Dynamic Adaptive Gradient Operators for Noise-Resilient Edge Detection and Image Enhancement

Gradient-based edge operators are widely utilized in image edge processing but are highly sensitive to noise, often limiting their effectiveness. Traditional noise reduction techniques, while mitigating noise, frequently introduce image blurring, resulting in a loss of fine details. To address this issue, we propose a novel Dynamic Adaptive Gradient Operator algorithm. Central to this approach is a dynamic weighting mechanism, denoted as D, which adaptively adjusts the gradient response to suppress noise while preserving fine image details. This algorithm enhances the performance of classical edge operators, including Laplacian, Sobel, Prewitt, and Isotropic operators. Experimental results evaluated using Structural Similarity Index Measure (SSIM) and Peak Signal-to-Noise Ratio (PSNR) demonstrate that the enhanced operators achieve an average SSIM of 0.97 and a PSNR of 33.34, significantly outperforming their traditional counterparts. Notably, the improved Laplacian operator achieves a 16.19% increase in SSIM and a 1.75% increase in PSNR. Compared to conventional gradient operators, the proposed algorithm reduces distortion and noise while effectively preserving detailed image features, underscoring its potential for advancing image edge processing.