Visual–inertial state estimation with camera and camera–IMU calibration

Abstract

In the last two decades, the Visual–Inertial Odometry (VIO) has recently received much attention for efficient and accurate ego-motion estimation of unmanned vehicle systems. In particular, the VIO includes only an Inertial Measurement Unit (IMU) and a camera. In this paper, we present a novel calibration approach for accurate deployment of monocular VIO. For this purpose, the hybrid optimization algorithm is used for calibrating the camera intrinsic and camera–IMU extrinsic calibration, automatically and without knowing the mechanical configuration. It is a professional work to carefully calibrate the intrinsic and extrinsic parameters, and it is required to repeat this work when the mechanical configuration of the sensor suite changes. Quantitative comparisons our method with the offline conventional calibration method on the KITTI dataset verify the efficacy and accuracy of the proposed method. We also demonstrate the performance of the proposed approach in large scale outdoor experiments.