In this paper, we present a knowledge-based circle object fuzzy multi-threshold edge detection method based on wavelet transform. By observing the characters of edges, background and noise in the wavelet transform image and analyzing the wavelet edge detection's principle, we combine the wavelet edge detection technique and the fuzzy logic technique together to detect the object's edge. A fuzzy operator which takes not only the wavelet's module magnitude into account, but also the wavelet gradient direction is designed to take edges out of the image. Then, a series of different fuzzy thresholds are set to complete the fuzzy enhancement task In the end, a knowledge-based criteria for the circle edge is applied to resolve the circle object edge and eliminate those edges that we do not want. A series of simulation experiments have been conducted by using the proposed method. We compare the performance of our method and the traditional fuzzy edge detection method. The experiment result shows that the proposed method works well for images and produces good results in comparison to other methods.
展开▼