Prediction of minerals in rock images using Markov logic

Mining engineers are frequently faced with problems of deciding on the best option to access the ore bodylocation and value of the ore body. The main problem occurred in those farther areas where no facilities are available for life. Many technical and financial risks are also involved. In that case, special experts and equipment are also needed to accessand locate the mineral. Moreover, a large amount of data is required to assist with this process. In developing countries, due to lack of modern technologies, it is difficult and costly to mine these precious natural resources. To provide a simple and low cost solution, we propose an IT based framework for this problem. In our proposed research, we use special feature of satellite images of hilly areas to classify the areas that possibly contain minerals and metals. The proposed system uses Mean shift algorithm to identify various features of colored images. Markov logic is used to classify each color according to their weights. The initial experiments show that the results are positive.