Please login first

PolInSAR coherence based decomposition for scattering characterization of urban area
Awinash Singh * 1 , Shashi Kumar 2 , Kavita Jhajharia 1
1  Manipal University Jaipur, India
2  Indian Institute of Remote Sensing (IIRS), ISRO, Dehradun


Polarimetric SAR data based scattering retrieval has been widely used to characterize manmade and natural features. It has been found that PolSAR data has the capability to retrieve scattering information contributed by different features within a small area or single resolution cell. Generally, it has been found that the urban structures are contributing the high double-bounce scattering, but due to closely spaced urban structure, multiple reflections of the SAR waves from the walls of the buildings give the appearance of the volume scattering. The overestimation of volume scattering from urban structure could be reduced by the adoption of interferometric coherence in decomposition modeling. The PolInSAR coherence constitutes the full collection of polarimetric and interferometric information. The urban buildings are considered as permanent scatterers which is usually not affected by the temporal and volume decorrelation. Therefore, they show high coherence magnitudes. The prime focus of this research was the implementation of PolInSAR coherence in the decomposition modeling to minimize the overestimation of volume scattering from the urban structure. This study has used the CoSSC product of the TanDEM-X mission. The PolInSAR data over Dehradun, India were acquired in bistatic mode. All the possible combinations of PolInSAR coherence were generated from TerraSAR-X and TanDEM-X. The Pauli basis based PolInSAR coherence has shown the capability to distinguish different features according to their nature. To find the appropriate coherence for decomposition modeling the optimization was performed on PolInSAR data to select the optimal coherence. The results obtained from PolInSAR coherence based decomposition modeling had shown the dominance of double bounce scattering in the urban area for closely spaced structures also. The study strongly recommends the use of PolInSAR coherence in the decomposition modelling to minimize the ambiguity in the scattering retrieval from an urban area due to close spaced buildings.

Keywords: Polarimetric SAR; PolInSAR; CoSSC, TanDEM-X, Urban structure