小波统计法的表面缺陷检测方法  

Surface defect inspection based on wavelet statistical analysis

作　　者：张学武[1] 吕艳云[1] 丁燕琼[1] 梁瑞宇[1,2] 

机构地区：[1]河海大学计算机及信息工程学院,江苏常州213022 [2]东南大学信息科学与工程学院,江苏南京210096

出　　处：《控制理论与应用》2010年第10期1331-1336,共6页Control Theory & Applications

基　　金：国家自然科学基金资助项目(60872096);河海大学青年科技创新基金资助项目(XZX/07/3002.3);常州市科技创新基金资助项目(CN2008033)

摘　　要：针对铜带表面缺陷的特点,基于小波统计方法设计了对铜带表面缺陷检测的系统.首先把铜带表面图像分为互不重叠的子图像,再把子图像分为多个小波处理单元,对每个小波处理单元进行db4紧支集正交小波一级分解,在此基础上进行HotellingT2统计检测缺陷.最后利用支持向量机进行缺陷分类.实验中将基于小波的统计方法和基于灰度的差影法进行比较,结果证明本文提出的方法识别率高,特别对于用一般算法识别率较低的"起皮"缺陷达到96.7%的识别率.According to the characteristics of defect image on copper snips surface, we design a surface defect detection system on the basis of wavelet-based multivariate statistical approach. First, the surface image is divided into sub-images; each sub-image is further segmented into multiple wavelet processing units. Then, each wavelet processing unit is decomposed by 1-D db4 wavelet function. The multivariate statistics of Hotelling T2 is then applied to detect the defects, and Support-Vector-Machines（SVM） is used as the defect classifier. The defect detection performances of the proposed approach are compared with those of the grayscaledifference method. Experimental results show that the proposed method has higher performances on identification; the recognition rate for the ripple defects achieves 96.7% which is unattainable by common algorithms.

关 键 词：缺陷检测 小波统计 强反射金属 机器视觉 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]