Abstract:
Objectives To overcome the limitation of the existed video synopsis methods which are unable to express the global trajectory of video object in multi-camera environment, this paper proposes a multi-video synopsis method in geographic scene considering the virtual viewpoint area of camera.
Methods First, on the basis of video spatialization, the virtual viewpoint area of the camera in geographic scene is calculated. Then, based on the data organization of multi-video objects, cameras are divided into multiple sets to obtain the virtual viewpoint areas of multi-camera. Finally, multi-camera video synopsis is achieved in the virtual geographic scene.
Results The simulation results show that the same video object captured by multiple cameras can be comprehensively expressed with less time, comparing to the play time accumulated by single-camera video synopsis.
Conclusions The proposed method can express the global trajectory of video objects in multi-camera environment, and get a higher time compression ratio than the single-camera video synopsis method.
下载:
