EXPLORING THE POTENTIAL OF POLARIMETRIC AND INTERFEROMETRIC SAR VARIABLES FOR DEFORESTATION MAPPING

Abstract

Synthetic Aperture Radar (SAR) serves as an important data source for continuous monitoring, especially across the tropics. The structural and dielectric properties of targets can be studied with the help of the wide range of SAR derivatives. This study aims at testing and understanding the potential of polarimetric and interferometric SAR variables for deforestation mapping across tropical and sub-tropical areas. Statistical change detection method Cumulative Sums of Change is used in this study for understanding the potential of backscatter and coherence derived from C-band Sentinel-1 SAR data for deforestation monitoring in Haldwani Forest Range in India. A rule-based decision making approach is applied on the change map generated with backscatter using the coherence to reduce the number of false positives due to shorter wavelength. With backscatter-only approach, overall accuracy 0.58 and kappa coefficient of 0.23 were obtained with VH polarization. However, with coherence-based compensation approach an improvement of 39.6% (AO=0.81) and 120.8% (κ=0.53) was observed in the overall accuracy and kappa coefficient. © 2025 IEEE.