本文提出了一种新颖的干涉条纹图非线性滤波方法，可以选择性地在频域图像的不同区域选用不同的滤波方式，在滤掉图像中大部分散斑噪声的同时，能够减少信息的损失，使图像的内部边界仍然保持清晰。文中首先简介了正交小波变换的原理，然后介绍了去除散斑噪声的具体算法，最后给出了计算机模拟去噪声的结果 This paper presents a novel non-linear filter fringe pattern filtering method, which can selectively use different filtering methods in different regions of the frequency domain image selectively, and can reduce the information loss while filtering out most speckle noise in the image , So that the image’s internal boundaries remain clear. The paper first introduces the principle of orthogonal wavelet transform, and then introduces the specific algorithm to remove speckle noise, and finally gives the results of computer simulation of noise