摘要
基于机器视觉方法检测织物表面图案几何形状,研究棉质纺织品表面图案边缘的检测效果,通过对比找到适合工业应用的边缘检测算法。选取多种典型的边缘检测算子为考察对象,首先对各自边缘检测的结果图像及其信息熵进行对比,然后分析各算子在目标区域分割、内部条纹及局部图案细节等多方面的表现,证明Sobel算子得到的图像边缘与背景区域的灰度差异明显,灰度信息熵最小,有利于图像几何信息的识别;而且Sobel算子对织物表面及背景区域的纹理不敏感,识别结果:图像中噪声及后续处理难度较小,适合棉质袜子表面几何形状的检测。
Based on the machine vision method to detect the geometric pattern of the fabric surface pattern,the detection effects of the cotton fabric edge were investigated to find the edge detection algorithm which is suitable for industrial applications. A variety of typical edge detection operators were selected. The result images of each edge detection and their information entropy were compared. Then,the performance of each operator in the segmentation of the target region,the internal stripes and the details of local patterns were analyzed. The results showed that there is a significant difference in the gray level between edge and background of the edge image obtained by Sobel operator. It is due to the gray has the smallest information entropy,which is most conducive to the identification of image geometric information. Moreover,Sobel operator is suitable for the detection of the surface geometry of cotton socks,because it is insensitive to the texture of the fabric surface and the background area,which brings it easy for the subsequent processing.
作者
杨鹏程
杨社强
肖渊
刘洋
YANG Pengcheng;YANG Sheqiang;XIAO Yuan;LIU Yang(School of Mechanical and Electrical Engineering,Xi′an Polytechnic University,Xi′an,Shaanxi 710048,China)
出处
《毛纺科技》
CAS
北大核心
2018年第9期79-83,共5页
Wool Textile Journal
基金
陕西省科技厅自然科学基础研究计划-青年人才项目(2015JQ5196)
陕西省青年人才托举计划(20160124)
西安工程大学博士科研启动金(BS1401)
关键词
织物表面质量
几何形状检测
边缘检测
图像分析
fabric surface quality
geometric shape detection
edge detection
image analysis
