LOCALIZATION CORRECTIONS FOR MOBILE LASER SCANNER USING LOCAL SUPPORT-BASED OUTLIER FILTERING
- 1Department of Real Estate, Planning and Geoinformatics, Po. Box 15800, 00076 AALTO, Finland
- 2Robotics and Telematics, University of Würzburg, Würzburg, Germany
Keywords: Localization, Intrinsic localization, Noise Filtering, Outlier Removal, Mobile Laser Scanner
Abstract. Following the pioneering work introduced in [Lehtola et al., ISPRS J. Photogramm. Remote Sens. 99, 2015, pp. 25–29], we extend the state-of-the-art intrinsic localization solution for a single two-dimensional (2D) laser scanner from one into (quasi) three dimensions (3D). By intrinsic localization, we mean that no external sensors are used to localize the scanner, such as inertial measurement devices (IMU) or global navigation satellite systems (GNSS). Specifically, the proposed method builds on a novel concept of local support-based filtering of outliers, which enables the use of six degrees-of-freedom (DoF) simultaneous localization and mapping (SLAM) for the purpose of enacting appropriate trajectory corrections into the previous one-dimensional solution. Moreover, the local support-based filtering concept is platform independent, and is therefore likely to be widely generalizable. The here presented overall method is yet limited into quasi-3D by its inability to recover trajectories with steep curvature, but in the future, it may be further extended into full 3D.