This is a highly technical paper, but it proposes an alternative approach to obtaining high accuracy trajectories for mobile lidar data collection that could be advantageous in urban settings. It combines stereo cameras with lidar and an IMU, but in a unique way that reduces the reliance on the IMU.
An integrated stereo visual-LiDAR odometry and reduced IMU is proposed in this paper in which first the stereo visual odometer estimates ego motion which is used to register point clouds then the GICP algorithm is used to refine the ego motion estimation, and, finally, the forward velocity and azimuth obtained by a visual-LiDAR odometer are integrated with reduced IMU in an Extended Kalman Filter (EKF) to provide final navigation solution. The proposed system outperforms the simulated Reduced Inertial Sensor System (RISS).
