Research on coal gangue recognition technology based on YOLOv8

Efficient identification and separation of coal and gangue is the key technology to improve the utilization rate of coal resources and reduce environmental pollution. Aiming at the problem that the traditional coal gangue recognition method relies on manual experience, low efficiency and insufficient robustness, this study proposes an intelligent coal gangue recognition method for complex industrial scenarios based on YOLOv8 target detection algorithm. Firstly, a set of coal gangue collection system is built, and the image data set is collected by industrial camera. Secondly, the model of YOLOv8 is trained, and the coal and gangue are detected in real-time based on YOLOv8. Experiments show that the overall mAP50 of the coal gangue recognition system based on YOLOv8 reaches 98.7 %, the mAP50 of coal and gangue reaches 98.7 % and 98.6 % respectively, and the real-time reaction time is 36ms. It can effectively distinguish coal and gangue, and provide theoretical basis and technical support for the subsequent intelligent separation of coal gangue.