Hyperspectral Data Mining for Characterising Granite Type Uranium Deposits in South China

Based on hyper spectral data mining techniques, field spectra data is used to study the diagnosable spectral signatures of uranium mineralization factors in this paper, and the spectral features of pluton, ore-controlling fault and hydrothermal alteration including alkali metasomatism, hematitization, chloritization and hydromicazation are acquired and analyzed, the diagnosable spectra identification symbol has been developed. Furthermore, the radioactive anomaly characteristics is the key symbol to distinguish the uranium deposits from other metallic ore deposits, the radionuclide including uranium, thorium, potassium and radium are relative abundant in the uranium deposits. The author uses integration technology of remote sensing imagery and aerial radioactivity data to extract the anomaly information which relate to uranium deposits, and the practical result is achieved. Above studies and experiences provide the new approaches to the exploration of uranium resources.