Study on spacial distribution information extraction of uranium based on the spaceflight hyperspectral remote sensing in Dongsheng area in inner Mongolia

Hyperspectral remote sensing technology has both higher spectral resolution and spectral overlay without a break, providing abundant ground information, optimizing the rock mineral identifies and extraction conditions, strengthening the detectability to ground and the resolution capability to ground objects of remote sensing, heightening the quantitative level of remote sensing technology. Using hyperspectral remote sensing technology for uranium information's quantitative extraction has significant meanings. This paper chose DongSheng area in Inner Mongolia as research area, adopting mineraligation theories of sandstone type uranium deposit as academic basis, using hyperion data gained by EO-1 satellite which has the highest spectral resolution in sensors loaded by satellites at present as information source. On the basis of data preprocessing (including spectrum rebuilding, noise elimination, bandwave optimum), we utilized uranium mineral's spectra belt that has diagnostic absorption characteristic, combining spectral curves actually measured in the field, making use of information extraction techniques which based on regional distinctions in space, like principal component analysis, MNF transformation and on distinctions of pixel level, like waveband operation, spectrum matching, Spectral Angle Mapper Classification to declutch and extract minerals containing uranium and special distribution information of stratigraphic. And then we located the space position distribution of uranium. Consequently, we got five uranium mineral bonanzas in the research area from the angle of remote sensing. The research provides the theories basis for the development of uranium resources in DongSheng area.