Edge detection based on the Shannon Entropy by piecewise thresholding on remote sensing images

Edge detection is one of the most important concepts used in processing of remote sensing images. The aim of edge detection is to mark the points of an image at which the rate of brightness changes sharply. Sharp changes in image features often represent important events and changes in environmental properties. In other words, edges can be defined as the boundary between two regions separated by two relatively distinct grey level properties. Most classic mathematical methods for edge detection are based on deriving original image pixels such as Laplacian gradient operator. In remote sensing images, because of the high variation rate, the edge detection operators may have some weaknesses in correct detection of the scope of complications. This study provides a novel approach for detecting the edges based on the features of remote sensing images. In this method, at first, thresholds of different regions of the image were determined in a piecewise manner; then, by using the proposed methods, appropriate thresholds were extracted, and finally, the boundary between these regions was extracted using Shannon entropy. The obtained results were compared with some standard algorithms and it was observed that the method was efficiently able to detect edges.