Comparison of Edge Detection Methods Using Road Images

Abstract

Edge detection is a method in image processing that provides valuable information about images. This paper focuses on edge detection in the context of infrastructure, specifically on asphalt roads. It compares the Canny, Prewitt, Sobel, Roberts, and Laplacian of Gaussian algorithms as image processing techniques. Each algorithm yields distinct results, which are evaluated by comparing the processed images to the original images. The assessment utilizes Mean Squared Error (MSE), Peak Signal to Noise Ratio (PSNR), and Structural Similarity Index Measure (SSIM) to measure the algorithms' performance. By employing road images as data for processing, this study aims to identify the algorithm that produces the clearest edges in road images. The experimental results indicate that the Roberts algorithm demonstrates superior accuracy, achieving MSE values of 0.176, PSNR of 7.5, and SSIM of 0.001.