Identifying industrial processes through VNIR-SWIR reflectance spectroscopy of their waste materials

Industries dump important amounts of wastes which must be characterized for estimating environmental quality and identifying industrial processes, which can they be related to some specific mineral combinations. Land disposal is an usual practice and allows for hyperspectral remote sensing characterization. Many studies have shown that spectroscopy is a powerful tool for mining waste minerals monitoring. Nevertheless, little has been done regarding the spectral features of industrial wastes. The aim of this study is to perform spectral measurements, in-situ and in the laboratory, of industrial wastes to extract their optical and spectral characteristics. This preliminary study will then be used for interpretation of hyperspectral images acquired over test sites. We focus here on two industrial sites because of their similarity in the visible and near infrared (VNIR) that can not be distinguished using color images only : (1) Neutralization Plant Seche-Ecoservice (France) and (2) Land disposal area of the Bauxite refinery Alteo-Environment. The spectral features of the spectrum are extracted using the first and the second derivative. With this preliminary study, we are able to conclude that imaging spectroscopy is an efficient tool for the characterization of such industrial wastes.