Research map / 3D vision / Depth and stereo
Infrastructure-based Multi-Camera Calibration using Radial Projections
Yukai Lin, Viktor Larsson, Marcel Geppert, Zuzana Kukelova, Marc Pollefeys, Torsten Sattler
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add robust rig averaging (ablation) | Inf+RD · GoPro Helmet / Indoor short sequences | Complete framesets (%): 54.5 -> 98.3 | Improved |
| add robust rig averaging (ablation) | Inf+RD · GoPro Helmet / Outdoor short sequences | Complete framesets (%): 67.6 -> 97.7 | Improved |
| add robust rig averaging (ablation) | Inf+RD · Pentagonal / Indoor short sequences | Good calibrations (%): 23.0 -> 43.1 | Improved |
| add robust rig averaging (ablation) | Inf+RD · Pentagonal / Outdoor short sequences | Complete framesets (%): 28.8 -> 79.2 | Improved |
| delay intrinsics estimation using 1D radial cameras instead of RD solvers (ablation) | Inf+RD+RA · GoPro Helmet / Indoor short sequences | Good calibrations (%): 75.6 -> 79.0 | Improved |
| delay intrinsics estimation using 1D radial cameras instead of RD solvers (ablation) | Inf+RD+RA · GoPro Helmet / Outdoor short sequences | Good calibrations (%): 45.5 -> 48.3 | Improved |
| delay intrinsics estimation using 1D radial cameras instead of RD solvers (ablation) | Inf+RD+RA · Pentagonal / Outdoor short sequences | Good calibrations (%): 38.3 -> 41.5 | Improved |
| delay intrinsics estimation using 1D radial cameras instead of RD solvers (ablation) | Inf+RD+RA · full sequences | Calibration accuracy | No clear effect |
| estimate intrinsics (Inf+1DR+RA) instead of using ground-truth intrinsics (Inf+K) | Inf+K · Pentagonal / Outdoor | Translation error (cm): 2.217 -> 0.423 | Improved |
| estimate intrinsics (Inf+1DR+RA) instead of using ground-truth intrinsics (Inf+K) | Inf+K · Pentagonal / Outdoor | Rotation error (deg): 0.295 -> 0.567 | Hurt |
| refine intrinsics in final bundle adjustment (ablation) | Inf+K · GoPro Helmet / Indoor | Reprojection error (px): 0.283 -> 0.270 | Improved |
| use map-based calibration instead of pattern-based Kalibr for localization | robot localization · garden dataset | Median position error (cm): 3.77 -> 3.56 | Improved |
Closest papers by meaning
- Automatic Camera Calibration Using Multiple Sets of Pairwise Correspondences (TPAMI 2017 · similarity 0.85)
- MamV2XCalib: V2X-based Target-less Infrastructure Camera Calibration with State Space Model (ICCV 2025 · similarity 0.84)
- UC-NERF: Neural Radiance Field for Under-Calibrated Multi-View Cameras in Autonomous Driving (ICLR 2024 · similarity 0.84)
- SOAC: Spatio-Temporal Overlap-Aware Multi-Sensor Calibration using Neural Radiance Fields (CVPR 2024 · similarity 0.83)
- PRaDA: Projective Radial Distortion Averaging (CVPR 2025 · similarity 0.83)
- Adaptive Fusion of Single-View and Multi-View Depth for Autonomous Driving (CVPR 2024 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
