Project VeloSLAM

Summary

Registration of 3D point clouds captured by moving platforms in large scale outdoor environments is an essential part of many applications such as automatic modeling of environments, dynamic map generation and 3D reconstruction. Due to the fact that inertial measurement units (IMU) and onboard units (OBU) are very expensive, we aim at developing a new method which is independent to the IMU or OBU systems. The existence of such a system would enable field workers to apply 3D mapping methods in different contexts, i.e., without the need of having equipped special vehicles. Since one scan of the Velodyne Lidar covers a 360 x 26.8 deg. field of view of the environment, there is enough information to estimate the relevance of each scan. Therefore, we want to setup joint research to register precisely overlapping 3D point cloud captured by the Velodyne HDL-64 Lidar into a complete model without any prior pose information. More precisely, the following steps will be executed:

1. Features will be extracted from the range, resp. reflectance image produced by this scanner and they will be used for obtaining a rough guess for the registration.

2. Methods from the current software package 6DSLAM will afterwards be used to register the scans in a global frame of reference.

3. In addition so-called ortho-image which can be obtained by a surveying aircraft or a satellite will be studied for their usability in an enhanced registration process.

By doing so, we will incorporate background knowledge into the SLAM system. The hereby used method will be a novel probabilistic framework to exploit prior, heterogeneous information to improve the quality, robustness and speed of 6D SLAM. Then, with this technology, we develop the new method of automatic modeling of large scale outdoor environments with Velodyne HDL-64 sensor scanner independent to the expensive IMU system.

Funding

VeloSLAM is funded by the National Natural Science Foundation of China (NSFC) under the grant number 41050110437 (International Cooperation and Exchange Project). For the project duration Andreas Nüchter is a fellow of the NSFC.

Project period: 2011

Partners

• Transportation Research Center, Wuhan University

  

• Automation group at Jacobs University gGmbH

  

Publications