LiDAR-GUIDED REDUCTION OF SPECTRAL VARIABILITY IN HYPERSPECTRAL IMAGERY

Hyperspectral unmixing has attained a great importance in recent decades in remote sensing applications. Due to some external effect (illumination conditions) or internal effects (concentration of chlorophyll), spectral variability exists in hyperspectral images. This spectral variability causes significant errors in abundance estimates. In this paper, we propose a new framework that incorporates feature extraction with Digital Surface Model (DSM) clustering informationto suppress the effect of spectral variability in hyperspectral unmixing. In this way, meaningful material abundance estimates are obtained. Experiments are conducted on simulated data. Results show that spectral variability can be reduced with the aid of LiDAR data in hyperspectal unmixing.