The Application Of Wavelet Analysis In Image Edge Detection

Edge Detection is one of the basic contents on image processing and analysis, on which people have not as yet found a satisfactory way out. The edge which is one of the fundamental features of an image, embodying the other features such as position, outline , etc., is widely used in image analysis and processing like feature description, image segmentation, image enhancement, pattern recognition and image compression. So Edge Detection is the research hotspot in the technology of image processing and analysis all the while, for which the new theory and methods are put forward constantly.The traditional edge detectors such as Roberts,Sobel etc., mainly use differential method detect edges, are sensitive to noise, and often strengthen noise when they detect edges.Canny edge detector may smooth edges, then obtain not-signal pixel edges.Wavelet multiscale edge detection algorithm is effective to remove noise from images without blurring the edges.It can detect finer edges when scale is small, also can suppress noise when the scale is enlarged.It can obtain accurate signal pixel edges.The main research of this paper as followings:(1)The paper expatiates on the main theory of wavelet analysis and classical theory of digital image edge detection, and gives the experiment results,compares and evaluates their advantage and disadvantage.(2)In fact, most images is laced with noise.In order to get the edge of the image, removing noise should be resolved first. Both edge and noise information are high-frequency information, we should try to remove noise before maintain more edge information, but the loss of edge information is evident and inevitable in the denoising process. In this paper, an image threshold denoising method is presented: Combine wavelet image threshold denoising and edge detection. Experimental results presented in this paper show that our method can keep image's edges from damaging when images which include different variance noises are denoised. (3)In this paper, we go deep into the detecting noises image edges with wavelet analysis. There are two methods: sole scale detection and multiscale detection. The key point of wavelet sole scale edge detection is to select a proper threshold. After VisuShrink universal method and BayesShrink Bayes threshold are introduced, and further their advantages and disadvantages are compared, we present a new threshold, and experimental results show that new threshold is advantaged. Multiscale detection method mainly uses different scale through the valid combination of the edge detection operators, and detects image edge correctly. This paper studies the edge detection algorithm which is the wavelet domain scale multiplication. The algorithm multiplys the wavelet coefficients at adjacent scales, it can suppress noises and including causal relation at different scales, it think that the maximum of product by shifting and multiplying the wavelet coefficients is edge.Finally, the paper puts forwards the scale integration laws in 4 neighboring scale wavelet coefficients edge detection result of charts,and the results analysis of the experiment pictures indicate that the method has a better accuracy and precision in the Edge Detection.