基于自适应形态学滤波的ICT图像缺陷提取  

Defect Extraction of ICT Images Based on Adaptive Morphological Filtering

作　　者：方黎勇[1] 李柏林[1] 何朝明[1] 

机构地区：[1]西南交通大学机械工程学院,四川成都610031

出　　处：《西南交通大学学报》2009年第1期41-44,65,共5页Journal of Southwest Jiaotong University

基　　金：国家863计划资助项目(2003AA4Z3200)

摘　　要：针对ICT图像缺陷现有提取方法的不足,提出了一种自适应的数学形态学滤波方法.该方法首先对整张图像进行分水岭变换,根据区域大小自适应地选取结构元素区域.然后对该区域边界进行椭圆拟合,以确定结构元素模板.根据形态学原理运用此模板对图像进行闭运算,从而获得图像背景.最后,用背景图像减去原始图像,即可获得缺陷的图像.试验结果表明,该方法能获得更好的图像缺陷提取效果.To overcome the shortcomings of the current defect extraction methods for ICT （industrial computerized tomography） images, an adaptive filtering method based on the mathematical morphology was proposed. Firstly, with this method, the primary region of a structural element is automatically chosen in the light of region size after the watershed transformation of full ICT images. Secondly, the ellipse fitting operation is applied to the boundary of the primary region to get the final structural element. Then the background images are gotten by the closing operation of the mathematical morphology on the original images using the final structural element. Finally, the defect images are obtained by subtracting the original image from the background images. The experimental results show that compared with the current methods, the proposed method can obtain a good result of defect extraction of ICT images.

关 键 词：数学形态学滤波 分水岭算法 缺陷提取 结构元素 椭圆拟合 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]