Coal Gangue Recognition Algorithm Based on Improved YOLOv5
Fangjun Gui, Shuo Yu, Hailan Zhang, Hongda Zhu
Abstract
Coal gangue sorting is an important part of the construction of smart mines. At present, coal gangue identification and detection algorithms are mostly based on laboratory simulations for testing. Under actual working conditions, there are problems such as missed detection and poor real-time performance. In response to this problem, an improved YOLOv5 coal gangue recognition algorithm was proposed, and a live image acquisition system was designed. First, Mosaic data enhancement is used to increase sample diversity, Add attention moudle to improve detection performance, and the camera is mounted for data collection in actual working conditions. The experimental results show that the improved algorithm has an accuracy rate of 95.5%, with high recognition accuracy, and has important reference significance for practical applications.