Computer Science · Computer Vision and Pattern Recognition
Filtering-based phase unwrapping
Qian Kemao, Wenjing Gao, Haixia Wang
See on the mapAbstract
Phase unwrapping is a usually necessary step in optical interferometry in order to construct a continuous phase distribution. A straightforward strategy is to filter the noise effectively, followed by a simple phase unwrapping process. A windowed Fourier filter is proposed to realize this strategy, which has the following properties: (i) it removes the noise effectively; (ii) it preserves the phase near edges; (iii) it identifies invalid areas; (iv) the amplitude of filtered exponential phase field can be used as a quality map for quality-guided phase unwrapping; (v) the filtering result is insensitive to the threshold; (vi) the window size in the filter can be switched according to different requirement. A large window size is useful for noise filtering while a small window size is useful for protecting signal details.