Research of Image Denoising Method Based on Part Adaptive Total Variation and Median Filter

According to the advantage of median filter and total variation algorithm, a new median filter adaptive total variation model based on part gradient of image pixels is established in the paper. At first, the new model uses the gradient information of every pixel of the image to find out the isolated noise point and then uses the median filter to remove it. Second, the adaptive total variation algorithm to deal with the new noise image is adopted. The new method proposed in the paper not only surmounts the disadvantage of generating artificial edge but also has the advantages of denoising and edges preservation of TV model. The experiments show that the PSNR of the improved method increases almost 1.0dB compared with the adaptive model at the same noisy level.