Development of an Inversion Code, Icare, Able to Extract Urban Areas Ground Reflectances

Automatic mapping of urban materials from remotely sensed radiance images remains difficult because of the complex physical phenomena induced by relief. Indeed, shadows and environment effects disrupt the radiance incoming the sensor. Moreover, the measured radiance depends on the illumination conditions of the observed area. This is quite problematic for multi-temporal analysis. Images are often taken in different illumination conditions thus shadows are not located in the same place. This paper describes a new physical model, ICARE, able to solve the radiative transfer inversion problem over urban areas, in the reflective domain (0.4 μm – 2.5 μm), from high spatial and spectral resolution images. This new approach takes into account the complex relief of 3D structures, the spatial heterogeneity of the scene and atmospheric effects, in order to extract rigorously the ground surface reflectance, even in shadows areas. The resolution method consists in modelling separately the irradiance and radiance components at ground and sensor levels. The validation of ICARE is checked through the CAPITOUL trial, carried out in Toulouse (France, 2004). Measurements were performed using two airborne Pelican image systems consisting in 8 high spatial (20 cm) and spectral (30 nm) resolution cameras. Results are presented over typical urban structures in Toulouse center completed by an error budget. They demonstrate that extracted reflectances correspond very well to ground reflectance measurements. Moreover, there is a good continuity between reflectance obtained over sunny and shadowed materials. That proves that shadows areas and environment effects are well corrected. To quantify the gain brought by ICARE, a comparison between classification based on radiance and reflectance images, obtained by inversion, is introduced.