农机化研究2025，Vol.47Issue(12)：45-52,8.DOI:10.13427/j.issn.1003-188X.2025.12.006
基于改进YOLOv7的丘陵地区茶树叶部病虫害识别方法
Identification Method of Diseases and Pests on Tea Leaves in Hilly Areas Based on Improve YOLOv7
摘要
Abstract
To address the problems existing in the detection and recognition of diseases and pests on tea leaves in hilly areas,such as small image targets,serious missed and false detection,and high manpower consumption,inaccurate recognition,low efficiency,a method of tea leaf diseases and insect pests recognition in hilly areas based on improved YOLOv7 was proposed.Firstly,the AC-E-ELAN module was introduced to promote the model to obtain rich image fea-ture information and enhance the learning and reasoning ability of the model.Then,DCNv2 and CBAM modules were added to improve the model's ability to extract small features and resist interference.Finally,the CARAFE upsampling operator and WIoU loss function were applied to improve the efficiency and effect of model recognition.The improved YOLOv7 model was compared with capsule network,residual dense network,YOLOv7,YOLOv8 to identify four common diseases and insect pests in tea leaves.The results proved that the improved YOLOv7 model outperformed other models in all evaluation indicators,and had higher recognition accuracy and ability.The research provided new ideas and methods for improving the production quality and yield of tea trees,and offered references and guidance for promoting the monito-ring and early warning of crop pests and diseases,unified prevention and control,and the development of smart agricul-ture.关键词
茶树叶部/病虫害识别/改进YOLOv7/深度学习/丘陵地区Key words
tea leaf section/identification of pests and diseases/improve YOLOv7/deep learning/hilly area分类
信息技术与安全科学引用本文复制引用
彭炜峰,罗江华..基于改进YOLOv7的丘陵地区茶树叶部病虫害识别方法[J].农机化研究,2025,47(12):45-52,8.基金项目
重庆市教委科技研究项目(KJZD-M202503801) （KJZD-M202503801）
