| Components will produce unexpected inner defects when production or producessing which needs to check defects in component whether within the scope of permissible to decide whether the component can be used. As CT scan can only get 2D slice-image, if want to get 3D information of defect, it needs engineer's experience and spatial visualization which has great limitations. In order to get more accurate information of the slice-images for 3D reconstruction, such as defect position, shape, size, density distribution. The edge detection and interpolation are important parts of 3D reconstruction. Their results have a direct impact on reconstruction. This paper takes this as a background to research edge detection and interpolation algorithm of inner defect in component systematically.In edge detection, aimed at present edge detection methods can not detect edge accurately, edge vague and exist false edges, edge detection method based on improved Sobel operator and clustering mothod was proposed.In aspect of slice images interpolation, for the present shape interpolation algorithms to non-similar shapes,no coincide location on adjacent slices, concave contours for linear interpolation products falsity interpolation, takes a long time computation and takes up large storage space, a algorithm was proposed for shape interpolation based on skeleton points to avoid falsity interpolation. In the case of image interpolation for 3D volume models, present methods either lack of the capability of interpolation gray levels and shapes at the same time, or need higher computation cost. In order to solve the problem, the interpolation method based on the mathematical morphology and geodesic distance function was proposed.Edge detection and interpolation experiments of inner defects of the components are done through VC++ simulation experiment. The results indicate feasibility and effectiveness of the edge detection and interpolation algorithm of this paper. |
PDF Full Text Request