In this thesis, we establish a new iterated denoising algorithm by analyzing Partial Differential Equation (PDE) techniques in image processing. Our new algorithm smooths noisy image with two phases. In the first stage, new algorithm substitutes iterated scheme based on PDE with adaptive Wiener filtering in the wavelet domain and gains good effect by adjusting related parameters. The simulates show the parameters in one test still is appropriate for other some denoising images,which is superior to PDE techniques. In second stage, we use PDE to reduce edge artifacts generated in the first stage. Numerical experiments show that new algorithm improves the reconstructed image quality in denoising.
|