Circular Targets for 3D Alignment of Video and Lidar Sensors

This paper presents a novel approach for solving the problem of 3D alignment between video and lidar sensors. The proposed method concerns intelligent vehicle applications, where the relative distance between sensor frames can be significant. Circular calibration targets are used in order to make full use of the perception properties of both lidar and video cameras, which greatly simplifies the calibration task. The method determines the relative pose in rotation and translation of the sensors, using sets of corresponding circular features acquired for several configurations of the targets. A performance analysis in simulation and an error propagation analysis are carried out. The calibration procedure is tested on different configurations, and the calibration accuracy and estimation of confidence intervals are evaluated on real data. keywords: Extrinsic Calibration, Lidar, Vision Sensor, Circular targets, Intelligent Vehicles.