PLANE-BASED RANGE CALIBRATION METHOD FOR GEOSLAM ZEB-HORIZON HANDHELD LIDAR INSTRUMENT

Abstract

A plane-based calibration method was proposed for Geoslam Zeb-Horizon handheld LiDAR in this study. The Velodyne VLP-16 multibeam LiDAR is embedded in Zeb-Horizon as the data collector. Since the range accuracy of VLP-16 contribute significantly to the final object reconstruction, it is intensively studied with self-calibration experiments. The calibration of Zeb-Horizon was performed in a underground parking lot which provide a large variety of planar features with different range. The values of the plane parameters were obtained from plane fitting by least square method with the point cloud from the ground-based LiDAR and divided into calibration planes and check planes. Point clouds lie on the surfaces of planar features collected kinematicaly by Zeb-Horizon was extracted manually and removed noise point clouds by RANSAC algorithm.The systematic error parameters solved in this study included the range scale factor and the rangefinder offset those have been meationed in the past literature to improve the accuracy of VLP-16. However, Zeb-Horizon can output neither the parameters of the laser emitters, nor the information about each point generated from which laser emitter. Therefore, this study assumed that all laser emitters have consistent ranging errors, that is, only one set of systematic error parameters under the assumption that there was no angular observation error. The results suggest that the check plane accuracy improved average 41% when adding systematic error parameters to the adjustment mathematical model. The residuals and standard deviation of the adjustment were also significantly reduced.