Denoising method for image with mixed noises

The invention discloses a denoising method for an image with mixed noises, which comprises the following steps of: performing wavelet packet decomposition on the noise-including image; converting the image subjected to the wavelet packet decomposition to a wavelet domain, and dividing the image into a low-frequency sub-band image and a high-frequency sub-band image; filtering the low-frequency sub-band image adopting a neighborhood averaging method; and firstly detecting satisfied special points in the noise of the high-frequency sub-band image by using a cluster analysis method, filtering to remove pulse noise by adopting median filter, and denoising white Gaussian noise on the image by adopting threshold processing in a wavelet contraction step according to the equivalence of a Gaussian curvature high-order diffusion of and wavelet contraction method. When the denoising method is adopted, the mixed noises in the image can be removed, and the double functions of high-frequency characteristics and edge shapes can be well kept, thus a better effect is kept.