TTLD-YOLOv7：非结构化环境下茶树病害的检测算法

俞淑燕; 杜晓晨; 冯海林; 李颜娥

doi:10.13305/j.cnki.jts.2024.03.012

茶叶科学 >

2024 , Vol. 44 >Issue 3: 453 - 468

DOI: https://doi.org/10.13305/j.cnki.jts.2024.03.012

研究报告

TTLD-YOLOv7：非结构化环境下茶树病害的检测算法

俞淑燕 ,
杜晓晨 ,
冯海林 ,
李颜娥

展开

浙江农林大学数学与计算机科学学院,浙江杭州 311300

俞淑燕,女,硕士研究生,主要从事茶树病虫害方面的研究。

收稿日期: 2024-03-17

修回日期: 2024-04-17

网络出版日期: 2024-07-08

基金资助

浙江省“三农九方”科技协作项目(2022SNJF036)

收起

TTLD-YOLOv7: An Algorithm for Detecting Tea Diseases in An Unstructured Environment

YU Shuyan ,
DU Xiaochen ,
FENG Hailin ,
LI Yan′e

Expand

College of Mathematics and Computer Science, Zhejiang Agriculture and Forestry University, Hangzhou 311300, China

Received date: 2024-03-17

Revised date: 2024-04-17

Online published: 2024-07-08

Fold

摘要

茶树病害对茶树种植业和相关行业的影响极为严重。在动态而复杂的茶园环境中检测疾病的传统方法效率低下,检测效果不尽人意。本研究提出一种基于YOLOv7-tiny的模型,增强了茶树病害的细微检测能力。通过整合CoordConv和ECA信道关注机制,本模型在卷积特征图中实现了更高的空间识别能力,并降低了背景噪声对特征识别的影响。进一步的改进包括采用归一化瓦瑟斯坦距离度量和去耦头,以提高对小病斑的检测能力。使用K-means算法根据茶树病斑的特殊性生成了新的锚框,提高了模型的精确性和通用性。对比分析表明,该模型优于现有模型Faster R-CNN、SSD、YOLOv5s、YOLO-Tea、YOLOv7-tiny和YOLOv7,平均精确度提高5.39个百分点,达到了93%。改进后的模型可应用于茶树病害监测。

关键词： YOLOv7-tiny; 茶树病害; 目标检测; 自然环境

本文引用格式

俞淑燕 , 杜晓晨 , 冯海林 , 李颜娥 . TTLD-YOLOv7：非结构化环境下茶树病害的检测算法[J]. 茶叶科学, 2024 , 44(3) : 453 -468 . DOI: 10.13305/j.cnki.jts.2024.03.012

Abstract

Tea diseases have an extremely serious impact on tea plantations and related industries. Traditional methods for disease detection in the dynamic and complex tea plantation environment are inefficient and unsatisfactory. This study proposed that a YOLOv7-tiny-based model enhanced the fine-grained detection of tea tree diseases. By integrating CoordConv and ECA channel attention mechanisms, this model achieved higher spatial recognition capability in convolutional feature maps and reduced the effect of background noise on feature recognition. Further improvements included the use of a normalized Wasserstein distance metric and decoupled heads to improve the detection of small spots. A new anchor frame was generated using the K-means algorithm based on the specificity of tea spots to improve the accuracy and generalizability of the model. Comparative analysis shows that the model outperforms the existing models Faster R-CNN, SSD, YOLOv5s, YOLO-Tea, YOLOv7-tiny, and YOLOv7, with an average accuracy improvement of 5.9 percentage points to 93%. The improved model could be applied to tea disease monitoring.

Key words： natural environment; object detection; tea diseases; YOLOv7-tiny

参考文献

[1] Du M J, Li X X, Cai D S, et al.In-silico study of reducing human health risk of POP residues’ direct (from tea) or indirect exposure (from tea garden soil): improved rhizosphere microbial degradation, toxicity control, and mechanism analysis[J]. Ecotoxicology and Environmental Safety, 2022, 242: 113910. doi: 10.1016/j.ecoenv.2022.113910.
[2] 张洪, 张孟婷, 王福楷, 等. 4种间作作物对夏秋季茶园主要叶部病害发生的影响[J]. 茶叶科学, 2019, 39(3): 318-324.
Zhang H, Zhang M T, Wang F K, et al.Effect of four intercrops on the occurrence of major foliar diseases in summer and fall tea plantations[J]. Journal of Tea Science, 2019, 39(3): 318-324.
[3] 边磊, 何旭栋, 季慧华, 等. 基于机器视觉的小贯小绿叶蝉智能识别的研究与应用[J]. 茶叶科学, 2022, 42(3): 376-386.
Bian L, He X D, Ji H H, et al.Research and application of intelligent identification of Empoasca onukii based on machine vision[J]. Journal of Tea Science, 2022, 42(3): 376-386.
[4] Yang N, Yuan M F, Wang P, et al.Tea diseases detection based on fast infrared thermal image processing technology[J]. Journal of the Science of Food and Agriculture, 2019, 99: 3459-3466.
[5] 杨奉水, 王志博, 汪为通, 等. 人工智能识别茶树病虫害的应用与展望[J]. 中国茶叶, 2022, 44(6): 1-6.
Yang F S, Wang Z B, Wang W T, et al.Application and prospect of artificial intelligence identification of tea Pests and diseases[J]. China Tea, 2022, 44(6): 1-6.
[6] Liu Q, Zhang Y, Yang G.Small unopened cotton boll counting by detection with MRF-YOLO in the wild[J]. Computers and Electronics in Agriculture, 2023, 204: 107576. doi: 10.1016/j.compag.2022.107576.
[7] Nath M, Mitra P, Kumar D.A novel residual learning-based deep learning model integrated with attention mechanism and SVM for identifying tea plant diseases[J]. International Journal of Computers and Applications, 2023, 45: 471-484.
[8] Sun Y Y, Jiang Z H, Zhang L P, et al.SLIC_SVM based leaf diseases saliency map extraction of tea plant[J]. Computers and Electronics in Agriculture, 2019, 157: 102-109.
[9] 林彬彬, 邱新法, 何永健, 等. 茶树病害智能诊断识别算法研究[J]. 江苏农业科学, 2019, 47(6): 85-91.
Lin B B, Qiu X F, He Y J, et al.Research on Intelligent diagnosis and recognition algorithm of tea tree diseases[J]. Jiangsu Agricultural Sciences, 2019, 47(6): 85-91.
[10] Yin C H, Zeng T W, Zhang H M, et al.Maize small leaf spot classification based on improved deep convolutional neural networks with a multi-scale[J]. Agronomy, 2022, 12: 906. doi: 10.3390/agronomy12040906.
[11] 孙道宗, 刘欢, 刘锦源, 等. 基于改进YOLOv4模型的茶叶病害识别[J]. 西北农林科技大学学报(自然科学版), 2023, 51(9): 145-154.
Sun D Z, Liu H, Liu J Y, et al.Tea Disease Recognition Based on Improved YOLOv4 Modeling[J]. Journal of Northwest A&F University (Nature Science Edition), 2023, 51(9):145-154.
[12] Xue Z Y, Xu R J, Bai D, et al.YOLO-Tea : a tea disease detection model improved by YOLOv5[J]. Forests, 2023, 14: 415. doi: 10.3390/f14020415.
[13] 叶荣, 马自飞, 高泉, 等. 基于改进YOLOv5s-ECA-ASFF算法的茶叶病害目标检测[J]. 中国农机化学报, 2024, 45(1): 244-251.
Ye R, Ma Z F, Gao Q, et al.Tea disease target detection based on improved YOLOv5s-ECA-ASFF algorithm[J]. Journal of Chinese Agricultural Mechanization, 2024, 45(1): 244-251.
[14] Bao W X, Zhu Z Q, Hu G S, et al.UAV remote sensing detection of tea leaf blight based on DDMA-YOLO[J]. Computers and Electronics in Agriculture, 2023, 205: 107636. doi: 10.1016/j.compag.2023.107637.
[15] Soeb J A, Jubayer F, Tarin T A, et al.Tea leaf disease detection and identification based on YOLOv7 (YOLO-T)[J]. Scientific Reports, 2023, 13(1): 6078. doi: 10.1038/s41598-023-33270-4.
[16] Chen J Y, Liu H, Zhang Y T, et al.A multiscale lightweight and efficient model based on YOLOv7 : applied to citrus orchard[J]. Plants, 2022, 11: 3260. doi: 10.3390/plants11233260.
[17] Redmon J, Divvala S, Girshick R, et al.You only look once: unified, real-time object detection[C]//IEEE. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas: IEEE, 2016: 779-788.
[18] Wang C Y, Bochkovskiy A, Liao H Y M. YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[C]//The Institute of Electrical and Electronics Engineers. 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Vancouver: IEEE, 2023: 7464-7475.
[19] Liu R, Lehman J, Molino P, et al.An intriguing failing of convolutional neural networks and the CoordConv solution[J]. Advances in Neural Information Processing Systems, 2018, 31: 1-12.
[20] Zhang H T, Tian M, Shao G P, et al.Target detection of forward-looking sonar image based on improved YOLOv5[J]. IEEE Access, 2022, 10: 18023-18034.
[21] IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Seattle, WA, USA: IEEE, 2020: 11534-11542.
[22] Ni H J, Shi Z W, Karungaru S, et al.Classification of typical pests and diseases of rice based on the ECA attention mechanism[J]. Agriculture, 2023, 13: 1066. doi: 10.3390/agriculture13051066.
[23] Xu C, Wang J W, Yang W, et al.Detecting tiny objects in aerial images: a normalized Wasserstein distance and a new benchmark[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2022, 190: 79-93. doi: 10.1016/j.isprsjprs.2022.06.002.
[24] Wang J W, Xu C, Yang W, et al.A normalized gaussian Wasserstein distance for tiny object detection[J]. arXiv preprint arXiv:2110.13389, 2021: 1-12. doi: 10.48550/arXiv.2110.13389.
[25] Song G L, Liu Y, Wang X G.Revisiting the sibling head in object detector[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Seattle, WA, USA: IEEE, 2020: 11563-11572.
[26] Wu Y, Chen Y P, Yuan L, et al.Rethinking classification and localization for object detection[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Seattle, WA, USA: IEEE, 2020: 10186-10195.
[27] Meng C X, Wang Z N, Shi L, et al.SDRC-YOLO : a novel foreign object intrusion detection algorithm in railway scenarios[J]. Electronics, 2023, 12: 1256. doi: 10.3390/electronics12051256.
[28] Li G B, Shi G L, Jiao J.YOLOv5-KCB : a new method for individual pig detection using optimized K-means, CA attention mechanism and a Bi-directional feature pyramid network[J]. Sensors, 2023, 23: 5242. doi: 10.3390/s23115242.
[29] Xue J L, Cheng F, Li Y Q, et al.Detection of farmland obstacles based on an improved YOLOv5s algorithm by using CIoU and anchor box scale clustering[J]. Sensors, 2022, 22: 1790. doi: 10.3390/s22051790.

Options

文章导航

模态框（Modal）标题

摘要

本文引用格式

Abstract

参考文献