Defect segmentation algorithm for low contrast slice images of cone-beam computed tomography

According to the particularities of Cone-Beam Computed Tomography (CBCT) slice images, the low contrast segmentation problem in the whole image is translated into the higher contrast segmentation problem in the local topological structures by obtaining the gray topological structures in the slice image. Then the class of the current template region is judged by the two thresholds and four detecting templates in each topological structure. The detecting result of the topological structure is the summation of the detecting results in the four directions. The final detecting result is the summation of the detecting results of all topological structures. This algorithm was adopted to processing the CBCT slice image of wax model of hollow turbine blade. The experiment result shows that the low contrast defects in the slice image are extracted effectively.