A Comprehensive Analysis of Watershed Image Segmentation Based on Various Edge Detection Approaches

Edge detection is a critical problem in computer vision and image processing. The issue with picture segmentation has been quite concerning for the researchers. A spike in intensity from one pixel to the next can significantly alter the picture quality and image segmentation since edges in digital images are places with substantial intensity contrasts. The objective of image segmentation is to divide an image into useful parts with respect to a certain application. There are many edge detection operators available for digital picture object boundary extraction and image segmentation. Each operator is designed to have a particular type of edges sensitivity. Entropy, which is the main concerting operator, is among them, along with Sobel, Roberts, Prewitt, Laplacian of Gaussian (LoG), and Canny. In this study, the peak signal-to-noise ratio (PSNR), mean square ratio (MSE), and execution times are also explored along with entropy, a statistical measure of unpredictability that may be utilized to define the texture of the input picture. The goal of the current study is to evaluate the performance of several edge detection operators and compare their results. This is done for a picture using the Python programming language. A thorough study of numerous edge detection methods, including Sobel, Prewitt, Roberts, LoG, and Canny, is conducted in this work. Experimental findings show that the clever edge detector performs better than others.