A Cluster-Analysis and Convex Hull-Based Fast Large-Scale Phase Unwrapping Method for Single- and Multibaseline SAR Interferograms

Abstract

For synthetic aperture radar (SAR) interferometry (InSAR), phase unwrapping (PU) is an important and difficult step. Due to the high computational complexities of the classical and skilled PU methods, the size and number of interferograms to be processed together must be considered in InSAR applications. However, with the rapid development of InSAR technology, the scale of interferometric data has significantly increased, which has brought the following two difficulties to InSAR processing: excessive memory consumption and unacceptably long computing time. To solve these problems, a fast large-scale PU method based on cluster analysis and convex hull (CCFLS) is proposed in this article. It was developed for single-baseline InSAR and refined under the multibaseline two-stage programming approach InSAR framework. The main idea of CCFLS is to reduce the consumption of computing resources by discarding the PU processing of worthless low quality areas. The key to this work is to determine the discarded region at a low computational cost while ensuring the same PU accuracy for the remaining region as global processing. The theoretical analysis demonstrates the advantage of the CCFLS method for large-scale InSAR data processing, and experiments also verify that CCFLS can greatly reduce the consumption of computing resources while ensuring the PU accuracy of the solved region.

Keywords

• Large-scale (LS)
• multibaseline (MB)
• phase unwrapping (PU)
• single -baseline (SB)
• synthetic aperture radar (SAR) interferometry (InSAR)