Camera Calibration Workbench

Camera Calibration Workbench is a PyQt6 + OpenCV desktop application for chessboard-based camera calibration, image undistortion, and pixel-to-world coordinate conversion.

It supports:

• standard pinhole camera calibration,
• OpenCV fisheye camera calibration,
• calibration parameter export/import,
• batch undistortion,
• fisheye scale tuning and principal-point offsets,
• coordinate conversion from pixel/depth to world coordinates.

Features

• Friendly desktop UI with non-blocking background calibration tasks.
• Structured numeric controls for chessboard dimensions, grid size, fisheye scale, and principal-point offsets.
• Live runtime log capture from OpenCV calibration steps.
• JSON calibration parameter files.
• Optional automatic undistortion after calibration.
• Compatible launcher for the original frontendapp_pyqt6.py entry point.
• uv-based local environment setup.

Installation

Requirements:

• Python 3.10+
• uv
• Linux desktop environment capable of running Qt applications

Clone and run:

git clone https://github.com/Rosatus/CameraCalibrationWorkbench.git
cd CameraCalibrationWorkbench
./run_app.sh

The launcher creates venv/ with uv when needed and installs the project in editable mode.

Manual setup:

uv venv venv
uv pip install --python venv/bin/python -e .
source venv/bin/activate
python -m camera_calibration_workbench

Usage

1. Choose a directory containing chessboard calibration images.
2. Select camera type: standard or fisheye.
3. Set inner corner count.
   • Example: a 12 by 9 square board has 11 by 8 inner corners.
4. Set the physical grid size in meters.
5. Choose an output directory.
6. Click Start Calibration.

The application writes:

calibration_params.json
undistorted_images/

when automatic undistortion is enabled.

Fisheye Tuning

For fisheye images, the UI exposes:

• scale: adjusts the output field of view.
• dx, dy: offsets the principal point before saving parameters or undistorting.

Typical scale values:

0.8-0.9   preserves more visual content
1.0       default OpenCV-style output
>1.0      tighter view with more cropping

Project Layout

src/camera_calibration_workbench/
  app.py          PyQt6 desktop UI
  calibrate.py    OpenCV calibration and undistortion implementation
  coordinate.py   pixel/depth to world-coordinate conversion
  workflow.py     UI-friendly workflow helpers
tests/            backend regression tests
docs/             metric and usage notes

Development

uv venv venv
uv pip install --python venv/bin/python -e ".[dev]"
source venv/bin/activate
pytest

Run the application from source:

python -m camera_calibration_workbench

License

This project is licensed under the GNU Affero General Public License v3.0 or later. See LICENSE.