LIDAR SYSTEM CALIBRATION USING OVERLAPPING STRIPS

Abstract

A LiDAR system calibration procedure estimates a set of parameters that represent biases in the system parameters and measurements. These parameters can be used to improve the quality of any subsequently-collected LiDAR data. Current LiDAR calibration techniques require full access to the system parameters and raw measurements (e.g., platform position and orientation, laser ranges, and scan-mirror angles). Unfortunately, the raw measurements are not usually available to end-users. The absence of such information is limiting the widespread adoption of LiDAR calibration activities by the end users. This research proposes alternative methods for LiDAR system calibration, without the need for the system raw measurements. The simplified method that is proposed in this paper uses the available coordinates of the LiDAR points in overlapping parallel strips to estimate biases in the system parameters and measurements (more specifically, biases in the planimetric lever-arm offset components, boresight angles, ranges, and mirror-angles). In this approach, the conventional LiDAR equation is simplified based on a few reasonable assumptions; the simplified LiDAR equation is then used to model the mathematical relationship between conjugate surface elements in overlapping parallel strips in the presence of the systematic biases. In addition, a quasi-igorous calibration method is also proposed to deal with non-parallel overlapping strips. The quasi-rigorous method can handle heading angle and elevation variations of platform trajectories since it also makes use of time-agged point cloud and trajectory position data. To illustrate the feasibility and the performance of the proposed calibration methods, experimental results from simulated and real datasets are introduced.