A new fuzzy edge detection method for image enhancement

A new method for detecting object edges, called the fuzzy Sobel method, is proposed. The fuzzy logic methodology is applied to extract the feature value for an image and a Sugeno-type fuzzy reasoning strategy is adopted for edge enhancement. Such a method can improve the drawbacks of conventional approaches, for instance, the Prewitt and Sobel methods. In the two traditional methods, they use fixed parameters for all kinds of images so that they might have success in one image but fail in another one. The fuzzy Sobel method, on the other hand, can find four threshold values automatically and then use them to construct membership functions for each detected image. Therefore, the fuzzy Sobel method can determine the parameters adaptively for different images. The edges detected in the Sobel method are often vague, but the fuzzy Sobel method can enhance edges and makes the resultant image clear.