A method to improve mineral identification accuracy based on hyperspectral Data

Abstract

To improve the mineral identification accuracy of the rapid quantificational identification model, the noise was filtered in fragment based on the wavelength of altered mineral absorption peak and the regional spectral library that fitted for the study area was established. The filtered spectra were analyzed by the method with regional spectral library. Compared with the originally mineral identification result, the average efficiency rate was improved by 5.1%; the average accuracy rate was improved by 17.7%. The results were optimized by the method based on the position of the altered mineral absorption peak. The average efficiency rate would be improved in the future to identify more accurate minerals.