Obstacles Extraction from a Video Taken by a Moving Camera

In automatic collision avoidance systems, the ability to detect obstacles is important. This paper proposes a method of automatic obstacles detection employing a camera mounted on a vehicle. Although various methods of obstacles detection have already been reported, they normally detect moving objects such as pedestrians and bicycles. In this paper, a method is proposed for detecting obstacles on a road, even if they are moving or static, by the use of background modeling and road region classification. Background modeling is often used to detect moving objects when a camera is static. In this paper, we apply it to a moving camera case in order to obtain foreground images. Then we calculate the camera motion parameters using the correspondence of feature points between two consecutive images and detect the road region using motion compensation. In this road region, we carry out regional classification. We can delete all objects which are not obstacles in the foreground images based on the result of the regional classification. In the performed experiments, it is shown that the proposed method is able to extract the shape of both static and moving obstacles in a frontal view from a car.