Study on detecting impurities in coal based on improved YOLO v8
Aiming at the serious influence of impurities in crude coal such as iron wire,woven bag,wood and mesh on the operation of intelligent dry selection equipment and subsequent production processes in coal preparation plant,a method for identifying impurities on hand-sorting belt based on improved YOLO v8 was proposed.A global attention mechanism was introduced to enhance cross dimensional feature interaction in images,and a weighted bidirectional feature pyramid network structure was introduced to improve the model's multi-scale detection ability for clutter by adaptively controlling the fusion between feature maps of different scales.On this basis,WIoU loss function was used to replace CIoU loss function to improve the balance of sample quality in the process of model training and improve the performance of the model.The data set of impurities in coal was expanded by data enhancement,and the results of improved YOLO v8 were verified by experiments.The results of test showed that the average detection accuracy of impurities in coal on hand-sorting belt was obviously improved by the improved algorithm compared with the original YOLO v8,which laid a foundation for intelligent impurity removal in advance before raw coal preparation.
impurity removal of coalobject detectionfeature fusionattention mechanismloss function
万方数据
维普
