Eye Pi

Eye Pi is a motion detection application written in C++ and Node.js for the Raspberry Pi with a camera attached.

Current implementation

• OpenCV detection of movement.
• C++ to Node communication using Node.js Native abstractions.
• Uses cmake-js and CMake, rather than node-gyp for easier OpenCV compiling.
• Streaming of camera frames to the browser using WebSockets.

Usage

Firstly, OpenCV needs to be installed, either built from source or retrieved. You can find OS specific installation instructions here

Then:

npm install

If no errors, then:

npm start

Open up http://localhost:3000, and view your camera streaming to the browser!

Todos

• Test on Raspberry Pi Zero.
• Use faster and more efficient serialisation and deserialisation from C++ to Node (currently using JSON with Base64 encoding for the data, which is slow).
• Use GridFS and MongoDB to store video and motion detection events.
• Timeline view of motion detection events and scrubbing (in the browser).
• Better motion detection implementation (currently very simple).

License

MIT