棉纺织技术2025，Vol.53Issue(6)：66-73,8.
基于改进CLAHE与场景分割的生丝疵点检测算法
Raw silk defect detection algorithm based on improved CLAHE and scene segmentation
摘要
Abstract
In order to improve the detection accuracy and efficiency of raw silk defects,a raw silk defect detection algorithm based on machine vision was proposed.Firstly,in the preprocessing stage,the improved limited contrast histogram equalization(CLAHE)algorithm was used to equalize the image brightness distribution and preliminarily enhance the image quality.Secondly,the scene segmentation algorithm was used to solve the problem of uneven low illumination that may existed in raw silk defect detection in darkroom,and the artificial bee colony optimization algorithm was used to segment and locate the suspected defect area.Finally,according to the raw silk characteristics of the suspected defect area,the local area directional gradient histogram and local area threshold segmentation were used to further accurately locate the defect.The defect detection test was carried out on the raw silk images collected in the actual artificial detection environment,and it was compared with the artificial detection method and other defect detection algorithms.The results showed that the comprehensive positive detection rate of the algorithm was 93.1%,and the detection time of a single sheet was 0.693 s.It is considered that the algorithm has high accuracy and strong robustness in the case of raw silk images under uneven low illumination conditions.关键词
生丝/疵点检测/改进CLAHE算法/场景分割/人工蜂群优化算法/方向梯度直方图Key words
raw silk/defect detection/improved CLAHE algorithm/scene segmentation/artificial bee colony optimization algorithm/directional gradient histogram分类
轻工业引用本文复制引用
曾凡高,李子印,汪小东,叶飞,姚晓娟,杨言语..基于改进CLAHE与场景分割的生丝疵点检测算法[J].棉纺织技术,2025,53(6):66-73,8.基金项目
国家市场监督管理总局科技计划项目(2022MK048) （2022MK048）
浙江省市场监督管理局青年科技项目(QN2023444) （QN2023444）
浙江省基础公益研究计划项目(LGN20F50001) （LGN20F50001）
