Self calibration of multiple LIDARs and cameras on autonomous vehicles

•Solution for the automatic calibration of multiple LIDAR, Cameras and other 3D sensors, with minor user intervention.•Applicable both in static sensor-rich setups and in mobile systems, such as autonomous cars and other ADAS contexts.•Accessible methodology for the community since it uses standard algorithms and libraries in a ROS framework and a simple ball.

Autonomous navigation is an important field of research and, given the complexity of real world environments, most of the systems rely on a complex perception system combining multiple sensors on board, which reinforces the concern of sensor calibration. Most calibration methods rely on manual or semi-automatic interactive procedures, but reliable fully automatic methods are still missing. However, if some simple objects could be detected and identified automatically by all the sensors from several points of view, then automatic calibration would be possible on the fly. The idea proposed in this paper is to use a ball in motion in front of a set of uncalibrated sensors allowing them to detect its center along the successive positions. This set of centers generates a point cloud per sensor, which, by using segmentation and fitting techniques, allows the calculation of the rigid body transformation among all pairs of sensors. This paper proposes and describes such a method with results demonstrating its validity.