Research On The Technique Of Edge Detection Based On Fuzzy Logic

Edge detection is an important part of image processing field,which is the fundamental basis of image analysis such as image segmentation, extraction of texture and shape feature, etc. In image processing and analysis field, because of the fuzziness that the image edge itself has, obvious superiority is showed by applying the fuzzy theory to edge detection of image. Therefore, the technique of edge detection which is based on fuzzy theory is attached importance to by more and more researchers. In the thesis, the general theory of fuzzy logic is firstly introduced, and then some classical edge detection methods are studied, based on which, the relative algorithms about edge detection with fuzzy theory are made a research. The main research work in the thesis includes:(1) Edge detection of gray image with fuzzy theory is studied .On the one hand, considering the weakness of Pal and King algorithm of fuzzy edge detection, some improvement measures are proposed. The superiority of the improved algorithm is showed by the final edge detection results and the comparison on time costs. On the other hand, analyzing the relevance of pixels, an algorithm about fuzzy similarity is introduced. Fuzzy logic is applied from different angles by the two algorithms that are mentioned above. Seeing from the experimental results, the extracted edges with the two fuzzy algorithms are both better than traditional measures, which show the good performances of fuzzy edge detection algorithm.(2) Based on the mathematical morphology and fuzzy theory, a hybrid edge detection method based on morphology filter and gradient fuzzy GHMT(Gray Hit/Miss Transform) is proposed. Fuzzy logic is combined with gradient GHMT detector in the gradient fuzzy GHMT detector, which provided a powerful tool to process the uncertain information of image edges, but in the noisy circumstance, the detected edges are not satisfying. Considering this disadvantage, generalized morphological filters with multiple structuring elements integrating with gradient fuzzy GHMT detector is proposed, which improve the anti-noise capacity,and get better edges in noisy circumstance. The experimental results show that better performances can be achieved by using this hybrid method, especially detecting edges of noisy images.(3) Fuzzy logic is extended to the edge detection of color image. Color image is...