A review of enhanced total variation approaches

Abstract

This study focuses on the total variation model for denoising images, which are important for maintaining the resolution and quality of images used in environmental, agricultural, and geographical analysis. Although traditional denoising methods, such as median and Gaussian filters, are effective, these methods usually result in the loss of basic image details such as edges, topography, and resolution. These are qualities of great importance. However, the total variation model encounters the challenge of non-differentiability, an important process in denoising processing. The study provides a comprehensive overview of the total variation model and the improvements made to it. It also compares it with traditional methods, demonstrating its ability to offer clearer images while preserving important details.