The image is transformed from space domain to wavelet domain by using wavelet transform, and then the wavelet coefficients are processed in the wavelet domain to reduce noise, and finally the denoising image is obtained by inverse wavelet transform. In the wavelet domain, the larger coefficients in the high-frequency sub-band contain important image information, and the number is small, and a large number of smaller coefficients mainly contain noise, and the distribution is relatively uniform. Comparing the wavelet coefficients with one person or multiple thresholds, setting the wavelet coefficients to zero, maintaining or shrinking according to the comparison results can reduce low noise and preserve image details.