| 注册
首页|期刊导航|棉花学报|基于CID-YOLO的籽棉杂质检测方法

基于CID-YOLO的籽棉杂质检测方法

徐耀 李宁冉 薛胜 朱婷婷 陈东胜 倪超

棉花学报2025,Vol.37Issue(6):514-524,11.
棉花学报2025,Vol.37Issue(6):514-524,11.DOI:10.11963/cs20250049

基于CID-YOLO的籽棉杂质检测方法

Seed cotton impurity detection method based on CID-YOLO

徐耀 1李宁冉 1薛胜 1朱婷婷 1陈东胜 2倪超1

作者信息

  • 1. 南京林业大学机械电子工程学院,南京 210037
  • 2. 奎屯银力棉油机械有限公司,新疆 奎屯 833200
  • 折叠

摘要

Abstract

[Objective]This study aims to address the issues of impurity contamination in machine-picked seed cotton in Xinjiang and the low efficiency of manual sorting,we propose a highly efficient and precise method for detecting impurities in seed cotton to enhance cotton quality.[Methods]An improved CID-YOLO model is introduced based on YOLO v11.Firstly,a normalization-based attention mechanism module is introduced to enable the network to balance target channel features and pixel features while reducing parameter count and computational complexity,thereby improving detection accuracy.Secondly,according to the characteristics that the detected target is easy to appear at the edge,the lightweight adaptive convolution module is used to replace some convolution modules in the backbone network and neck network,which can better preserve the edge and structure information of the sampling phase,and effectively reduce the amount of parameters.Finally,an efficient layer aggregation network is used to replace the improved pyramid pooling module in the backbone network to enhance the ability of feature multi-scale extraction and information aggregation,so as to improve the perception performance of targets of different scale targets in complex background.[Results]Experimental results demonstrate that the CID-YOLO model achieves outstanding performance in seed cotton impurity detection,with a detection accuracy of 92.1%,recall of 89.5%,F1 score of 90.8%,and mean average precision of 92.8%.Its overall recognition effectiveness significantly outperforms existing YOLO series models.[Conclusion]The proposed CID-YOLO method enables rapid and precise identification of drip irrigation tapes,colored impurities,cotton stalks,and plastic mulch in machine-picked seed cotton.It provides effective technical support for real-time seed cotton impurity detection and holds promising application prospects.

关键词

籽棉/杂质/YOLO v11/CID-YOLO

Key words

seed cotton/impurity/YOLO v11/CID-YOLO

引用本文复制引用

徐耀,李宁冉,薛胜,朱婷婷,陈东胜,倪超..基于CID-YOLO的籽棉杂质检测方法[J].棉花学报,2025,37(6):514-524,11.

基金项目

新疆维吾尔自治区重大科技专项(2022A01009) (2022A01009)

棉花学报

OA北大核心CSCDCSTPCD

1002-7807

访问量0
|
下载量0
段落导航相关论文