Automatic positioning data correction for sensor-annotated mobile videos

摘要: Video associated positioning data has become a useful contextual feature to facilitate analysis and management of media assets in GIS social applications. Moreover, with today's sensor-equipped mobile devices, the location camera can be continuously acquired conjunction captured video stream without much difficulty. However, most sensor information collected from devices is not highly accurate due two main reasons: (a) varying surrounding environmental conditions during acquisition, (b) use low-cost, consumer-grade sensors current devices. In this paper, we enhance noisy generated by smartphones recording analyzing typical error patterns for real introducing robust algorithms, based on Kalman filtering weighted linear least square regression, respectively. Our experimental results demonstrate significant benefits our methods, which help upstream sensor-aided applications access content precisely.