A New Technique for Vehicle Tracking on the Assumption of Stratospheric Platform.
Eihan SHIMIZU, Takashi SHIDOJI, Takashi Fuse · Journal of the Japan society of photogrammetry and remote sensing · 1999
High resolution and continuous images at the same point can be acquired by stratospheric platforms. The stratospheric platform has great potential with wider scope of utilization. For instance, the stratospheric platforms may provide measurement of traffic flow. In this study, we argue the possibility of vehicle tracking by stratospheric platforms, to be specific, we estimate displacement vectors of vehicles by applying the probabilistic relaxation method to time-series images. We improve the probabilistic relaxation method by (1) using the color information of vehicles, and (2) using the labeling of each other. The proposed method can be applied to appearance and disappearance of vehicles. Through the application to simulation data and sample image, we confirm the effectiveness of the proposed method.