A FUZZY DIFFERENCE BASED EDGE DETECTOR

In this paper, a new algorithm for edge detection based on fuzzyconcept is suggested. The proposed approach defines dynamic membershipfunctions for different groups of pixels in a 3 by 3 neighborhood of the centralpixel. Then, fuzzy distance and -cut theory are applied to detect the edgemap by following a simple heuristic thresholding rule to produce a thin edgeimage. A large number of experiments are employed to confirm the robustnessof the proposed algorithm. In the experiments different cases such as normalimages, images corrupted by Gaussian noise, and uneven lightening imagesare involved. The results obtained are compared with some famous algorithmssuch as Canny and Sobel operators, a competitive fuzzy edge detector, and astatistical based edge detector. The visual and quantitative comparisons showthe effectiveness of the proposed algorithm even for those images that werecorrupted by strong noise.