A NON OVERLAPPING CAMERA NETWORK: CALIBRATION AND APPLICATION TOWARDS LANE DEPARTURE WARNING

摘要

In this paper, we present a new multi camera approach to Lane Departure Warning (LDW). First, a perspective removal transformation is applied to the camera captured images to convert them into bird's-eye view images. Then, the position of the two cameras relative to a reference point is accurately determined using a new calibration technique. Lane detection is performed on the front and rear camera images who results are combined using data fusion. Finally, LDW is implemented by determining the distance between the vehicle and adjacent lane boundaries. The proposed system was tested on real world driving videos and shows good results when compared to ground truth.