High-precision Recognition of Highway Throwing Objects Based on Cooperative Perception of Vehicle and Road
Aiming to improve the accuracy of real-time identification of throwing object,a highway throw-ing objects recognition model based on CABIFPN-YOLOv5 is proposed.Firstly,a dataset for image recogni-tion of throwing objects was constructed;Secondly,the YOLOv5 algorithm model is used for network training;Finally,a bidirectional feature pyramid BIFPN network with stronger feature fusion capability is used,while incorporating Coordinate attention mechanismto improve the backbone network,thereby improving the real-time detection accuracy of the model.Experimental results demonstrate that the proposed model based on the CABIFPN-YOLOv5 algorithm has a mean recognition accuracy mAP@0.5 and mAP@0.5:0.95 reaches 97.9％and 94.7％,with an FPS of 89.5 Hz,enabling real-time and high-precision identification of throwing objects.
throwing objectshigh-precision recognitionCABIFPN-YOLOv5 algorithm model
万方数据
维普
