Gold-related lithologic and mineral mapping from Hyperion and ASTER data in the south Chocolate Mountains, California

Remote sensing has the potential to provide detailed information on mineralogy and geochemistry comprising the Earth's surface; it is therefore extremely useful for mapping rocks, mineral assemblages and weathering characteristics. In this dissertation, lithologic and mineralogical information extraction and visualization from EO-1 Hyperion and EOS ASTER data for gold exploration are addressed.; One Hyperion image, two ASTER images and one Landsat ETM image were acquired for the south Chocolate Mountains area. An evaluation of the use for mapping gold-associated lithologies and minerals is undertaken. The Maximum Likelihood classifier and matched filtering detector were employed to extract lithologic and mineralogical information from the three data sets. Comparison of Hyperion, ASTER and Landsat ETM data indicates that the classifications of ASTER and Hyperion images provide mostly similar results, but Hyperion can discriminate classes with minor differences better than ASTER and ETM. On the other hand, due to the presence of more shortwave infrared and thermal bands, analysis of ASTER data can provide much better lithologic mapping than ETM data.; A group of Tertiary volcanic rocks (Tv, Tvb), Mesozoic granitic rock (gr), pre-Tertiary sericite schist (mso) and quartz biotite gneiss (mc) can be well identified in the ASTER and Hyperion images. The rock units so identified correlate very well with those on the reference geological map. The overall accuracy of the classification of the five significant rock types is over 80% and the Kappa coefficient is over 0.75. The alteration zones in the study area can be identified better from ASTER radiance data than from the interpretation of traditional information sources. Four alteration minerals can be extracted from ASTER reflectance data using the constrained energy minimization algorithm and reference spectra from the ASTER spectral library.; A GIS icon-based visualization technique was developed and implemented in a prototype software system called IconMapper. This technique extends raster images from pixels to icon constructs, which enables the user to co-visualize more than three geospatial data layers in a simple raster GIS data model. The visualization of alteration mineral abundance extracted from ASTER data proves the usefulness of this technique.