MINERALOGICAL EXPLORATION AND LITHOLOGIC CLASSIFICATION VIA SPECTRAL UNMIXING IN DEATH VALLEY, INYO COUNTY, CALIFORNIA

A 2013 ASTER satellite image of the Black Hills region of Death Valley, California, is used for mineral identification and mapping of Precambrian to Tertiary age sedimentary, igneous, and metamorphic rocks. Image processing techniques such as minimum noise fraction, pixel purity, and a spectral angle mapper allow for the extraction of spectral signatures leading to the identification of minerals in the image. K-Means, Unsupervised Classification is used to separate rock types by formation that produced eight distinct rock packages in the area. A mask was used to exclude data from vegetation. Spectral signatures are gathered by the n-D Visualizer that allows for the collection of pure endmembers in a pixel data cloud. Endmembers’ individual spectral signatures are viewed graphically on a Z-profile and compared to a library of spectral signatures. NASA’s Jet Propulsion Laboratory provided the ASTER Spectral Library. The minerals identified positively correlate to known rock types in the study area. After identification, a Spectral Angle Mapper (SAM) was used to create masks of each of the minerals found. A 3D surface view was created to illustrate the current topography. The surface view is generated by creating a hillshade and digital elevation model. Bands 3N and 3B were used for nadir and backwards viewing. A total of 37 ground control points (GCPs) tied the images together for accuracy. The study concludes that geological maps can be extracted from Aster imagery with high accuracy using remote sensing.