Anisotropic Diffusion based Impulse Noise Removal for Remote Sensing Images

In image processing and computer vision, image denoising is a crucial challenge that should be rectified by suppressing the noise-corrupted image and obtaining the image information. The random variation of brightness or colour information in acquired images is referred to as image noise. Image denoising is also useful in a variety of applications, such as image restoration, visual tracking, image registration, picture segmentation, and image classification, where recapturing the original image content is critical to achieving good results. To deal with additive noise, a myriad of image denoising methodologies have been proposed in recent times. Impulse noise, on the other hand, remains a challenging problem to solve using multiple ways. It is a sort of noise with either black or white noise pixels. We propose a novel concept of scale-space in this study, as well as a class of algorithms that implement it via a diffusion process. The primary purpose is to eliminate salt and pepper noise from remote sensing imagery using an anisotropic diffusion median filter. Our method ensures that region boundaries are kept as precise as possible. The findings of the experiments are depicted in a series of images. In terms of visual outcomes and performance metrics, the performance of the algorithm is validated by Structural Similarity Index Metric (SSIM) and Peak Signal to Noise Ratio (PSNR)