Research On Fuzzy-Threshold-Based Segmentation Method Of Images

Image segmentation is a process, which can extract targets or interesting areas from input image. It is also an important step for detecting and identifying objects. There are various kinds of methods, which can implement image segmentation, and the threshold method is a simple and practical one. In this dissertation, we investigate threshold method and fuzzy theory, and combine them as a whole to improve the accuracy of image segmentation. The works of the dissertation are summarized as follows:1) The development of threshold method in recent years is analyzed. And some classical methods, such as OTSU method, two-dimension OTSU self-adaptive thresholding segmentation algorithm, two-dimension entropy method, etc. are discussed. Besides, we make some improvements for the first two methods.2) The combination of fuzzy theory and threshold method is discussed. And the threshold method based on ambiguity, fuzzy-set method based on interval estimation, fuzzy-set method based on genetic algorithm are expatiated and implemented.3) After studying on the self-adaptive fuzzy threshold segmentation method, we give a method about self-adaptive selection of window-width to improve the correctness of threshold selection. The dissertation obtains a new histogram transformation method by combining two existing histogram transformation methods, which can improve efficiency of fuzzy threshold segmentation method.4) Aiming at segmentation of image with small targets, the dissertation puts forward an algorithm of approaching segmentation. According to the guide rule of the gray-level statistics of the target and background in the image, the algorithm segments the gray-level statistic histogram several times to get the best threshold.