OPTIMUM INDEX FACTOR (OIF) FOR ASTER DATA: EXAMPLES FROM THE NEOPROTEROZOIC ALLAQI SUTURE, EGYPT

The Advanced Space-borne Thermal Emission and Reflectance Radiometer (ASTER) sysetm consists of three subsystems: (1) Visible and near infrared (VNIR) operating in 3 bands with 15m spatial resolution. (2) Short wave infrared (SWIR) operating in 6 bands with 30m spatial resolution. (3) Thermal infrared (TIR) operating in 5 bands with 90m spatial resolution. With its high spatial resolution in the VNIR subsystem and wide spectral range, ASTER will greatly improves our ability for the identification of geological materials with remotely sensed data. However, it is difficult to select the 3 bands that will give the best results for extracting lithological information from ASTER data. The 14 bands of ASTER data allow for 364 combinations of red-green-blue (RGB) color overlays. The optimum index factor (OIF), based on total variance within bands and correlation coefficient between bands, is a statistical approach to rank all possible 3-band combinations. Three-band combinations with high total variance within bands and low correlation coefficient between bands will have high OIF. These combinations are expected to have the maximum extractable lithological information since they use bands with the least redundancy in the remotely sensed data. We used the OIF technique to rank all 3-band combinations of ASTER data covering part of the Neoproterozoic Allaqi suture that exposed in the arid region of southern Egypt. This suture is an E-W trending deformation zone defined by S-verging fold and thrust belt and is dominated by ophiolites, shelf sediments, arc volcanics, and granitoids. Our study indicates that RGB color overlay using bands 3 (VNIR), 6 (SWIR), and 13 (TIR) has the highest OIF. Generally, combining bands from VNIR, SWIR and TIR yields high OIF because bands from these subsystems have low correlation coefficient between them. OIF results are further supported by qualitatively evaluating reflectance spectra curves of common rock types in the Allaqi suture. Serpentine, gabbro, and marble have spectral feature in band 3 (VNIR), rhyolite, biotite schist, and graphite have spectral feature in bands 6 and 8 (SWIR), and granite have spectral feature in bands 11 and 13 (TIR). Principal Component transformation can be used to reduce the redundancy in highly correlated bands when VNIR, SWIR and TIR subsystems are used separately.