| The image segmentation technology plays an important part in image processing.The computer X-ray Tomography(Computed Tomography, CT) image has a high definition resolution, if the pathological and anatomical structure is presented clearly by CT important information of an image tend to focus on the edge. We can estimate the lesion size by extracting the edge of the image, and it is very important to understand the outline of the lesion site of the CT image, not only the next image analysis will be simplified, the pathological diagnosis will be much more easily. Three kinds of improved edge detection algorithm based on wavelet transform is put forward in this paper, and comparing with the traditional edge detection method, and the improved method proved to be effective. In this paper, the main research contents and innovation points are as follows:(1) Firstly make simulation experiment based on traditional algorithm and analyzes and summarizes their respective advantages and disadvantages.(2) In view of the traditional fixed threshold of wavelet multi-scale edge detection which can make some weak edge is deleted, without considering the whole image mode, this paper adopts adaptive threshold method according to the characteristics of the image edge by setting dynamic window.(3) Based on a single threshold of wavelet edge detection, weak edge will be filtered because of uneven gray or noise, etc. In order to solve this problem, in the improved algorithm we replace single threshold method with dual threshold, the experimental results show that the improved algorithm makes image edges more smooth, at the same time also can eliminate noise very well.(4) Traditional wavelet transform can only transform the rows and columns of the image. By defining the direction of the wavelet, more direction of the whole image can be detected, and though the simulation results, the edge can be more smooth based on directional wavelet transform, and has less noise. |
PDF Full Text Request