The determination of hardness and approximate mineral composition of rocks and classifying these lithologies aids in controlling various processes in the plant, such as reducing the grinding process, which accounts for about 50% of its energy consumption. In this paper, a new method for rock lithological classification is presented, based on color as well as 3D laser based features. The method uses color and laser range images, acquired from rocks on a conveyor belt, to compute Gabor and LBP (Local Binary Pattern) features. Various Gabor and LBP features are tested, including rotation invariant features. The images are tessellated into sub images in which the features are computed. The classification is performed in two stages. In the first stage, the sub-images are classified by using a support-vector machine (SVM) classifier. In the second stage, the classification is improved by a voting process among all the sub-images of each rock. The method was tested on a database with five different rock lithologies taken from a copper mine which has been used in previous studies, allowing comparison with our new results. The results show that the classification performance was improved significantly by adding the 3D laser texture features, and using a combination of rotation invariant Gabor and LBP features, achieving a classification accuracy of 99.24% on the database. Using the CMIM (Conditional Mutual Information Maximization) feature selection method showed that only 10% of the total extracted features are required to achieve the maximum correct classification rate and that using the 3D laser features, (for the first time in our rock classification method to the best of our knowledge) is important for maintaining high classification performance. (C) 2017 Elsevier B.V. All rights reserved.