Detecting objects in complex urban area using depth information

Vision-based driver assistance system in complex urban area is highly demanding. The basic requirement for the system is the capacity of detecting potential obstacles. In particular, the automobile, it is more dangerous than other objects. Facing the sensitive area called as “interesting area” in front of the driver within a certain range, this paper presents an object detection method fully using depth information of the scene, which based on the fact that potential obstacles may overlap in original image and could be separated in depth map in terms of their lateral and longitudinal positions. The testing result and related process details are described intensively in the paper, and our work is testified effective.