CaliPy - Camera Calibration Application

CaliPy is a Qt-based calibration application written in Python. It displays results generated by bbo(lab)-calibcam in the background to check camera calibration.

Currently supported calibration patterns:

• ChAruco Boards

Currently supported camera models:

• Pinhole Camera Model
• Omnidirectional Camera Model

Installation

CaliPy depends on the following Python packages:

• NumPy
• PyYAML
• OpenCV
• ImageIO with FFMPEG support
• Construct
• PyQt5
• PyQtGraph
• Bbolab/calibcamlib
• Bbolab/svidreader
• PyInstaller (optional, for standalone build)

Current best practise is to use conda env.

Conda

You can install all the required dependencies into a conda environment with the following command:

conda env create -f environment.yml

Then run the cali.py script with python cali.py from the newly created environment.

Alterntively if you have the pyinstaller package installed (i.e. via pip) you can turn CaliPy into a self-contained executable with the following command:

python3 setup.py standalone

The resulting CaliPy folder inside the dist subfolder can then be packed and distributed without the need to install any dependencies.

Usage

1. This is a GUI to view bbo-lab calibcam results. The software offers some commandline options:

   1. --system_file is for loading any calipy session files containing source info and video paths.
   2. --videos and --pipelines can be alternatively used to provide paths to video files, and any pipelines to be applied on these videos. The final output of the pipeline is used to display calibration.
   3. --calib_file is for loading calibraiton results from bbo-calibcam. If no video information is provided using the above commandline options, the software will attempt to load videos from the calib_file. Then, the video links in the calib_file should be active.

   Note: if no proper commandline options are provided, an empty GUI is loaded.

2. Add cameras with appropriate identifiers in the 'Cameras' dock on the top left side of the window.

3. In the 'Sources' dock, add a session and add a recording for each one of the cameras to the session. Multiple such sessions can be added.

4. This is a GUI to view bbo-lab calibcam results. On the Menubar, use Result > Load .npy to load calibcam calibration files.

5. Ideally, the software should match the already loaded videos to the 'rec_file_names' in the calibcam dictionary.

Format

Use File > Save.. to save the session and the added sources as a .system.yml file. The session can be reopened by selecting it under File > Open...

Bbo-calibcam files are either .yml or .npy files containing a dictionary. Use Result > Load Calib to load them.