Visual modeling of mineral potential exploration using support vector machine

With the advent of big data in geosciences, exploration studies have entered new dimensions. Big data means high resolution image information. Since these data in geosciences have a very large volume and variety, it is necessary to use big data analysis approaches in this field. In this study, the application of support vector machine in machine vision in the field of mineral potential exploration is investigated. In recent years, image classification has attracted a lot of attention in machine vision, whose processes include pre-processing and segmentation, feature extraction and related class identification. In this study, geological maps and remote sensing images are used to model the exploration of minerals potentials, and Alexnet architecture is used to automatically extract features, and field information is used to learn the algorithm. In the next step, support vector machine is used for modeling in order to identify structure factors in the occurrence probability of minerals potentials. Algorithms and evaluation indicators are programmed in MATLAB environment at each stage. The accuracy obtained using this method is 71% on the test data. According to the previous study conducted by the authors in identifying mineralization structures, the average accuracy of image data classification using convolutional neural network algorithms is 65%, the spectral angle mapper method in identifying alteration zones is 70% and applying filters in identifying faults is 28%. As can be seen, the method used in this research is highly accurate. Its advantages include reducing costs and speeding up the decision-making processes.