Material reflectance retrieval in urban tree shadows with physics-based empirical atmospheric correction

Material reflectance retrieval from high spatial resolution acquisitions over urban areas requires an accurate modeling of the signal accounting for the 3D environment. Especially in tree shadows, the solar radiation incident to the ground contributing to the estimation of the reflectance has many origins linked to the plant structure and its composition. In this paper, the 3D atmospheric correction code, ICARE, limited to opaque structures like buildings, is improved thanks to an empirical correction factor taking into account the porosity of a tree crown. The validation of this method is assessed through a dataset combining a hyperspectral image and a 3D model of the scene.