文摘
In video surveillance application, in order to deploy the cameras effectively, we need to know the information of the monitored scene, including the position, shape and location of the obstacle in the scene. Also, the area where the target appears frequently. Aiming at the complex structure and difficult modeling of the scene, a scene obstacle model construction method based on moving target tracking is proposed. The scene observation model can be obtained by using the spatial projection relationship to determine the position of the obstacle and interesting area. This method only needs to track the target motion, and does not need to reconstruct the 3D structure and image mosaic, which can effectively solve the online adaptive modeling of the scene.