Raspberry Raven Pi Security Camera Does Double Duty

The worst thing about holiday decorations is that while you could leave them up all year, your neighbors probably won’t like you very much for it. Christmas lights on your house are one thing, but as far as Halloween decorations go, [MisterM]’s raven security camera is one of the few exceptions to this rule.

Nevermore will [MisterM] wonder who goes there. As soon as this raven lays its beady red LED eyes on whatever is lurking in the garden, it comes to life with a bit of head swiveling and some random sounds. The bird either goes CAW! or quotes Christopher Lee’s reading of Edgar Allen Poe’s “The Raven”.

Inside this bird’s chest cavity is a Raspberry Pi 2 and standard camera, a servo to swivel the head, and an audio amplifier and speaker. This bird is running MotionEye on top of the Raspi OS so it can run a script whenever it senses motion.

We like that [MisterM] was able to find right-sized bits of plastic to mount the servo in the neck and the horn to the head. It just goes to show that not everything needs a 3D printer, a CNC, or woodworking. Check out the scary demo after the break.

Want to scare the whole neighborhood? Check out the science behind good-looking house projections.

11 thoughts on “Raspberry Raven Pi Security Camera Does Double Duty

    1. lol

      I was just thinking about the bit in “American Gods” where Shadow is following one of Wednesdays ravens

      the bird stops

      Shadow says “Hey! Quoth something!”

      the raven looks at him and says

      “Procreate elswhere!”

      then keeps flying

  1. Brilliant fun, but I think that the Raven’s voice is ‘Magpie’ chatter rather that the “Cronk Cronk ” of an actual raven. Maybe that’s a deliberate tribute to the rasp… pi?

  2. I like the design aesthetic but scene-change motion detection is usually worthless outdoors. Much better to combine something like a PIR and a microwave motion sensor in a logical ‘and’ to detect people, or use something like the Melexis or Panasonic passive IR ‘human presence’ detectors.

    I’ve been experimenting with the yolo2 inference model on a Jetson Nano for outdoor people detection. It works better than scene change detection, but still has issues with backgrounds and lighting unless you train for your actual surroundings.

Leave a Reply to Mike Giles Cancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.