产品表面图像中的缺陷自动检测方法研究  

Research on method of auto-detecting products surface image’s defects

作　　者：岳贤军[1] 

机构地区：[1]南通大学电子信息学院

出　　处：《微计算机信息》2007年第18期297-299,共3页Control & Automation

基　　金：南通大学自然科学基金资助项目(03040080)

摘　　要：仔细设计产品表面缺陷检测的图像识别算法是基于图像处理的产品表面缺陷自动检测系统快速并正确地检测缺陷的关键。以E型磁环表面缺陷的自动检测为例,分析了产品表面图像的特征,用梯度直方图自适应阈值分割的方法对图像进行增强和分割;并提出了一种新的快速识别算法对分割后的目标图像进行检测,实验结果表明了该算法的有效性和实用性。To design recognition algorithm Carefully is key factor that product surface defects auto-detecting system works perfectly ant accurately. Taking example for auto-detecting surface defects of magnetic tach of E-shape, characters of surface Image of product an analyzed; Image is enhanced and segmented synchronously using the method of auto-finding threshold based on grads-histogram; thus novel algorithm is presented to process the target Image. The Results of experiments show that the new algorithm is effective and practicable.

关 键 词：缺陷检测 图像处理 特征识别 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]