On-line non-overlapping camera calibration net
This addresses camera calibration for robotics or autonomous systems, but appears incremental as it builds on existing PoseNet and online estimation techniques.
The paper tackles camera calibration for non-overlapping cameras by proposing an online method that estimates ego-motion and inter-camera pose using a PoseNet-based network, with experiments on simulations and the KITTI dataset showing effectiveness in simulation.
We propose an easy-to-use non-overlapping camera calibration method. First, successive images are fed to a PoseNet-based network to obtain ego-motion of cameras between frames. Next, the pose between cameras are estimated. Instead of using a batch method, we propose an on-line method of the inter-camera pose estimation. Furthermore, we implement the entire procedure on a computation graph. Experiments with simulations and the KITTI dataset show the proposed method to be effective in simulation.