Extraction of polymetallic mineralization information from multispectral Thematic Mapper data using the Gram-Schmidt Orthogonal Projection (GSOP) method

In order to accelerate the reconnaissance of mineral resources in Xingjiang Uygur Autonomous Region of China, multispectral Thematic Mapper (TM) data were used to delineate hydrothermal alteration related to polymetallic mineralization. In addition to the use of commercial image processing packages, new image processing software based on geological environment and spectral information have been developed. The Gram-Schmidt Orthogonal Projection (GSOP) method used in linear algebra for signal processing was adopted for our study. The kernel part of the algorithm is that the GSOP uses least mean square deviation of mathematical expectation in linear vector space (Hilbert space). In a vector space, one base axis is established and other transitional vectors are inputted to evaluate the effectiveness of GSOP. Satisfactory results have been achieved by applying this method to differentiate hydrothermal alterations in the Kangxiwa area of the Kunlun mountains, western China. The red colour in the image shows Pb-Zn-Cu-Au-Ag polymetallic mineralization zones, which were confirmed by field investigations and sample analyses in 1998.