Hyperspectral remote sensing for mapping and detection of Egyptian kaolin quality

This study aims to use spectral analysis and hyperspectral Hyperion remote sensing images of the Egyptian Abu Zenima Carboniferous and Cretaceous kaolin deposits, located in West-Central Sinai Peninsula, for mapping the spatial distribution of their qualities determined by their mineralogical and geochemical parameters. Mineral quantification has been made by X-ray diffractometry and chemical analysis, and kaolinite structural order-disorder degree was measured by mean of the Hinckley (HI), Stoch (IK) and Liètard (R2) indices. The geochemical characteristics classified the studied samples into ferruginous and non-ferruginous deposits. The mineralogical composition discriminated the studied sample grades into kaolin (>75% kaolinite), silty kaolin (75-50% kaolinite) and kaolinitic siltstone (<50% kaolinite). Abu Zenima kaolin quality is mainly influenced by quartz and Fe-Ti minerals contents. The Carboniferous kaolin samples are characterized by ordered kaolinite (HI or R2 > 1 and IK < 0.7), while the Cretaceous kaolins exhibited mainly disordered kaolinite (HI or R2 < 1 and IK > 0.7). Five EO-1 Hyperion Level 1 GST radiometrically and geometrically corrected images of April and May 2011 were used to identify the spatial distribution of Carboniferous and Cretaceous kaolin grades and the structural characteristics of kaolinite. The position, depth, full-width-half-maximum and 22SP-Index of the absorption features were calculated for the continuum-removed spectra in the range 350-2500 nm. Prominent absorption features spectra occur around ~1400, ~1900, ~2200 and 2300 nm, and there are shifting and changes in their position and morphology with the kaolinite richness and the structural order-disorder degree. Spectral Angle Mapper (SAM) supervised classifications proved successfulness for identifying the kaolinite spatial distribution on the Hyperion images using the measured spectra of kaolins with different qualities and structural order-disorders.