Research On Optical Image Edge Detection Based On Wavelet Transform And Morphology

Optical images are very important for people to obtain information, while it is one of basic character of an image that contains a lot of information of image, which is important cause to process optical images, such as image segmentation, image data compression, image recognition. Because noise and edge of image are high frequency signal, it is difficult to distinguish between noise and edge. And it is difficult to make a decision between denoising and locating edge. Therefore, edge detection of image is always a difficult and hotspot task in image processing. Edge detections of image based on wavelet domain and morphology are researched, and some improved algorithms are proposed. The paper is organized as follows:First of all, we introduce the research significance and situation of edge detection of image. Then, two difficulty in edge detection and evaluation standard of edge detection are introduced, etc. And some fundamental knowledge of wavelet and morphology are introduced. These provide theoretical basis for the subsequent improved algorithm.Secondly, in order to made thresholds have the characteristic of local Adaptive, a new adaptive double thresholds are proposed. For the pixel of value of module between two thresholds, this threshold method make use of the difference between edge and noisy to filter candidate edge points. And this process improves veracity of edge detection. By this, we can wipe off noisy and hold detailed information of images, and improve visual effect of image edge detection.Thirdly, by synthesizing multi-scale structrral element in denoising method and sharp method of morphology, we can wiped off noisies effectively, and sharpen detailed edges of image. Based on this, by adopting edge variance to adaptively syncretize edge information of different direction, which were detected by different structrral elements, we can gain the edge image almost without noisy, and hold edge information perfectly. And the edges have good continuity.At last, through analyzing the relativity of wavelet coefficient of signal and noisy with different scales, and studing the advantages and disadvantages of the traditional scale product function, this paper constructs a new scale product function to modify wavelet coefficients. And through synthesized it with method of morphology, the shortcoming of using one of two method separately was overcame, and make use of edge information of image adequately. At the same time, the noisys were wiped off and the detailed information of images were hold, so the effect of image edge detection was improved more.