Coal Gangue Recognition during Coal Preparation Using an Adaptive Boosting Algorithm

The recognition of coal and gangue is the premise foundation intelligent sorting. Adaptive boosting (AdaBoost) algorithm-based identification has not been studied in depth. This paper proposed a image algorithm strong classifier based on AdaBoost with genetic (GA)-optimized support vector machine (SVM). One thousand images were collected on-site expanded to five via rotation exposure adjustment. 12 gray-level gradient co-occurrence matrix texture features extracted construct feature vector, establishing training dataset test dataset. Selection SVM kernel function, GA optimization parameter setting, base number was discussed. effects AdaB-GA-SVM other classifiers different investigated. results indicated that accuracy GA-SVM best when function RBF population number, crossover probability, mutation probability 80, 0.9, 0.005, respectively. excellent effective classification performance highest 95%, precision rate 92.8%, recall 97.3%, KS values 0.79.