Study Of The Path-Following Phase-Unwrapping Algorithm

Phase unwrapping is a key technique in data processing of Synthetic aperture radar interferometry (InSAR) and its accuracy directly impacts on the precision of digital elevation models (DEM ) and deformation monitoring, so it is still a hotspot in the research of InSAR. In the paper achievements and problems of current phase unwrapping algorithms are analyzed and summarized, whereas main study is laid on the advantages of branch-cut phase unwrapping algorithm, quality-guided path following algorithm and mask cut algorithm. Therefore, Goldstein's phase unwrapping algorithm is studied in two fields of the linking strategy of branch-cuts and introduction of quality maps in guiding the branch-cuts' placement to solve the problems that there are a large number of redundancies in branch-cuts and quality maps are not considered in placement of the branch-cuts.1. Goldstein's branch-cuts are optimized by using minimum spanning tree, as a result, the redundancies in the branch-cuts are effectively reduced, cycles and cross branch-cuts are eliminated. It is demonstrated that solutions are improved evidently.2. The shortest path based on degrees is put forward. The new algorithm makes the length of the branch-cuts shorter than Prim algorithm and operation faster.3. Quality information of interference pattern is applied to Goldstein's phase unwrapping algorithm. Firstly, Goldstein's algorithm is executed after bad quality pixels are masked by mask-cut. Secondly, a quality-map is separated into regions of different quality value according to thresholds and then weighted branch-cuts are optimized.Experiments prove that the length of branch-cuts and magnitude of discontinuities are remarkably reduced and unwrapped results are evidently improved with branch-cuts optimized.