干旱区生态环境知识资源中心

About twenty Saharian desert regions have been selected a few years ago in order to carry out in-flight calibration of the different instruments operating in the visible and near-infrared spectral domain. Since then, CNES has collected an important number of measurements acquired by these instruments of interest (SPOT, AVHRR, SeaWiFS, Polder, Vegetation, MODIS, MISR) over the selected desert areas (SADE database).

The present work fits into a global assimilation approach which aims to improve both the characterization of the calibration sites and the cross-calibration of optical satellite sensors. This work is particularly devoted to the spectral characterization of the selected site using the SADE database. The method is based on the use of a spectral model of ground surface reflectance at global scale. It is assumed that this model can be derived from laboratory reflectance measurement (i.e. "small scale" measurement). Then, instead of reversing the top of atmosphere measurement into ground reflectance, the ground reflectance model is transported at the top of atmosphere for comparison to available measurement, and the parameters adjustment is done at this level. A top of atmosphere simulated reflectance dataset (corresponding to various usual multispectral sensors) is used in a first step to assess for the relevancy of the proposed method.