Visual classification method based on CNN for coal-gangue sorting robots

Abstract

Sorting coal and gangue is an important means to improve the heating capacity of coal, realize the graded sale of coal and maximize the production benefits of coal mines. At present, the sorting of coal and gangue is still mainly manual, but the manual sorting efficiency is low and labor intensity is high, so it is particularly important to develop a coal and gangue sorting robot. The key technology of coal gangue sorting robot is the identification and classification of coal. This paper proposes a visual classification method based on CNN, constructs a visual depth neural network FCCN(Fast coal classification net), and implements a visual coal classification detection algorithm. Compared with the similar algorithm Mask R-CNN, it improves the classification accuracy, reduces the training time of the model, and has good application value.