Mineralization information extracted based on spectral angle mapping and SVM :A case study in Inner Mongolia, Ulan Chubb

摘要

Based on traditional Mineralization information extraction technology, use of spectral angle mapping method (SAM) and support vector machine (SVM) method, alteration information extraction are used in Inner Mongolia, Ulan Chubb. First, preprocessing the original ASTER remote sensing image, and then study of the ore-related alteration of geological, in view of the different altered information uses the spectral Angle mapping method of selecting the training sample. Second, using SVM method to solve optimal hyperplane, then determine the decision function, establish of forecasting model. Lastly, generalizing promotion recognition to identify other samples to get alteration of information. Compared Alteration of the information extracted and the original geological data found consistent with the data. View from the field investigation, found in varying degrees of mineralization, and get a good effect.