基于视觉的行人定位方法研究——北京故宫游客头戴相机视频记录的时空轨迹重建

Research on vision-based pedestrian location method:reconstruction of temporal and spatial trajectory recorded by head-mounted camera of tourists in the Forbidden City in Beijing

基于视觉的行人轨迹定位方法,不依赖于信号接收设施,具有在建筑环境行为学研究中普及的潜力。研究针对城市公共空间,应用街景地图数据库,探索了地理坐标系中用视觉匹配法从行人头戴相机视频中恢复轨迹的技术管线。

A technique known as vision-based localization does not rely on a signal reception infrastructure,and has the potential to be widely used in environmental behavior research.Our research focuses on the scenario of urban public space,and we examined the technical pipeline of employing the street view map database for rebuilding trajectory from pedestrian head-mounted camera video in the geographic coordinate system by a visual matching manner.