Robust Solution for Extrinsic Calibration of a 2D Laser-Rangefinder and a Monocular USB Camera

Abstract

With today’s technological development, almost every robotic system such as UAVs, mobile robots or autonomous ground vehicles are equipped with LIDARs and cameras. These sensors provide data which can be used to estimate the position and orientation of mobile robotic systems in the environment. The combining of the sensory data, commonly known as sensor fusion, contributes to the improvement of the environmental perception of the robots. In order to fuse this data, an extrinsic calibration of the sensors used is needed. In this thesis, a custom robust extrinsic calibration method of a 2D LIDAR and a USB camera is presented. The proposed approach uses a small calibration target with AR tags in the overlapping FOV of the sensors and projections of vectors onto a common plane to estimate the pose of the camera in reference to the LIDAR based on few constraints. The result of the calibration is a transformation matrix that can be later used by other applications. The calibration process is fast and simple to perform and is intended to be improved and used as a part of the greater UAV control system of the company Pro-Drone.