Uncertainty modelling of refraction effects in non-central camera calibration (13719) |
| Yu Lan, Mario Kolling, Alexander Dorndorf and Jens-André Paffenholz (Germany) |
Mr. Yu Lan
Scientific Assistant
TU Clausthal
Institute of Geotechnology and Mineral Resources
Geomatics for Underground Systems
Clausthal-Zellerfeld
Germany
|
| |
| Corresponding author Mr. Yu Lan (email: yu.lan[at]tu-clausthal.de, tel.: +491745978383) |
| |
|
| [ abstract ] [ paper ] [ handouts ] |
|
Published on the web n/a
Received 2025-09-16 / Accepted n/a |
| This paper is one of selection of papers published for the FIG Congress 2026 in Cape Town, South Africa in Cape Town, South Africa and has undergone the FIG Peer Review Process. |
FIG Congress 2026 in Cape Town, South Africa
ISBN n/a ISSN 2308-3441
URL n/a
|
Abstract |
| With the rapid adoption of LiDAR-camera multi-sensor system (MSS) in industrial and intelligent perception, accurate calibration of camera intrinsics is critical for reliable sensor fusion and robust scene understanding. Target-based camera calibration with ChArUco board or checkerboard is, however, limited by the quality of corner measurements. In practice these measurements carry uncertainty that propagates through the calibration pipeline and can bias both intrinsics and extrinsics. Moreover, in this work, in order to enable the camera to operate in extreme environments, the camera is enclosed in a protective housing with a front-mounted dome port. The dome port introduces refraction that perturbs ray geometry and invalidates standard pinhole mappings from 3D points to the image plane.
To address these challenges and improve LiDAR-camera sensor fusion, this study incorporates a physically grounded refraction model for an in-air dome-port camera into the imaging model. Uncertainty is explicitly modelled by assigning a covariance matrix to each detected ChArUco corner, and these uncertainties are propagated to the estimated intrinsics, distortion coefficients, extrinsics, and the decentring between the dome port centre and the camera optical centre. The resulting parameter estimates are accompanied by statistically justified covariance matrices, enabling uncertainty-aware residual weighting during optimization. This yields more consistent LiDAR-camera alignment and improves the accuracy of downstream state estimation.
|
| |
| Keywords: Photogrammetry; LiDAR-camera multi-sensor system ; camera calibration; ChArUco board; refraction model; uncertainty propagation |
