An optimal color image edge detection approach

Abstract

Detecting edges in color image is currently a very emerging topic in pattern recognition research. In this paper, an optimal approach for color image edge detection is introduced. The proposed technique is based on canny edge detector. The color computations are done in HSV color space. Here, the deficiencies of the traditional Canny edge detector are removed by replacing the Gaussian filter with an adaptive median filter and a type-2 fuzzy set-based technique is used for double threshold calculation in the final step of edge detection. The experimental results show the superiority of the proposed approach.