基于遗传算法与二维最大熵的编织袋缺陷检测  

Woven Bag Defect Detection Based on Genetic Algorithm and Two-Dimensional Maximum Entropy

作　　者：黄露 夏军勇 吴庆华[1] 钟飞[1] HUANG Lu;XIA Junyong;WU Qinghua;ZHONG Fei(School of Mechanical Engineering,Hubei University of Technology,Wuhan 430068,Hubei)

机构地区：[1]湖北工业大学机械工程学院,湖北武汉430068

出　　处：《轻工机械》2021年第5期69-73,78,共6页Light Industry Machinery

基　　金：湖北省技术创新专项(重大项目):全自动纸塑复合袋成型装备研发(No.2018AAA026);湖北工业大学博士启动基金(No.BSQD2017001)。

摘　　要：编织袋图像存在的灰度不均匀、噪声污染大等问题影响了缺陷检测的精度和效率,为此,课题组提出一种基于改进遗传算法与二维最大熵的编织袋缺陷快速检测方法。先对编织袋图进行预处理,消除图中存在的背景噪声以及细微像素点;接着利用与二维最大熵结合的改进遗传算法快速选取图像分割的最佳阈值,提高分割速度与精度;最后利用连通域标记对缺陷进行统计与定位。实验结果表明:该方法对编织袋缺陷的分割精度与速度优于迭代阈值法、一维最大熵法以及结合一般遗传算法的二维最大熵法。新方法能够精准、高效地检测出编织袋的质量缺陷。In view of the problems of uneven grayscale and large noise pollution in the woven bag image that affects the accuracy and efficiency of defect detection,a fast defect detection method for woven bags based on improved genetic algorithm and two-dimensional maximum entropy was proposed.Firstly,the woven bag image was preprocessed to eliminate the background noise and fine pixels in the image,and then the improved genetic algorithm combined with the two-dimensional maximum entropy was used to quickly select the best threshold of image segmentation to improve the segmentation speed and accuracy.Finally,the connected domain markers were used to count and locate the defects.Experimental results show that the accuracy and speed of the proposed algorithm for detecting woven bags are better than the iterative threshold method,the one-dimensional maximum entropy method and the two-dimensional maximum entropy method combined with general genetic algorithm.The new method can detect the quality defects of woven bags accurately and efficiently.

关 键 词：缺陷检测 编织袋 改进遗传算法 二维最大熵 连通域 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TH39[自动化与计算机技术—计算机科学与技术]