ORIGINAL PAPER Lithological classification and chemical component estimation based on the visual features of crushed rock samples Farzaneh Khorram 1 & Amin Hossein Morshedy 2 & Hossein Memarian 1 & Behzad Tokhmechi 3 & Hamid Soltanian Zadeh 4 Received: 22 April 2016 /Accepted: 19 July 2017 # Saudi Society for Geosciences 2017 Abstract Lithological classification and monitoring of ore quality are challenging issues in mine operation, especially to maintain desired feed for processing minerals plants. Quantification of visual features is an innovative method to analyze rock components. In this paper, an analysis of vision-based rock type and classification algorithm is pro- posed for fast, inexpensive, and reliable identification pro- cess compared with common chemical analysis. To evaluate the proposed algorithm, samples were collected from the Novin limestone mine in Iran. Based on chemical and thin-section studies, the samples were classified into three different groups as calcium carbonate, dolomite, and dolo- mitic limestone. The limestone samples were crushed, sieved, and, respectively, divided into three size fractions as 2.5–7, 1.5–2.5, and 0.1–1.5 cm. A set of 58 images as large and medium size samples and 78 images as fine size samples were taken in the laboratory environment. Features were extracted from the captured images and reduced by applying principal component analysis (PCA). The support vector machine (SVM) and Bayesian techniques were used for classification. The best classification accuracy was about 80 and 90% in limestone and dolomite rock samples, respec- tively. Then, a multi-layer perceptron (MLP) neural network was employed to predict chemical compositions percentages. The determination coefficient within the range of 0.76 to 0.85 was observed and predicted values confirmed good performance of the grade estimation. The outputs illustrated that the proposed intelligent and automated technique can be successfully applied to monitor ore grade and classify lithol- ogy in different stages of mining projects. Keywords Image processing . Neural network . Ore grade prediction . Lithological classification . Support vector machine (SVM) Introduction In mining operation, raw materials are exploited from de- posits without a prior knowledge about their composition. To identify minerals, there are physical and chemical prop- erties such as mineralogy, porosity, texture, hardness, degree of liberation, and the vast range of mineral’ s attributes (Barker 2014). Over the years, different methods have been developed for the purpose of mineral identification; tradi- tional methods have been upgraded to innovative techniques with fast, inexpensive, and accurate performance. Development and improvement in image processing algo- rithms are applied to extract and quantify visual features, which can promote an automatic system to recognize dif- ferent patterns in the minerals. Several researches have pre- sented novel methods to classify and quantify mineral com- position in the microscopic and macroscopic scales (Dorador and Rodríguez-Tovar 2016; Jouini et al. 2015; Marschallinger 1997; Saxena et al. 2017). * Farzaneh Khorram farzanehkhorram@ut.ac.ir 1 School of Mining Engineering, College of Engineering, University of Tehran, Tehran, Iran 2 Department of Mining and Metallurgical Engineering, Yazd University, Yazd, Iran 3 School of Mining, Petroleum and Geophysics Engineering, Shahrood University of Technology, Shahrood, Iran 4 School of Electrical and Computer Engineering, College of Engineering, University of Tehran, Tehran, Iran Arab J Geosci (2017) 10:324 DOI 10.1007/s12517-017-3116-8