A comparative analysis between an hybrid MUSIC-LS algorithm and the standard MUSIC for volume height estimation in tomography-SAR

This paper deals with an application of Multi-baseline-SAR (Synthetic Aperture Radar) that estimates the biomass in forested area. Tomography is a technique that acquires N images of the study area at different altitudes to retrieve the backscattering signal parameters. This work focuses on inversion algorithms applied to SAR signal. The proposed procedure is based on the fusion of MUSIC (Multiple Signal Classification) and LS (Least Squares) inversions. We used the MUSIC inversion as a reference for the evaluation of the algorithm while the LIDAR height was exploited as ground truth for the inversion. Results of the inversion are validated qualitatively and quantitatively by comparing between the estimated volume backscatter signal (MUSIC and MUSIC-LS) and the DSM (Digital Surface Model) obtained from LIDAR (Laser Detection and Ranging) data.