Skip to Main Content
Edge structures which are boundaries of object surfaces are essential image characteristic in computer vision and image processing. As a result, edge detection becomes part of the core feature extraction in many object recognition and digital image applications. This paper presents a new hybrid edge detector that combines the advantages of Prewitt, Sobel and optimized Canny edge detectors to perform edge detection while eliminating their limitations. The optimum Canny edges are obtained through varying the Gaussian filter standard deviation and the threshold value. Simulation results show that the proposed hybrid edge detection method is able to consistently and effectively produce better edge features even in noisy images. Compared to the other three edge detection techniques, the hybrid edge detector has demonstrated its superiority by returning specific edges with less noise.