基于局部处理的X射线图像裂纹缺陷自动检测  

Automatic detection of crack defects in X-ray image based on local processing

作　　者：娄联堂[1] 何慧玲 石胜平 LOU Liantang;HE Huiling;SHI Shengping(College of Mathematics and Statistics,South⁃Central University for Nationalities,Wuhan 430074,China;Technological Research Institute,CRRC Yangtze Corporation Limited,Wuhan 430212,China)

机构地区：[1]中南民族大学数学与统计学学院,武汉430074 [2]中国中车集团长江车辆有限公司工艺研究所,武汉430212

出　　处：《中南民族大学学报（自然科学版）》2020年第1期98-102,共5页Journal of South-Central University for Nationalities：Natural Science Edition

基　　金：国家自然科学基金资助项目(60975011);中央高校基本科研业务费专项资金资助项目(CZW15051;YZZ13003)

摘　　要：研究了X射线图像裂纹缺陷自动检测问题.首先根据裂纹缺陷特征及位置信息提取感兴趣区域,使检测到的缺陷区域准确地包含实际缺陷;然后提出了一种结合局部方差和局部直方图均衡化的图像增强方法,实现了图像的对比度增强,使裂纹缺陷肉眼可见;接着利用阈值分割和形态学方法对裂纹缺陷进行分割提取;最后根据裂纹缺陷的几何特征实现了裂纹缺陷的检测.实验结果表明:相比于仅利用局部方差或局部直方图均衡化对图像做一次对比度增强,此方法能有效增强缺陷特征,更利于缺陷检测.The problem of automatic detection of crack defects in X-ray images is studied. Firstly, the region of interest is extracted according to the feature and location information of crack defects, so that the defect areas detected accurately contain the actual defects. Then, an image enhancement method combining local variance and local histogram equalization is proposed to enhance the contrast of the image and make the crack visible to the naked eye. Next, threshold segmentation and morphological analysis are used to segment and extract the crack defects. Finally, the detection of crack defects is realized according to the geometric characteristics of crack defects. The experiment results show that this method can effectively enhance the defect features and is more conducive to defect detection than only using local variance or local histogram equalization to enhance the image contrast once.

关 键 词：裂纹缺陷自动检测 感兴趣区域 局部增强 阈值分割 形态学处理 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]