Abstract— This paper briefly reviews recent research in the area of texture classification with a focus on the problem of marble classification, and proposes a simple approach for this purpose. It aims at the automatic analysis and classification of marbles and granites. This is an important requirement for this field of industry in today’s life and in the future, into groups according to quality: Class A excellent, Class B good, and Class C fair. The proposed approach consists of two phases: training and classification. In the training phase, we apply the k-means clusters unsupervised learning method to categorize large number of marble samples into three classes A, B, and C. The output of the first phase is a set of thresholds used as inputs for the second phase. For classification, we have used Sobel edge detector. Experimental results show that the proposed approach can achieve promising practical results.