A web application run on a Raspberry Pi that uses a camera to calibrate and checkpoint a light path.

Camera setup and corresponding web interface.

Homography demo video

Features include:

- live stream 
- raw, DNG or jpeg capture
- homography preview (when viewing a checkerboard)
- web view and capture API

Project organization

The Flash web server is in flaskserver, with flaskserver/app.py being the entrypoint. flaskserver/camera.py and flaskserver/cv.py are modules used by the web server for interacting with the camera and calculating homographies. There are some design notes in docs/web-design.rst.

Setup

The project relies on the Picamera2 Python package. The package documentation recommends installing it using apt:

sudo apt install python3-picamera2 --no-install-recommends

The other requirements for the project are listed in requirements.txt. If you are using a virtual environment, create it with access to the system packages so that it can have access to picamera2. For example:

python -m venv --system-site-packages venv
source venv/bin/activate
pip install -r requirements.txt

Run

From the flaskserver directory:

python ./flaskserver/app.py

Install web server as service

To make the Raspberry Pi start the web server on boot, a service must be registered. scripts/copy_service_config.sh shows an example of how this can be done, copying config/lightpath-camera.service to /etc/systemd/system, with that service file instructing scripts/service.sh to be run on startup.