Research Of Pedestrian Detection,Tracking And Pose Estimation In Urban Scene

Pedestrian detection and pose estimation have a very wide range of applications in computer vision.It has become a hot topic attracting much attention.However,due to the differences between human appearance and serious occlusion of pedestrians,the traditional image processing methods cannot meet the needs of practical applications.This paper focuses on computer vision technology of artificial intelligence.It proposes object detection algorithms based on deep learning and applies them to urban scenes to complete pedestrian localization tasks.At the same time,depth estimation and object tracking methods are used to achieve the pose estimation of pedestrians effectively.This paper first establishes a dataset containing three types of pedestrian:common pedestrians,cyclists,and traffic police.With a total number of 1455 images,more than 7,000 pedestrians,the dataset meets the research needs of deep learning methods.Secondly,this paper improves the single-stage detector RetinaNet and uses the training method of transfer learning to achieve fast and accurate recognition of pedestrians.The final average recognition accuracy rate is more than 95%,with mAP reaching 72.16%.And the recognition time of a single image only needs 0.04 seconds.Then,this paper proposes a loss function with a penalty term,which can effectively solve the problem of pedestrian occlusion commonly encountered in urban scenes,Also it further improves the pedestrian detection mAP to 73.45%,while the precision is 95.03%,and the recall is 86.37%.Finally,based on the detection results,an IoU-based pedestrian tracking and depth estimation methods are used to design the pose estimation algorithm for the descending pedestrian in urban scenes.It satisfies the demand of pedestrian pose estimation in urban scene.