Share Email Print
cover

Proceedings Paper

3D scene reconstruction with monocular visual odometer based on inverse perspective mapping
Author(s): Cao Yu; Feng Ying; Li Wunan; Chen Yunjin; Wei Li-an; Lei Bing
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

With the utilization of high-accuracy inertial navigation system, the existing city 3D scene reconstruction system has complex equipment and high cost. For this concern, we improve the system based on monocular visual odometry and 2D laser scanner. It is monocular visual odometry algorithm that realizes scanner orientation and moving track calculating. Then, we actualize point cloud optimization according to the acquired track, aiming at obtain 3d reconstruction result of the target scene. Experiment result shows that: With a 280.27 meters long closed route, the location error of monocular visual odometer is 1.5 meter, error rate is 0.54%. Therefore, the method in this paper is high accuracy and low cost.

Paper Details

Date Published:
PDF
Proc. SPIE 8379, Laser Radar Technology and Applications XVII, 837918; doi: 10.1117/12.918748
Show Author Affiliations
Cao Yu, National Univ. of Defense Technology (China)
Feng Ying, National Univ. of Defense Technology (China)
Li Wunan, National Univ. of Defense Technology (China)
Chen Yunjin, National Univ. of Defense Technology (China)
Wei Li-an, National Univ. of Defense Technology (China)
Lei Bing, National Univ. of Defense Technology (China)


Published in SPIE Proceedings Vol. 8379:
Laser Radar Technology and Applications XVII
Monte D. Turner; Gary W. Kamerman, Editor(s)

© SPIE. Terms of Use
Back to Top