CT图像中一类裂缝类缺陷提取与可视化

Extraction and Visualization of A Class of Crack Defect

文章首先对裂缝图像进行形态学去噪预处理。然后,针对CT图像中裂缝类缺陷和背景相连的情形,通过区域生长、边缘提取、斑点噪声消除、形态学膨胀和腐蚀等运算,优先封闭外轮廓,从而建立了裂缝类缺陷区和背景区之间的分界线。最后,通过图像差运算实现了裂缝类缺陷的二维特征提取,并将缺陷形态可视化。

For crack defect,images are preprocessed using the method of morphological denoising. As crack always connects background in micro - computed tomography images, after region growing, edge extraction, elimination of speckle noise, morphological dilation and erosion, the exterior outline is closed, boundary is founded between region of crack and region of background. Operation of image subtraction is used to extract two dimensional feature of crack, and two dimensional cracks are visualized.