Monitoring Land Subsidence in a Rural Area Using a Combination of ADInSAR and Polarimetric Coherence Optimization

This paper investigates a combination of advanced differential synthetic aperture radar interferometry (ADInSAR) with different coherence optimization methods. After the launch of satellites with polarimetry capabilities, differential synthetic aperture radar interferometry (DInSAR) is feasible to generate polarimetric DInSAR to enhance pixel phase quality and increase coherent pixel (CP) density. The first method proposed in this paper, modified coherence set-based polarimetry optimization (MCPO), is a modification of a known single-baseline coherence optimization method to optimize coherence of all interferograms simultaneously. The second method, coherence-set based polarimetry optimization (CPO), was presented by Neumann et al., and is an existing revision of the single-baseline coherence optimization technique. The final method, exhaustive search polarimetry optimization, is a search-based approach to find the optimized scattering mechanism introduced by Navarro-Sanchez et al. The case study is the Tehran basin located in the North of Iran, which suffers from a high-rate of land subsidence and is covered by agricultural fields. Usually such an area would significantly decorrelate but applying polarimetric ADInSAR allows us to obtain a more CP coverage. A set of dual polarization TerraSAR-X images with 9 × 9 and 15 × 15 as multilook factors were used within the polarimetric ADInSAR procedure. All three coherence optimization methods with two different multilook factors are shown to have increased the density and phase quality of CPs. Moreover, the estimated deformation rates were evaluated using available levelling measurements. MCPO, which is presented in this paper, works more successful than CPO in terms of CPs density, phase quality and deformation accuracy.