Targetless Lidar-camera Calibration via Cross-modality Structure Consistency

Abstract

<p>Hand-eye calibration is popular in targetless sensor calibration, as it just requires the motion information of each sensor. However, hand-eye calibration is not applicable in degenerate scenes, where sensor motion lacks sufficient rotation. In this article, we propose a novel motion-based method to address this problem. The robustness and accuracy of our method are verified on six sequences of the KITTI Odometry dataset. Furthermore, our codes will be public on https://github.com/gitouni/Spatial-Temporal-LiDAR-camera-Calibration after acceptance.</p> <p><br></p>