期刊名称：Journal of Automation, Mobile Robotics & Intelligent Systems (JAMRIS)
印刷版ISSN：1897-8649
电子版ISSN：2080-2145
出版年度：2013
卷号：7
期号：2
页码：30
出版社：Industrial Research Inst. for Automation and Measurements, Warsaw
摘要：A common method used to obtain 3D range data witha 2D laser range finder is to rotate the sensor. To combinethe 2D range data obtained at different rotaƟonangles into a common 3D coordinate frame, the axis ofrotaƟon relaƟve to the mirror center of the laser rangefinder should be known. This axis of rotaƟon is a line in 3Dspace with four degrees of freedom. This paper describesa method for recovering the parameters of this rotaƟonalaxis, as well as the extrinsic calibraƟon between the rotaƟonalaxis and a camera. It simply requires scanningseveral planar checkerboard paƩerns that are also imagedby a staƟc camera. In parƟcular, we use only correspondencesbetween lines in the laser scans and planes inthe camera images, which can be established easily evenfor non-visible lasers. Furthermore, we show that suchline-on-plane correspondences can be modelled as pointplaneconstraints, a problem studied in the field of robotkinemaƟcs. We use a numerical soluƟon developed forsuch point-plane constraint problems to obtain an iniƟalesƟmate, which is then refined by a nonlinear minimiza-Ɵon that minimizes the ,,line-of-sight" errors in the laserscans and the reprojecƟon errors in the camera image.To validate our proposed method, we give experimentalresults using a LMS-100 mounted on a pan-Ɵlt device ina nodding configuraƟon.
关键词：laser range finder; extrinsic calibraƟon; pointplane;constraints; minimal condiƟons; maximum likelihood;esƟmate
