Research on Pedestrian Detection Method Based on YOLOv5
Aiming at the problems that YOLOv5 is prone to missing targets and low detection accuracy when detecting pedestrians,an improved pedestrian detection model based on YOLOv5 network is proposed.First of all,the backbone network uses the SPD-GConv module constructed by the combination of SPD(Space-to-Depth)module and Ghost convolution for downsampling to reduce the loss of fine-grained feature information.Secondly,the multi-scale detection ability of the model is enhanced by adding small size detection layer.Then,the original CIoU loss function is replaced by α-EIoU loss function to improve the accuracy of pedestrian target location.The Crowdhuman dataset is used for training and testing.The experimental results show that the recall rate and average accuracy of the proposed algorithm are 4.7%and 3.5%higher than those of the original algorithm,respectively,which can effectively improve the accuracy of pedestrian detection in remote targets and dense scenes.
pedestrian detectionYOLOv5SPD-GConvmulti-scale detectionloss function
万方数据
维普
