Edge Detection And Interpolation Algorithm Research Of Inner Crack Defect

CT image 3D reconstruction has become a hot research at present at home and abroad. It gained two-dimensional gray images in non-destructive and reconstructed three-dimensional images which conformed to people's vision. Because the inner crack defect has the attribute of irregularity and discontinuity, the simple edge detection and image interpolation are not fully applicable to such defects. This paper takes this as a background to carry on edge detection and interpolation technique of inner crack defect in consecutive slices systematically.Firstly, in aspect of slice images edge detection, the paper introduces related knowledge of the image edge detection, expatiates systematically the classical edge detection algorithms and analyzes the problems in the classical algorithms. A new image edge detection algorithm which is based on wavelet multi-scale multiplication is concerned with doing multi-scale wavelet decomposition to the image with multi-scale of edge information and the modulus maximum value of wavelet transform, and then making a multiplication of consecutive scale wavelet coefficient to enhance edge and achieving the final image edge with double threshold to remove noise components.Secondly, in aspect of slice images interpolation, it's necessary to preprocess them. This paper introduces gray interpolation and contour interpolation mathematical descriptions and algorithms of defects. An image interpolation algorithm based on the theory of wavelet transforms is concerned with seeking the optimum corresponding points from the low frequency wavelet sub-images obtained through wavelet transform from the original images and then gaining interpolation points'gray values via the optimum corresponding points'gray values which generates gray values in the interpolation image, and the new interpolation image is arisen through the wavelet inverse transform in the end. Finally, edge detection and interpolation experiments of the inner crack defects are tested through VC++ simulation experiment. The result indicates that this algorithm avoids edge blurring , and ensures the accuracy of edge continuation and positioning.