Research Of Main Rock Forming Minerals Abundance Inversion With Thermal Infrared Hyperspectral Data

The thesis with "The Research and demonstration of thermal infrared hyperspectral mineralization and alteration mineral extraction" project, carry out mixed pixel decomposition method research for thermal infrared hyperspectral(TASI) data. Use four kinds of algorithms, which are the Unrestricted least squares(origin), the full constrained least square method(FCLs), based on the selection of the optimal endmembers FCLs algorithm(OES-FCLS) and Variable FCLS algorithm based on least square error(LES-FCLS), to mixed pixel decomposition with the thermal infrared hyperspectral data in Gansu Liuyuan huaheitan and comparative evaluation the results of the four methods. The main contents and results are as follows:(1) Compared the common methods of the spectral extraction by using the simulation data, but need more artificial intervention. VCA algorithm for End element extraction is litte lower accuracy than PPI, can be automatically, the fastest running efficiency, suitable for large areas of hyperspectral data.(2) In the common four kinds of mixed pixel decomposition algorithm based on the least squares, the UCLS algorithm has a weak anti noise ability, as a result of and for one and non- zero, can not show the true performance of mineral abundance. FCLS algorithm is better than UCLS in the anti noise performance, the abundance of map noise points has been reduced, and the abundance of the value and for one, non zero limit, the basic can reflect the abundance of real mineral abundance.(3) On the basis of the theory of end variables,carry out FCLS algorithm(OESFCLS) based on optimal terminal selection and Variable FCLS(LES-FCLS) algorithm based on least square error. OES-FCLS, LES-FCLS have a strong anti noise, the most close to the real abundance map and less noise, more pure. Mixed pixel decomposition of linear spectral decomposition algorithm based on the variable theory of the end element is more accurate than the least squares based spectral mixture solution(UCLS, FCLS),but the more algorithm is reduced efficiency。(4) Carry out a comprehensive comparative evaluation through the decomposition effect、 the operation efficiency and the anti noise performance. Selection of FCLS algorithm(OES-FCLS) is used as a method of mixed pixel decomposition for thermal infrared hyperspectral(TASI) data. At the same time also get the common rock forming mineral abundance map of the Gansu Liu Yuanhua black beach.