Simulation Of Bidrectional Reflectance On Mixed Minerals And Granite Based On Hapke Photometric Model

Quantitative research minerals spectra may rely on Hyperspectral remote sensing technology, at the same time photometric model is one of the means of quantifying minerals feature by remote sensing. The article used IDL programming language and developed a program based on Hapke photometric model, using "nonlinear least-squares" method to simulate reflectance spectrum. Measured by a variety of rock powder spectrums and simulated spectrums comparison, using correlation coefficient R and root mean square error RMSE to evaluate the results. Finally we get the model has high accuracy and feasibility of simulating mixed minerals reflectance spectrum. Then taking advantage of the model, we attempted to simulate the Gansu Liuyuan region granitite reflectance spectrum, and according to the simulation spectrum characteristics, we tried to identify the scope of granite by CASI/SASI hyperspectral data. The result were excellent by on-the-spot investigation. Thesis achieved the following results:1. Based on Relab spectral library and our laboratory data about multi-minerals, multi-angles, multi-particle size spectral reflectance, we did rigorous and detailed study and evaluation on Hapke AMSA model and IMSA model simulating rock powder sample reflectance spectrum, the results demonstrate that the AMSA model is more accuracy and easier to debug; meanwhile the parameters of AMSA model are highly coupling, the more accurate parameter the higher the precision.2. In the laboratory, we measured different mixing minerals spectra, such as different mixing proportion and different mixed minerals, then using the AMSA model simulated the mixed minerals reflectance spectra. The results show that the model has high accuracy in the mixed particulate mineral, but it’s not good at flaky minerals unless adjusting the parameters of the model; Hapke model has a high practical value in the study of mixed minerals spectra, it can be used to research "non-linear" effect of mixing mineral spectra and quantify minerals composition and mineral abundances.3. Based on existing workspace comprehensive analysis report of the mineral composition of granite, using AMSA model approximate simulated granite reflectivity "pure spectrum", and combined with a variety of measurement spectra of granite, then summarizing the spectral characteristic rule. Finally, we used of ‘matched-filtering’ of hyperspectral recognition technology to carry out the granite scope. The result has high accuracy contrast to geological data, it proves that the process used to recognize granite is operability.In conclusion, the research results can improve people awareness of the Hapke photometric model and contribute to hyperspectral mapping in other region.