Automated processing of planetary hyperspectral datasets for the extraction of weak mineral signatures and applications to CRISM observations of hydrated silicates on Mars

Abstract

Near infrared imaging spectrometers are key tools to investigate planetary surfaces in the Solar System. By coupling spectral and spatial information, they give access to the composition and morphology of the planets' surfaces which in turn provide insight into the geological state and history of the body. Processing and interpreting their datasets is however challenging owing to the very large amount of data they produce, a small subset of which contain relevant information, but also to numerous sources of errors, due to the instruments themselves or to observational biases, which further complicate the extraction of interesting but subtle spectral features. Collectively, these limitations have motivated the development of a set of tools that tackle these issues to facilitate the extraction of mineralogical information. The tools described here are successfully applied to the CRISM imaging spectrometer orbiting Mars in the search for hydrated silicates. An automated extraction of the hydrated silica signatures is performed at high accuracy and the discovery of a new mineral on Mars, epidote, is reported thanks to these new data reduction and analysis strategies.