camera-intrinsic-calibration

camera-intrinsic-calibration

A pure rust camera intrinsic calibration tool.

Installation

# cargo install cli
cargo install camera-intrinsic-calibration
# Or pip install cli
pip install camera-intrinsic-calibration
# Or pip install cli with rerun
pip install 'camera-intrinsic-calibration[rerun]'

Or download from the latest release.

CLI Usage

# camera-calibration-rs
ccrs -h

# run intrinsic calibration on TUM vi dataset
# Download and untar
wget https://vision.in.tum.de/tumvi/exported/euroc/1024_16/dataset-calib-cam1_1024_16.tar
tar xf dataset-calib-cam1_1024_16.tar

# [Optional] export RUST_LOG=trace
ccrs dataset-calib-cam1_1024_16 --model eucm

Visualize details after calibration

# use cargo to install rerun
cargo install rerun-cli --version 0.27.2
# or use pip to install rerun
pip install rerun-sdk==0.27.2
# visualize result
rerun results/20YYMMDD_HH_MM_SS/logging.rrd

Supported formats

Dataset format

• Euroc (default)

  dataset_root
  └── mav0
      ├── cam0
      │   └── data
      │       ├── {time_stamp}.png
      │       ├── {time_stamp}.png
      │       └── {time_stamp}.png
      └── cam1
          └── data
              ├── {time_stamp}.png
              ├── {time_stamp}.png
              └── {time_stamp}.png
  

• General --dataset-format general

  dataset_root
  ├── cam0
  │   ├── any_file_name.png
  │   ├── any_file_name.png
  │   └── any_file_name.png
  └── cam1
      ├── any_file_name.png
      ├── any_file_name.png
      └── any_file_name.png
  

  Images can be either .png or .jpg, but .png is preferred if possible. PNG is lossless compression, while JPG is not.

Camera models

• Extended Unified (EUCM)
• Extended Unified with Tangential (EUCMT)
• Unified Camera Model (UCM)
• Kannala Brandt (KB4) aka OpenCV Fisheye
• OpenCV (OPENCV5) aka plumb_bob in ROS
• F-theta (FTHETA) by NVidia

Examples

cargo run -r --example convert_model

Calibrate your own camera

Please follow the tutorial.

Acknowledgements

Links:

• https://cvg.cit.tum.de/data/datasets/visual-inertial-dataset
• https://github.com/itt-ustutt/num-dual
• https://github.com/sarah-quinones/faer-rs

Papers:

• Kukelova, Zuzana, et al. "Radial distortion homography." Proceedings of the IEEE conference on computer vision and pattern recognition. 2015.

TODO

• Multi-camera extrinsic
• More calibration info