Hyperspectral and IFSAR data for 3D urban characterization

AVIRIS and AIRSAR data over a urban environment are evaluated. The 2D maps available from hyperspectral classification are used to improve the detection and recognition of built zones starting from the IFSAR DEM. The improvements are twofold: best footprint detection provides a way to better reconstruct the 3D profile of individual buildings. Moreover, different terrain covers (e.g. trees) are recognized and discarded. Examples are provided over the UCLA campus in Los Angeles.