Fast phase denoising using stationary wavelet transform in speckle pattern interferometry

Phase fringe patterns in speckle pattern interferometry need to be denoised before they are processed further to ensure phase unwrapping reliability and phase determination accuracy. A fast phase denoising method combining the sine–cosine transform and the stationary wavelet transform, with a detailed, unified and adaptive parameter setting of the wavelet function, decomposition level, threshold selection and threshold function, is proposed for a speckle interference image. Owing to the ability of the sine–cosine transform to retain an abrupt phase change, and the stationary wavelet transform to quickly remove the phase noise, the proposed approach can be used to denoise the phase fringe pattern effectively and quickly. These characteristics are verified by simulation as well as practical experiments. Compared with the popular phase denoising method—the windowed Fourier transform with a window size of 20—the proposed method can save the running time by at least 16 times in a practical application test. This speed advantage will be more important when a high-resolution camera is used in dynamic measurement.