An LDA and RBF-SVM Based Classification Method for Inertinite Macerals of Coal.

In view of the complicacy and the diversity of inertinite macerals of coal, a classification method based on linear discriminate analysis (LDA) and support vector machine (SVM) is proposed. Firstly, according to differences of texture and intensity among macerals, inertinite macerals are represented with texture related features as energy, entropy, moment, local smooth and intensity related features as contrast, mean, standard deviation, 3-order moment deviation. Then, by using LDA, the initial features are further extracted by means of maximizing the inter-class dispersion as well as minimizing intra-class scatter. Finally, a classifier based on SVM with radial basis function (RBF-SVM) is built for the automatic classification of inertinite macerals. Experimental results show that, textures employed in this paper can present inertinite macerals effectively; with LDA and RBF-SVM, the dimension of feature space is reduced and the accuracy of classification is increased obviously.