Devansh Dhrafani*, Yifei Liu*, Andrew Jong, Ukcheol Shin, Yao He, Tyler Harp, Yaoyu Hu, Jean Oh, Sebastian Scherer

1. Obtain access keys by filling out a simple survey.
2. Paste the keys into download_data.py.

pip install minio tqdm
python download_data.py [--outdir] [--unzip] [--data]

• --data: default is all. Select:
  • depth for our stereo thermal & depth collection
  • thermal for the collection featuring prescribed fire and smoke
  • rosbags for LiDAR and IMU data
  • reconstruction for the reconstructed point clouds and trajectories from SLAM

Calibration file can be found in config/firestereo.yaml.

More details about each dataset sequence can be found in data_description.md.

python browse_data.py [--dir] [--options] [--skip] [--align_contrast] [--process]

• --dir: path to single thermal/depth directory or directory containing (thermal) img_left,img_right
• --options: visualize thermal or depth
• --skip: skip every nth frame
• --process: apply 16-bit to 8-bit processing with minmax or firestereo
• Navigate with:
  • <- -> for left, right
  • [ ] for left by n, right by n
  • p for printing current frame number
  • d for delete current frame
  • l for display horizontal lines across stereo pair
  • esc for exit

Visualize the thermal images with our 16-bit to 8-bit converstion and pre-processing pipeline by running python browse_data.py with --process firestere option. This will apply the same processing pipeline as in the paper.

If you find this work useful, please consider citing:

@article{firestereo,
  author={Dhrafani, Devansh and Liu, Yifei and Jong, Andrew and Shin, Ukcheol and He, Yao and Harp, Tyler and Hu, Yaoyu and Oh, Jean and Scherer, Sebastian},
  journal={IEEE Robotics and Automation Letters}, 
  title={FIReStereo: Forest InfraRed Stereo Dataset for UAS Depth Perception in Visually Degraded Environments}, 
  year={2025},
  volume={10},
  number={4},
  pages={3302-3309},
  doi={10.1109/LRA.2025.3536278}
}