基于最大熵快速迭代算法和边缘算子的织物疵点分割  

Fabric Defect Segmentation Based on Fast Iterative Algorithm of Maximum Entropy and Marginal Operator

作　　者：赵静[1] 赵东晓 高伟 吴付英[1] 

机构地区：[1]青岛理工大学临沂校区机电系,山东临沂273400 [2]临沂市特殊教育中心,山东临沂273400 [3]青岛恒星职院自动化学院,山东青岛266100

出　　处：《纺织科技进展》2012年第3期40-42,83,共4页Progress in Textile Science & Technology

摘　　要：将信息熵引入图像处理中,把疵点图像分为背景和目标两部分,分别对两个区域进行处理;通过求最大熵值的快速迭代算法,在满足信息熵最大要求的前提下对织物疵点区域进行分割,然后利用Canny边缘检测算子对分割后的疵点图像进行边缘检测,从而达到识别织物疵点的目的。仿真实验结果表明,将最大熵快速迭代算法与边缘算子结合进行疵点分割识别的方法是有效的。It introduced information entropy into the image processing, divided the defect image into two parts： the background and objectives, and dealed with the two regions respectively. Through calculating the maximum entropy by fast iterative algorithm, under the premise of meeting the maximum demands of the information entropy, it segmented the region of fabric defect, then used the Can- ny edge detection operator to defect the image edge , so as to achieve the purpose of identifying fabric defects. The simulation exper- ment results showed that defect segmentation based on combining fast iterative algorithm of maximum entropy with the edge operator recognition was effective.

关 键 词：织物疵点 疵点分割 信息熵 边缘检测 图像处理 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]