基于图像最大熵的织物疵点检测方法  

Detection Approach of Fabric Defect Based on the Maximum Entropy of Images

作　　者：赵静[1] 于凤芹[1] 

机构地区：[1]江南大学通信与控制工程学院,江苏无锡214122

出　　处：《纺织科技进展》2010年第2期64-65,69,共3页Progress in Textile Science & Technology

摘　　要：对织物表面疵点自动识别方法进行了探讨。将信息熵引入图像处理中,先通过最大熵快速迭代算法对织物疵点区域进行分割,把疵点图像分为背景和目标两部分;然后找出疵点区域的中心并求出疵点区域在纬向和经向上的方差;最后通过两者的比值与设定常数的比较,判断出疵点类型。仿真实验表明该方法对常见织物疵点的检测是有效的。It studied the automatic identifying approach of the fabric surface defects. The information entropy was introduced into the image processing, it segmented the region of the fabric defect by using the fast iterative algorithm of the maximum entropy, divid- ed the image into two parts which were the background and the objectives, then the center of the faults region was found out and the variance of zonal and warp in the faults region was determined, finally by comparing the ratio of the two parts with the predetermined constants, the type of defects was determined. Simulation experiment showed that this method is effective to detect common fabric defects.

关 键 词：织物疵点 信息熵 疵点自动识别 疵点检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]