Dmitriev A. V., Chimitdorzhiev T. N., Dagurov P. N. Forest Anisotropy Assessment by Means of Spatial Variations Analysis of PolSAR Backscattering

Keywords:
radar imaging, polarization signature, fractal dimension, spatial variations
Pages:
19–27

Abstract

UDC 528.87+528.855

How to cite: Dmitriev A. V., Chimitdorzhiev T. N., Dagurov P. N. Forest anisotropy assessment by means of spatial variations analysis of PolSAR backscattering // Sibirskij Lesnoj Zurnal (Sib. J. For. Sci.). 2017. N. 3: 1927 (in English with Russian abstract).

DOI: 10.15372/SJFS20170302

© Dmitriev A. V., Chimitdorzhiev T. N., Dagurov P. N., 2017

The possibility to synthesize polarization response from earth covers at any desired combination of transmit and receive antenna polarizations is the significant advantage of polarimetric radar. It permits better identification of dominant scattering mechanisms especially when analyzing polarization signatures. These signatures depict more details of physical information from target backscattering in various polarization bases. However, polarization signatures cannot reveal spatial variations of the radar backscattering caused by volume heterogeneity of a target. This paper proposes a new approach for estimating volume target heterogeneity from polarimetric synthetic aperture radar (PolSAR) images. The approach is based on the analysis of a novel type of polarization signature, which we call fractal polarization signature (FPS). This signature is a result of polarization synthesis of initial fully polarimetric data and subsequent fractal analysis of synthesized images. It is displayed as a 3D plot and can be produced for each point in an image. It is shown that FPS describes backscattering variations or image roughness at different states of polarization. Fully polarimetric data of SIR-C and ALOS PALSAR at ascending/descending orbits were used for testing the proposed approach. The azimuthal dependence of the radar backscattering variations is discovered when analyzing backscattering from a pine forest. It correlates with the results of a field survey of trees branch distribution.


Return to list