.

Key words: Panax notoginseng, disease detection, YOLOv5, deep learning

中图分类号:

何恒, 周平. 基于深度学习的三七病害识别方法适应性研究[J]. 湖北农业科学, 2024, 63(8): 54-60.

HE Heng, ZHOU Ping. Research on adaptability of Panax notoginseng disease identification method based on deep learning[J]. HUBEI AGRICULTURAL SCIENCES, 2024, 63(8): 54-60.

参考文献

[1] 蒋妮, 覃柳燕, 叶云峰. 三七病害研究进展[J].南方农业学报, 2011, 42(9):1070-1074.
[2] 于慧妍. 三七研究进展[J]. 世界最新医学信息文摘,2017,17(39):21-22.
[3] 蔡琳,彭鹏. 三七药理作用的研究进展[J].山东化工,2021,50(3):70-71.
[4] 崔宇, 李晓栩, 黄缄. 三七和银杏叶片对高原脱习服大鼠心功能及血清炎症因子的影响及其机制[J]. 中国应用生理学杂志,2019,35(1):34-38.
[5] 李宏伟, 郭富贵, 刘洋, 等. 三七病虫害研究现状与展望[J]. 草业科学, 2019, 36(5):1415-1427.
[6] 张连娟, 高月, 董林林, 等. 三七主要病害及其防治策略[J]. 世界科学技术-中医药现代化, 2017, 19(10):1635-1640.
[7] 杨涛, 陈昱君, 段承俐, 等. 三七黑斑病抗性人工鉴定方法的建立[J]. 云南农业大学学报, 2006(5):549-553, 559.
[8] 王志敏,皮自聪,罗万东,等. 三七圆斑病和黑斑病及其防治[J]. 农业与技术, 2016, 36(1):49-51, 53.
[9] 李欣. 三七主要病害病原菌的快速检测及叶部病害病原菌的鉴定[D]. 昆明:昆明理工大学, 2020.
[10] BOCK C H, POOLE G H, PARKER P E, et al.Plant disease severity estimated visually, by digital photography and image analysis, and by hyperspectral imaging[J]. Critical reviews in plant sciences, 2010, 29(2): 59-107.
[11] BARBEDO J G A. Digital image processing techniques for detecting, quantifying and classifying plant diseases[J]. Springerplus, 2013, 2(1): 1-12.
[12] WANG Y T, DAI Y P, XUE J R, et al.Research of segmentation method on color image of Lingwu long jujubes based on the maximum entropy[J]. EURASIP Journal on image and video processing, 2017, 17(18): 6029-6036.
[13] 明浩,苏喜友. 利用特征分割和病斑增强的杨树叶部病害识别[J]. 浙江农林大学学报, 2020, 37(6):1159-1166.
[14] 王燕妮,贺莉.基于多分类 SVM 的石榴叶片病害检测方法[J].计算机测量与控制, 2020, 28(9):191-195.
[15] 刘永波, 雷波, 胡亮, 等. 机器视觉在 HSV 颜色空间下稻瘟病病程分级判定研究[J]. 农学学报, 2020, 10(10):83-90.
[16] BAI X B, LI X X, FU Z T, et al.A fuzzy clustering segmentation method based on neighborhood grayscale information for defining cucumber leaf spot disease images[J]. Computers and electronics in agriculture, 2017, 136: 157-165.
[17] 包晓敏, 王思琪. 基于深度学习的目标检测算法综述[J]. 传感器与微系统, 2022, 41(4):5-9.
[18] 吴雪, 宋晓茹, 高嵩, 等. 基于深度学习的目标检测算法综述[J]. 传感器与微系统, 2021, 40(2):4-7, 18.
[19] LIU Y, SUN P, WERGELES N M, et al.A survey and performance evaluation of deep learning methods for small object detection[J]. Expert systems with applications,2021,172(4): 114602.
[20] 周惠汝, 吴波明. 深度学习在作物病害图像识别方面应用的研究进展[J]. 中国农业科技导报, 2021, 23(5):61-68.
[21] GIRSHICK R, DONAHUE J, DARRELL T, et al.Rich feature hierarchies for accurate object detection and semantic segmentation[A].Proceedings of the 2014 IEEE conference on computer vision and pattern recognition[C]. 2014.580-587.
[22] GIRSHICK R.Fast R-CNN[A].Proceedings of the 2015 IEEE international conference on computer vision[C]. 2015.1440-1448.
[23] 王远志, 施子珍, 张艳红. 改进的Faster R-CNN用于苹果叶部病害检测[J]. 安庆师范大学学报(自然科学版), 2022, 28(2):26-30.
[24] 陈柯屹, 朱龙付, 宋鹏, 等. 融合动态机制的改进型Faster R-CNN识别田间棉花顶芽[J]. 农业工程学报, 2021, 37(16):161-168.
[25] 李就好, 林乐坚, 田凯, 等. 改进Faster R-CNN的田间苦瓜叶部病害检测[J]. 农业工程学报, 2020, 36(12):179-185.
[26] 曹春号. 基于卷积神经网络的三七叶片病害手机在线检测与预警装置研制[D]. 昆明: 昆明理工大学, 2020.
[27] LIU W, ANGUELOV D, ERHAN D, et al.SSD: Single shot multibox detector[A].European conference on computer vision[C]. 2016.21-37.
[28] 李善军, 胡定一, 高淑敏, 等. 基于改进SSD的柑橘实时分类检测[J]. 农业工程学报, 2019, 35(24):307-313.
[29] REDMON J,DIVVALA S,GIRSHICK R,et al.You only look once:Unified,real-time object detection[A].Proceedings of the 2016 IEEE conference on computer vision and pattern recognition[C]. 2016.779-788.
[30] 吕石磊, 卢思华, 李震, 等. 基于改进YOLOv3-LITE轻量级神经网络的柑橘识别方法[J].农业工程学报, 2019, 35(17):205-214.
[31] 董丽君, 曾志高, 易胜秋, 等. 基于YOLOv5的遥感图像目标检测[J]. 湖南工业大学学报, 2022, 36(3):44-50.
[32] 王超, 张运楚, 孙绍涵, 等. 改进YOLOv5算法的钢筋端面检测[J]. 计算机系统应用, 2022, 31(4):68-80.
[33] 邢晋超, 潘广贞. 改进YOLOv5s的手语识别算法研究[J]. 计算机工程与应用,2022,58(16):194-203.
[34] WANG C Y, LIAO H Y M, WU Y H, et al. CSPNet: A new backbone that can enhance learning capability of CNN[A]. Proceedings of the 2020 IEEE/CVF conference on computer vision and pattern recognition workshops[C]. 2020.390-391.
[35] 张静,农昌瑞,杨智勇.基于卷积神经网络的目标检测算法综述[J].兵器装备工程学报,2022,43(6):37-47.
[36] BOCHKOVSKIY A, WANG C Y, LIAO H Y M. Yolov4: Optimal speed and accuracy of object detection[J]. arXiv e-prints, 2020.DOI: 10.48550/arXiv.2004.10934.
[37] 马琳琳, 马建新, 韩佳芳, 等. 基于YOLOv5s目标检测算法的研究[J]. 电脑知识与技术, 2021, 17(23):100-103.