Comparison of Image Enhancement Algorithms Based on Denoising and Edge Detection

Abstract

Image enhancement technology is crucial in many fields, including noise reduction and edge enhancement. This paper compares median filtering and Gaussian filtering for image denoising by practicing. Median filtering removes salt-and-pepper noise well and preserves edges, suitable for discrete noise and detail-critical scenarios like remote sensing and medical imaging. Gaussian filtering smoothes Gaussian and random noise but blurs edges, applicable for background noise reduction. For edge enhancement, Sobel and Laplace operators are contrasted. Sobel operator highlights directional edges, useful in dynamic and noisy scenarios like surveillance. Laplace operator globally sharpens multi-directional edges, sensitive to details but noise, often used in medical image boundary extraction. Their combination improves edge detection. This research provides a solid theoretical foundation for optimizing the application of image enhancement techniques, offering valuable insights and guidance for selecting appropriate image processing algorithms in diverse fields such as medical image processing, autonomous driving, and other advanced technological domains. Future studies could focus on adaptive denoising algorithms to meet diverse image processing needs.