Application Research On Median Filtering Technique In Image Processing

In the process of image formation, image transmission, image reception and image processing, the disturbance of image can not be avoided, the noise would deteriorate the quality of images, which can make the images become obscured, so the image filtering techniques take an important position in image processing field. In many image-processing techniques, Median Filtering Technique is nonlinear Filtering Technique that can overcome the shortage of linear Filtering Technique, which can protect the detail information of images when restraining noises, so Median Filtering Technique is researched and applied broadly. Until now it is still the hot focus in image processing research.The research on the application of median filtering in the course of image processing is expounded in this thesis. The author analyzed the theory of the traditional median filtering technique. Based on defects of median filtering algorithm, three improved median filtering methods are presented in the thesis. Because each pixel of the image is processed in the traditional method, the gray scale value of uncontaminated pixels are also changed. In fact, the contaminated pixels are in a minority. Therefore, such method not only waste the time, but also dim the image. In this case, the filtering algorithm based on the pixel selected to remove the noise is presented in the thesis. Besides, the algorithm based on the extremum detection is presented, depending on many noise-detecting methods such as the detecting of spiced salt noise and gaussian noise. Finally, assimilating the thought of pixel classification, the author develops the self-adapting median filtering method and adopts advanced filtering window technique, self-adapting algorithm without detail information lost, to keep the balance between detail-protecting and noise-removing.Simulation experimental results evidenced efficiency of the several algorithms proposed by this thesis.