This paper addresses the problem of calibrating the six degrees-of-freedom rigid body transform between a camera and an inertial measurement unit (IMU) while at the same time estimating the 3D motion of a vehicle. A high-fidelity measurement model for the camera and IMU are derived and the estimation algorithm are implemented within the particle filter (PF) framework. Belonging to the class of Monte Carlo sequential methods, the filter uses the unscented Kalman filter (UKF) to generate importance proposal distribution. It can not only avoid the limitation of the UKF which can only apply to Gaussian distribution, but also avoid the limitation of the standard PF which can not include the new measurements. Moreover, the proposed algorithm requires no additional hardware equipment. Simulation results illustrate the ill effects of misalignment on motion estimation and demonstrate accurate estimation of both the calibration parameters and the state of the vehicle.

Date Published: 31 January 2013
PDF: 6 pages
Proc. SPIE 8759, Eighth International Symposium on Precision Engineering Measurement and Instrumentation, 875946 (31 January 2013); doi: 10.1117/12.2014423

Published in SPIE Proceedings Vol. 8759:
Eighth International Symposium on Precision Engineering Measurement and Instrumentation
Jie Lin, Editor(s)