Object-based Video Reconstruction for Route-oriented Intercity Highway Tasks with Static Cameras

For intercity highway with many branches, the video data for each monitoring node is very large and it is unlikely to fuse these data without time synchronization. In this paper an object-based video reconstruction approach is presented for route-oriented tasks of intercity highway monitoring with many branches using low cost distributed video capturing devices installed on the roadside posts. First of all, the temporary vehicular objects are extracted using the background subtraction and then the video for each branch is reconstructed by fusing these objects into a preset background sprite based on the map of the temporary traffic states. Since the video for each branch is coded into the related ID and its motion parameters of the tracked object, any route-oriented tasks like vehicular object tracking or traffic routing can be easily visualized, providing a good link between drivers and the monitoring center. Experimental results show its effectiveness.