Edge Detection Algorithm for Digital Images Using Bird Swarm Algorithm and Latin Square Design Method

Abstract

Edge detection is a fundamental stage in digital image processing, where connected and homogeneous regions are identified based on grayscale levels. In this study, two approaches were employed for edge detection in VAR arbitration images: the Latin Square Design (LSD) method and the Bird Swarm Algorithm (BSA).

The results revealed variations from one image to another regarding edge identification. However, both methods demonstrated high efficiency in detecting edges across images of varying sizes and grayscale levels. The LSD method, which relies on statistical criteria, outperformed in terms of PSNR (image quality) and RMSE (root mean square error), achieving the highest PSNR values and the lowest RMSE values. This makes it a preferred benchmark for evaluation.