IMPROVEMENT AND RESEARCH ON PEDESTRIAN AND VEHICLE IMAGE DETECTION BASED ON YOLOV5S ALGORITHM
An improved YOLOv5s model is proposed to address the complex urban traffic environment,as well as the issues of missed detection,false detection,and low accuracy in pedestrian and vehicle detection results.The improved model adopts the EIoU loss func-tion and introduces the FocalEIoU loss function.Finally,the effectiveness of the improved algorithm is tested by selecting the PASCAL VOC dataset.The final experimental results show that the improved YOLOv5s algorithm improves mAP by about 4%and also improves detection accuracy,verifying the effectiveness of this study.
YOLOv5simage processingobject detectionpedestrian and vehicle detection
万方数据
维普
