工业CT含噪图像中的裂纹探测  

Crack detection of industrial CT noisy image

作　　者：李宗剑[1,2] 曾理[1,3] 

机构地区：[1]重庆大学光电技术及系统教育部重点实验室ICT研究中心,重庆400044 [2]重庆大学光电工程学院,重庆400044 [3]重庆大学数理学院,重庆400044

出　　处：《光学精密工程》2010年第2期451-457,共7页Optics and Precision Engineering

基　　金：国家自然科学基金资助项目(No.60672098;No.60972104)

摘　　要：提出一种基于快速Beamlet变换的工业CT含噪图像的裂纹探测方法。首先,分析图像在单尺度下的Beamlet组成以及Beamlet间的相互关系,设计出一种快速Beamlet变换。基于快速Beamlet变换,引入一个关于吻合度的控制量,并结合Beamlet自身的多尺度树型结构,采取"自上向下"寻找目标函数最优值的思路,得到裂纹的探测结果。最后,结合探测结果的相邻区域的像素特性,提取出含噪裂纹的区域边界。分别对含有裂纹的CT图像,以及叠加方差为0.1的高斯白噪声图像,叠加强度为0.1的椒盐噪声图像进行探测实验。结果表明,与基于Laplace、Canny或小波的探测方法相比,该方法能有效探测到工业CT含噪图像中的裂纹。因为Beamlet是以线基的方式分析图像数据,所以该方法对噪声干扰具有很好的抑制能力,成功实现了工业CT含噪图像中的裂纹探测。A crack detection method of Industrial Computed Tomography（ICT） noisy images based on fast Beamlet transform is presented in this paper. After analyzing the composition and relation of Beamlet in a mono-scale, a fast Beamlet transform is proposed. On the basis of the Beamlet trans-form, a control variable about the relativity is introduced. Then, combining the tree-structure of Beamlet＇s mutiscale and a ＂top to bottom＂ inter-scale inhibition to optimize the object function, the crack is detected. Finally, considering the near pixels of detection result of the crack, the edge of crack domain is extracted. A numerical experiment on the images including an original ICT noisy image, a Gaussian white noise image with a variance of 0.1 and a salt ＆ pepper noise image with a superposition density of 0.1 is carried out. Compared with the methods of Laplace, Canny or wavelet, the proposed method can detect the crack of ICT noisy image more effectively. Because the Beamlet transform uses lines to analyze the image data, the proposed method has a robustness to noises.

关 键 词：工业CT 裂纹探测 快速Beamlet变换 噪声图像 

分 类 号：TG115.281[金属学及工艺—物理冶金] TP391[金属学及工艺—金属学]