Fast Extrinsic Calibration for 3D LIDAR

Abstract

3D LIDAR based environmental perception has been widely used in the field of robotics research especially for the rapid development of the vehicle's auto pilot. Coordinate transformation calibration between LIDAR and the self-driving vehicle's body is a prerequisite for environmental perception. Different from the usual LIDAR extrinsic calibration equipped with IMU, the transform matrix is calculated by using the vertical relationship between the flat ground plane and another flat plane perpendicular to the ground without depending on third-party sensors such as IMU or satellite positioning devices in the proposed method. The calibration process is divided into two steps. The first step is to determine the transform matrix from the LIDAR's frame to the self-defined ground plane's frame. The second step is to determine the transform matrix from the ground plane's frame to the vehicle's frame. The proposed calibration method is fast and valid and has been validated on a variety of LIDAR equipment including Pandar40, VLP-16, HDL-32E and so on. The calibration result is used for 3D-LIDAR based environmental perception and map reconstruction in the process of auto pilot. The experimental results show that the proposed calibration method is reliable and convenient.