Proceedings Volume 8037: Laser Radar Technology and Applications XVI
Orlando, FL, USA
Stationary lidar (Light Detection and Ranging) systems are often used to collect 3-D data (point clouds) that can be used for terrain modelling. The lidar gathers scans which are then merged together to map a terrain. Typically this is done using a variant of the well-known Iterated Closest Point (ICP) algorithm when position and pose of the lidar scanner is not accurately known. One difficulty with the ICP algorithms is that they can give poor results when points that are not common to both scans (outliers) are matched together. With the advent of MEMS (microelectromechanical systems)-based GPS/IMU systems, it is possible to gather coarse position and pose information at a low cost. This information is not accurate enough to merge point clouds directly, but can be used to assist the ICP algorithm during the merging process. This paper presents a method called Sphere Outlier Removal (SOR), which accurately identifies outliers and inliers, a necessary prerequisite to using the ICP algorithm. SOR incorporates the information from a low cost GPS/IMU to perform this identification. Examples are presented which illustrate the improvement in the accuracy of merged point clouds when the SOR algorithm is used.
S. E. Budge and K. von Niederhausern, “Automatic merging of lidar point-clouds using data from low-cost GPS/IMU systems,” in Laser Radar Technology and Applications XVI, M. D. Turner and G. W. Kamerman, Eds., vol. 8037. Orlando, Florida, USA: SPIE, Apr. 2011, p. 80370D. [Online]. Available: http://link.aip.org/link/?PSI/8037/80370D/1