Two-dimensional phase unwrapping based on adaptive wavelet denoising

Due to the disturbances by noises and speckles in practical interferograms, accurate phase unwrapping has been a problem in interference measurement. Using the multi-scale and localization properties of wavelet analysis, a new phase unwrapping algorithm based on adaptive wavelet de-noising is proposed. In this algorithm, the original interferogram is decomposed into a smooth approximation and a series of wavelet planes. A resolution level is determined adaptively, and a threshold is set to smooth the wavelet planes. Then the noise-reduction interferograms are reconstructed based on the approximation and the smoothed wavelet planes, so that the normal unwrapping processing can be applied to obtain the phase recovery result. The algorithm is shown to be simple, effective, and convenient to use. In the meantime, it is sufficiently accurate for unwrapping recovery of continuous and smooth phases under the condition of serious noise contamination.