Motion-based online calibration for non-overlapping camera views

The use of multiple cameras in vehicles becomes more and more attractive as hardware prices decrease rapidly. Multiple camera sensors can be used to cover the whole environment of a vehicle and for 3D scene reconstruction using stereo or structure from motion techniques. To be able to use all sensor informations in a common coordinate frame, it is necessary to know the relative positions and orientations of the cameras. Calibration procedures (offline or online) determine these parameters using point correspondences between the camera images. However, camera configurations to monitor the entire area around a vehicle often have non-overlapping fields of view due to cost reasons. In that case, common techniques based on corresponding image points are no more applicable. This contribution outlines a concept to perform an online calibration of multiple cameras on a mobile platform with non-overlapping fields of view. We use the motion of the cameras and local image features to define constraints that allow for the calculation of the calibration parameters.