Mobile robotics can be successfully applied for 3D environment exploration and reconstruction. Here, a vehicle tracking system is described based on combining laser and camera information aiming at building maps of unstructured environments. The proposed system integrates a ground-fixed CCD camera with a phase-shifting laser to follow the vehicle moving across the environment. The method comprises two stages. The first stage is a precalibration process, for the gross estimation of camera parameters using laser data. The second stage is the tracking of the robot for map building of the environment. During the second stage, the camera calibration is refined updating its parameters through a recursive optimization using real time measurements of the vehicle positions, which are derived by the laser, automatically driven by the camera in this stage. The novelty of the proposed method is that it performs a realtime, accurate photogrammetric calibration without employing any calibration grid, thus preserving flexibility and allowing survey of large environments. Detailed results obtained by the simulations and preliminary experiments are described, showing the effectiveness of the system in refining the calibration parameters and accurately reconstruct the environment.
Mobile Robotics Technique for 3D Environment Reconstruction
REINA, GIULIO
2005-01-01
Abstract
Mobile robotics can be successfully applied for 3D environment exploration and reconstruction. Here, a vehicle tracking system is described based on combining laser and camera information aiming at building maps of unstructured environments. The proposed system integrates a ground-fixed CCD camera with a phase-shifting laser to follow the vehicle moving across the environment. The method comprises two stages. The first stage is a precalibration process, for the gross estimation of camera parameters using laser data. The second stage is the tracking of the robot for map building of the environment. During the second stage, the camera calibration is refined updating its parameters through a recursive optimization using real time measurements of the vehicle positions, which are derived by the laser, automatically driven by the camera in this stage. The novelty of the proposed method is that it performs a realtime, accurate photogrammetric calibration without employing any calibration grid, thus preserving flexibility and allowing survey of large environments. Detailed results obtained by the simulations and preliminary experiments are described, showing the effectiveness of the system in refining the calibration parameters and accurately reconstruct the environment.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
