Performance analysis for coal texture classification

This paper presents a novel method to classify coal texture into the six major categories, namely, Anthracite, Lignite, Bituminous, Sub-bituminous, Graphite and Peat. Coal textures are stochastic in nature. The existing classification and retrieval algorithms work well for the classification of regular texture, but fail to give the same results for the stochastic textures. Stochastic textures are mixture of textures. These textures can be identified only by visually meaningful classifiers. Coal textures are classified by calculating the Tamura features, since they give near-eye perception. This computer vision based algorithm can be used for automated coal texture classification. The proposed method outperforms the other previously developed methods by providing the classification accuracy of more than 87% for all the types of coal.