Wavelet analysis is a new method to study image restoration, but it has been widely concerned. In recent years, wavelet restoration has become a hot research direction at home and abroad. The theory of multi-resolution has been widely studied, which not only makes the classical methods such as noise suppression, camera defocusing, CTLs (constra ined to ta l least squares) regularization, adaptive and so on can be better solved, but also is very convenient and flexible in dealing with non-stationary problems and protecting image edge information. Charles et al. Used wavelet to replace the gradient operator in the energy density function, not only improved the texture extraction effect, but also eliminated the redundant iterative steps, which all benefited from the multi-resolution structure characteristics of wavelet. Stephanak is et al. Used 2-D separable wavelet instead of traditional smoothing operator to improve the smoothing effect. Cao et al. Use wavelet analysis in image restoration, can feedback image parameters in time, and further guide the restoration operation. Premaratne et al. Proposed a method of wavelet decomposition, which can automatically determine whether the restoration is completed. Liu et al. Used nonparametric estimation theory to limit the regularization process, and then used multi-resolution theory to solve the problem. This estimation algorithm is obtained in wavelet domain and belongs to the complex regularization method.