Image denoising method can be divided into the spatial and frequency domain processing methods from different processing domains. The former is to deal with it in the two-dimensional space of the image itself. According to the different nature, it can be divided into linear processing method and nonlinear processing method. The latter is used to approximate the original signal function with a set of orthogonal function. The linear filtering algorithm in spatial domain is more mature, the digital analysis is simple, and the random noise effect is obvious, but it has obvious defects, such as the prior statistical knowledge of random noise, and the poor performance of the image edge details.