A Python package for controlled and automated image and video recording with the raspberry pi

pirecorder is a Python package, built on the picamera and OpenCV libraries, that provides a flexible solution for the collection of consistent image and video data with the raspberry pi. It was developed to overcome the need for a complete solution to help researchers, especially those with limited coding skills, to easily set up and configure their raspberry pi to run large numbers of controlled and automated image and video recordings using optimal settings.

A paper accompanying this package is published in the Journal of Open Source Software:

Jolles, J.W. (2020). pirecorder: controlled and automated image and video recording with the raspberry pi. J. Open Source Softw. 5, 2584. doi: 10.21105/joss.02584.

• Controlled recording using custom, easy-to-edit configuration files
• Record single images and videos, timelapses, and sequences of videos
• Configure camera settings interactively via a live camera stream
• Dynamically draw the region of interest for your recordings
• Automatic naming of files and folders with relevant and custom labels
• Easy scheduling and automating recordings in the future
• Direct control of all modules via simple terminal commands
• Convert (folders of) images and videos with resize, monitor, and label options
• Dedicated documentation website with detailed guides and tutorials
• Jupyter notebook tutorial files

pirecorder consists of a main PiRecorder module to run image and video recordings, stream and camconfig modules with interactive user interfaces for help setting up, calibrating, and configuring the camera, a schedule module for scheduling future recordings, and a convert module for the easy converting of (folders of) recorded images and videos.

To install the latest release, simply open a terminal window and enter:

pip install pirecorder

To install the latest development version, enter:

pip install git+https://github.com/jollejolles/pirecorder.git --upgrade

pirecorder builds strongly on the picamera package, uses numpy, pyyaml, and opencv for some of its core functionality, and relies on various utility functions of the accompanying pythutils package. Scheduling functionality is based on CronTab and the associated python-crontab package. All dependencies are automatically installed with pirecorder except for:

• OpenCV: has to be manually installed due to its various dependencies on raspberry pi. Click here for a quick install guide.
• FFmpeg: is only needed for the convert functionality of pirecorder. Click here and here for guides to install on raspberry pi and OS X respectively.

For detailed documentation and tutorials about pirecorder and all its functionalities, see the dedicated website jollejolles.github.io/pirecorder/.

1. quick guide
2. the pirecorder package
3. setting up your raspberry pi
4. installing pirecorder
5. position and calibrate the camera
6. configure recording settings
7. configure camera settings
8. record and schedule recordings
9. converting media
10. run from the command line

To test all functionalities of the pirecorder package, run the tests/test.py file (here), or alternatively run commands manually using the documented jupyter files here. Note that running the tests will require user input as some of the functionalities are interactive.

pirecorder is developed by Dr Jolle Jolles, a research fellow at the Max Planck Institute of Animal Behavior, and at the Zukunftskolleg, Institute of Advanced Study at the University of Konstanz.

For an overview of version changes see the CHANGELOG and for detailed changes see the commits page. Please submit bugs or feature requests to the GitHub issue tracker here.

Contributions to this package are welcomed via the usual pull request mechanism.

If you use pirecorder in your research, please cite the accompanying paper:

@misc{Jolles2020,
      title = {pirecorder: controlled and automated image and video recording with the raspberry pi},
      author = {Jolles, Jolle W.},
      year = {2020},
      volume = {5},
      number = {54},
      pages = {2584},
      doi = {https://doi.org/10.21105/joss.02584}
}

Released under a Apache 2.0 License. See LICENSE for details.