Identification of Coal and Gangue by Self-Organizing Competitive Neural Network and SVM

Based on the difference of the gray scale and texture in the images of coal and gangue, the theory of the gray-level histogram was introduced. 5 characteristic parameters which are considered as the classification features were extracted from the gray-scale histogram. By introducing the self-organizing competitive neural network algorithm and support vector machine (SVM) algorithm, the identification of coal and gangue was completed respectively. Finally, the identification results between self-organizing competitive neural network and SVM were made a comparison. The results indicate that the 5 characteristic parameters extracted from gray-level histogram are valid as foundation in identifying the coal and gangue and also that the SVM algorithm is superior to the self-organizing competitive neural network algorithm in identifying the coal and gangue. The method combining gray-level histogram and SVM is effective and provides a new method in intelligent identification for coal and gangue.