Influence of Salt and Pepper Noise on the Edge Detection of Images Using Modified Canny Edge Detector with S-Membership Function

Abstract

Edges in a digital image provide important information about the objects contained within the image since they constitute boundaries between objects in the image. The extraction of these features can be further used

for real time purposes like face recognition, computer vision algorithms etc. But it is somewhat difficult to extract

out all the edges efficiently without affecting the structural properties of an image. Edges in an image represent a swift change in the intensity even in the presence of noise. It is essential to study the edge detecting properties of an image when the noise in the image is abundant. Edge detection in noisy images is in agreement between denoising and edge preserving capability. Various smoothing filters are appropriately studied in the viewpoint of edge detection. This paper describes the influence of various intensity levels of noise in the edge detection of images and detailed statistical metrics for understanding the efficiency of the proposed method. In this work, two images such as butterfly and fruit are considered and salt and pepper noise is added to these images at various levels of intensity (5% to 50%). In the next stage a median filter is utilized to denoise the images and the edges are analyzed with the help of modified canny method using S-membership function. The statistical metrics of shows that the edges of the resultant image is strictly preserved till 25% of noise intensity levels. The severe degradation of edges is observed in the image for the noise intensity levels more than 25% of salt & pepper noise.