Distortion Reduction in Singularity-Spreading Phase Unwrapping With Pseudo-Continuous Spreading and Self-Clustering Active Localization

Singularity-spreading phase unwrapping (SSPU) realizes low calculation-cost phase unwrapping by utilizing its nature completely free from the combinatorial problem of pairing singular points (SPs), which is unavoidable in conventional branch-cut methods. However, the SSPU sometimes suffers from low-frequency global distortion. It occurs when the singularity is spread to excessively large area. To solve this problem, in this paper, we propose two methods, namely, fully isotropic SSPU and nonhollow SSPU, both of which realize pseudo-continuous spreading. By extending these methods, we also propose self-clustering active localization. Experiments demonstrate successful reduction of the low-frequency distortion, showing about 2-3 dB improvement of signal-to-noise ratio (SNR) in generated digital elevation model (DEM) with only 1/4-1/3 calculation time of that in conventional SSPU. The performance is totally discussed among a branch-cut method, conventional SSPU method, and proposed SSPU methods.