IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing (Jan 2023)
An Efficient Polarimetric Persistent Scatterer Interferometry Algorithm for Dual-Pol Sentinel-1 Data
Abstract
With time-series PolSAR images, polarimetric persistent scatterer interferometry (PolPSI) algorithms can obtain optimized interferograms with overall better phase quality than any single-pol channel for ground deformation monitoring. Moreover, the open access of Sentinel-1 PolSAR data makes it possible for applications of PolPSI over large regions worldwide. However, the optimum scattering mechanism usually has to be searched in a high dimension solution space by PolPSI techniques from each pixel. This is with very high or even unacceptable computational costs if satisfactory results are expected. To this end, an efficient and effective PolPSI algorithm named as TP-ESM is proposed with Sentinel-1 data in this study, which optimizes pixels' interferograms by a weighted sum of their corresponding VV and VH channel interferograms. The effectiveness of TP-ESM is tested together with two other PolPSI techniques [i.e., TP-MSM and exhaustive search polarimetric optimization (ESPO)] over Beijing with 46 Sentinel-1 PolSAR images. The results show that TP-ESM can obtain similar optimized interferometric phases with ESPO over high quality pixels, and the ground deformation monitoring pixel density improvement achieved by TP-ESM and ESPO w.r.t. conventional PSI approach (with VV data) is 35% and 43%, respectively. On the other hand, the TP-MSM approach is found not applicable to Sentinel-1 data. Considering the negligible computational cost of TP-ESM w.r.t. ESPO, it presents a quite good performance on both interferograms' optimizations and ground deformation monitoring. Moreover, the proposed TP-ESM outperforms ESPO on distributed scatterers pxiels' optimization, and it is anticipated to have promising performances on other PolSAR images acquired by other sensors.
Keywords
