Improved Canny Edge Detection Technique Using S-Membership Function

Abstract

Traditional Canny edge detection algorithm is sensitive to noise, hence it may lose the weak edge information after noise removal and show poor adaptability of fixed parameters like threshold values. In view of these problems, this paper reports on the modification of canny edge detection algorithm using s-membership function. Adaptability of threshold values are achieved through S-membership function and is given as input to default Canny algorithm. The grayscale images have been analyzed for default Canny and modified Canny algorithm. To understand the performance of these algorithms it is essential to evaluate various statistical metrics. The proposed work states that the detailed statistical results and the images obtained reveal the superior performance of the modified Canny algorithm over the default Canny edge detection algorithm. Further the images obtained from modified Canny algorithm shows the marked edges with efficient image edge extraction and provide accurate information for image measurement.