Processing and analyzing advanced hyperspectral imagery data

Abstract

The main objective of the current work is to recognize the dominant and predominant clay minerals of South Port Said plain soils, Egypt using the high advanced remote sensing techniques of hyperspectral data. Spectral analyses as one of the most advanced remote sensing techniques were used for the aforementioned purpose. Different spectral processes have been used to execute the prospective spectral analyses. These processes include 1-The reflectance calibration of hyperspectral data belonging to the studied area, 2- Using the minimum noise fraction (MNF) transformation. 3 -Creating the pixel purity index (PPI) which used as a mean of finding the most spectrally pure, extreme, pixel in hyperspectral images. Making conjunction between the Minimum Noise Fraction Transform (MNF) and Pixel Purity Index (PPI) tools through 3-D visualization offered capabilities to locate, identify, and cluster the purest pixels and most extreme spectral responses in a data set. To identify the clay minerals of the studied area the extracted unknown spectra of the purest pixels was matched to pre-defined (library) spectra providing score with respect to the library spectra. Three methods namely, Spectral Feature Fitting (SFF),Spectral Angle Mapper (SAM) and Binary Encoding (BE) were used to produce score between 0 and 1, where the value of I equal a perfect match showing exactly the mineral type. In the investigated area four clay minerals could be identified i.e. Vermiculite, Kaolinite, Montmorillinite, and Illite recording different scores related to their abundance in the soils. In order to check the validity and accuracy of the obtained results, X-ray diffraction analysis was applied on surface soil samples covering the same locations of the end-members that derived from hyperspectral image. Highly correlated and significant results were obtained using the two approaches (spectral signatures and x-ray diffraction).