Exploring the Effectiveness of Sobel, Canny, and Prewitt Edge Detection Algorithms on Digital Images

Edge detection is a fundamental process in image processing, crucial for identifying object boundaries and structural features within images. This study explores three classical edge detection techniques - Canny, Sobel, and Prewitt. Six test images were used to ascertain their performance based on five metrics: Recall, Precision, F1-Score, Structural Similarity Index (SSIM), and Figure of Merit (FoM) implemented using python. The experimental results indicate that the Canny operator consistently outperforms the others in terms of Recall, F1-Score, and FoM, demonstrating superior capability in detecting true edges with high sensitivity and robustness against noise. The Sobel operator achieves the highest Precision and SSIM scores, reflecting strong edge localization and structural preservation, although with lower overall edge detection effectiveness. The Prewitt operator offers balanced performance across all metrics, providing a compromise between detection quality and computational simplicity. These findings are consistent with general observations from the literature, where Sobel is recognized for its noise resistance and simplicity, making it suitable for fast, real-time applications, while Prewitt, offering a similarly straightforward implementation, exhibits slightly greater sensitivity to noise. The Canny operator, widely regarded as the optimal edge detector, remains the preferred method for applications requiring high precision, low error rates, and strong edge continuity. Consequently, Canny is best suited for high-accuracy edge detection tasks, Sobel excels in structure-preserving applications, and Prewitt is recommended for general-purpose, resource-constrained scenarios.