Abstarct: Mineral detection, in most mines of Iran, is conducted manually using a microscope by a skilled user expert. Considering the time-consuming nature of this process, such detection will not be economic for a large mine such as Sangan. In this study, minerals in Sangan Iron Ore mines in Khorasan Razavi province are identified using image processing techniques, and then the percentage of minerals in the image is calculated. For this purpose, the polished sections of this mine were studied. According to mineralogical studies performed by optical microscope, the main minerals (ore) include magnetite, hematite, gunges like pyrite, Chalcopyrite and Pyrrhotite. In this thesis, given the lack of a standard databaxse for this purpose, a databaxse was created for this mine. To identify these minerals, after reviewing the databaxse and the proposed methods, color and texture criteria were identified as two suitable properties for detection of these minerals. The feature vector, histogram of dataset images and texture features in the three color channels constitute the RGB of these images. The classification was carried out using the nearest neighbor method baxsed on the Jensen similarity criterion. The main advantage of this method is its accuracy compared to other classification methods. The evaluation was carried out on the dataset from each mineral, with the results indicating a detection rate of 89% for the proposed method in identifying minerals in Sangan mines. The Pixon method was used for segmentation in research.