
Polarimetric Analysis of ALOS PALSAR data (POL-SAR) over Test Areas in North-West Bulgaria – Polarimetric Descriptors, Decompositions and Classifications
Author(s) -
Zlatomir Dimitrov
Publication year - 2022
Publication title -
aerospace research in bulgaria
Language(s) - English
Resource type - Journals
eISSN - 2367-9522
pISSN - 1313-0927
DOI - 10.3897/arb.v34.e03
Subject(s) - polarimetry , remote sensing , synthetic aperture radar , scattering , radar , environmental science , l band , geography , computer science , physics , optics , telecommunications
This study is focused on utilizing full-polarimetric L-band radar data from ALOS PALSAR (JAXA) by means of Polarimetry (POL-SAR), over mountainous test sites in Bulgaria. General aim is to show feasibility of the Polarimetry to describe natural targets, which exhibits various scattering mechanisms in respect to their bio-physical and geometrical properties. Firstly, the importance of Covariance and Coherent matrices is shown, which is followed by calculation of the polarimetric coherences with their particular significance. The mathematical and physical model based decompositions are applied to describe backscattering media from scattering mechanisms. Radar indices resulted from H/A/α-decomposition showed radnomization of scattering mechanisms over forest areas, whilst two major scattering mechanisms are observed mainly in crop lands. Α comparison is made between polarimetric descriptors from acquisitions in different seasonality over mountainous forest and agricultural lands. Polarimetric segmentations and classifications are applied, with 8 (H/A) and 16 (H/A/α) components. Finally, a forest mask is proposed based on relevant polarimetric descriptors. Study showed good utilization and importance of the full-polarimetric L-band SAR data, derived from ALOS PALSAR, in natural targets and forest areas. This report resulted from a course GEO414 -“Polarimetrie”, held at the University of Jena, Lehrstuhl für Fernerkundung, in the framework of ERASMUS+, with the kind support of – Dr. T. Jagdhuber (DLR) and Prof. C. Schmullius.