Research On Algorithms Of Image Filtering And Edge Detection For Image Enhancement

Computer Vision (CV) which based on image processing has been used in many areas. Thus more efficient methods of image processing are required with the development of CV. This dissertation mainly concerns on the relative technologies of image filtering and edge detection for image enhancement. The researches aim at designing more accurate and efficient algorithms for CV.The main contents of the dissertation are listed below.1. Research on algorithm of impulse noise removal. Aiming at the typical impulse noise—Salt & Pepper noise, A adaptive switching median filter which is based on switch program is presented. The basic steps are: First, classifying and detecting the noise points according to the characteristics of the impulse noise, and then filtering the detected points and keep the signal points to their own gray values. In the noise detection stage, Because of adaptive detection window and adaptive threshold technologies are used, the detection of impulse noise is more reasonable and accurate. In the filtering stage, only the signal points are used with the median filter, the results show that the image details are effective protected.2. Research on algorithm of edge detection. An accurate algorithm for edge detection is presented based on gradient histogram of image. The edge strength and directions are obtained through the Kirsch operators, and the thresholds are calculated based on the statistic characters of gradient histogram. The algorithm is simple and flexible. Experiments show that the algorithm can detect edges with single-pixel width efficiently without calculating gradient directions.3. On the basis of the detected edges, making full use of the detected edges to enhance the images. The results show that the algorithm can enhance the image edges effectively and have a satisfactory effect.