Skip to main content
Public - Anyone can download Featured

xcalib A9 r02_s01 Camera-LiDAR Matching Caches

October 15, 2025
Version 0.1.0
1.7 GB

Description

Public HDF5 caches for camera-LiDAR cross-modal matching and targetless extrinsic calibration. The release accompanies xcalib and the Radar Lab position-encoding study.

Contents

Each split stores per-frame camera images, LiDAR point clouds, object detections, camera-LiDAR match matrices, and the metadata needed by the xcalib data loader.

Published splits

  • a9_r02_s01_train.h5
  • a9_r02_s01_val.h5
  • a9_r02_s01_test.h5 (held-out test split)

Source and use

The caches are derived from the TUM Traffic / A9 s110 intersection recording. Users must follow both this release's license and the upstream dataset terms.

Technical Details

File Format: HDF5
File Size: 1.7 GB
Sample Count: 282
Version: 0.1.0
License: CC BY-NC-ND 4.0

Tags

xcalib camera-LiDAR cross-modal matching extrinsic calibration infrastructure HDF5 TUM Traffic A9

Dataset Files

Files for this release are hosted externally.

Use the download action to open the published dataset repository.

Citation

Guo, L., Tang, J., Bang, T., Zhang, T., Harris, A., Sartipi, M., and Cao, S. (2026). Position Encoding in Detection-Based LiDAR-Camera Matching: A Diagnostic Study at Infrastructure Sites. IEEE Sensors Letters. https://doi.org/10.1109/LSENS.2026.3708462. Also cite the upstream A9 Intersection Dataset when using these derivative caches.

Download Dataset

This dataset is publicly available for download.

Related Resources

GitHub Repository
Source code, scripts, and reproducibility resources.
Documentation
Project documentation, setup steps, or supporting notes.

Dataset Information

Downloads: 1
Created: Oct 15, 2025
Updated: Jul 28, 2026
Version: 0.1.0
License: CC BY-NC-ND 4.0