中国农机化学报2026，Vol.47Issue(1)：87-93,7.DOI:10.13733/j.jcam.issn.2095-5553.2026.01.013
基于改进YOLOv7的果园苹果目标识别方法研究
Research on target recognition method for orchard apples based on improved YOLOv7
摘要
Abstract
Apple recognition is an important link in the automated apple picking.To solve the problems of low recognition efficiency,missed detection,wrong detection and low detection accuracy of apples in orchards,an improved algorithm based on YOLOv7 is studied.Firstly,CBAM attention mechanism is integrated into the backbone network of the detection model to solve the problem of overlapping shielding of apples and difficult identification of small targets in complex orchards,the model pays more attention to the semantic features of apple.Then,the CIoU loss function in the original YOLOv7 network model is replaced with the MPDIoU loss function to alleviate the problem of imbalance between positive and negative samples in the original loss function.Finally,the conventional convolutional layers in the backbone network and the head network are replaced with PConv,which has fewer parameters and less computational cost,thereby improving the detection speed of the model.The results show that,compared to the original YOLOv7 model,the enhanced model increases average precision by 3.9%,and decreases parameter and computational complexity by 9.2%and 11.4%respectively,and improves the detection speed to 87 f/s.The improved model can realize the accurate identification of orchard apples,effectively reduce the problem of apple error detection and leakage detection,and the detection speed can meet the real-time requirements.关键词
苹果识别/YOLOv7/果园/损失函数/CBAM注意力机制Key words
apple recognition/YOLOv7/orchard/loss function/CBAM attention mechanism分类
农业科技引用本文复制引用
Yu Jiacheng,Yang Linchu,Xia Zilin,Gu Jinan,Bao Han..基于改进YOLOv7的果园苹果目标识别方法研究[J].中国农机化学报,2026,47(1):87-93,7.基金项目
江苏省科技项目(BE2021016-3) （BE2021016-3）
