安徽农业科学2025，Vol.53Issue(12)：193-195,3.DOI:10.3969/j.issn.0517-6611.2025.12.042
基于改进YOLOv7的番茄果实成熟度检测方法
Maturity Detection of Tomato Based on Improved YOLOv7
摘要
Abstract
To accurately identify the maturity of tomato target in facility environments,improve detection efficiency and quality,and achieve intelligent harvesting.An enhanced YOLOv7 model is proposed for detecting the maturity of target tomato fruits.This approach enhances the model's focus on the target regions of input data by integrating the CBAM attention mechanism;employing the Soft-NMS algorithm effectively prevents missed detections due to high-density overlapping targets being suppressed,thereby enhancing detection performance;optimizing the original loss function EIOU and replacing it with SIOU,the experimental results show that the improved YOLOv7 model has a detection preci-sion of 93.1%,a recall rate of 90.8%,and a mean average precision of 94.8%.Compared with the original YOLOv7 and YOLOv5 models,it has improved in detection precision,recall rate,and mean average precision,providing technical reference for tomato harvesting in complex environments.关键词
番茄/YOLOv7/成熟度检测/目标识别Key words
Tomato/YOLOv7/Maturity detection/Target recognition分类
农业科技引用本文复制引用
谭荣英,李小明,李军辉,宋一明..基于改进YOLOv7的番茄果实成熟度检测方法[J].安徽农业科学,2025,53(12):193-195,3.基金项目
北京农业职业学院2023年度科技创新项目(XY-YF-23-07). （XY-YF-23-07）
