TY  - CHAP
U1  - Konferenzveröffentlichung
A1  - Hermann, Matthias
A1  - Grießer, Dennis
A1  - Gundel, Bernhard
A1  - Dold, Daniel
A1  - Umlauf, Georg
A1  - Franz, Matthias O.
T1  - Targetless Lidar-camera registration using patch-wise mutual information
T2  - 25th International Conference on Information Fusion (FUSION 2022), 4-7 July 2022, Linköping, Schweden
N2  - Targetless Lidar-camera registration is a repeating task in many computer vision and robotics applications and requires computing the extrinsic pose of a point cloud with respect to a camera or vice-versa. Existing methods based on learning or optimization lack either generalization capabilities or accuracy. Here, we propose a combination of pre-training and optimization using a neural network-based mutual information estimation technique (MINE [1]). This construction allows back-propagating the gradient to the calibration parameters and enables stochastic gradient descent. To ensure orthogonality constraints with respect to the rotation matrix we incorporate Lie-group techniques. Furthermore, instead of optimizing on entire images, we operate on local patches that are extracted from the temporally synchronized projected Lidar points and camera frames. Our experiments show that this technique not only improves over existing techniques in terms of accuracy, but also shows considerable generalization capabilities towards new Lidar-camera configurations.
KW  - Lidar-camera registration
KW  - Mutual information
Y1  - 2022
SN  - 978-1-7377497-2-1
SB  - 978-1-7377497-2-1
SN  - 978-1-6654-8941-6
SB  - 978-1-6654-8941-6
U6  - https://doi.org/10.23919/FUSION49751.2022.9841290
DO  - https://doi.org/10.23919/FUSION49751.2022.9841290
N1  - Volltext im Campusnetz der Hochschule Konstanz via Datenbank IEEE Xplore abrufbar.
SP  - 8 Seiten
S1  - 8 Seiten
PB  - IEEE
ER  -