Here’s a neat little Raspberry Pi project: a pet door which uses facial recognition to control which animals get into your house.
Attached to a regular cat or dog door, this system uses motion sensors to pick up an approaching animal. This turns on a webcam which uses OpenCV facial recognition to determine what kind of animal it is. If it’s a pet that matches the database, servos move to open the flap, and the animal can enter.
If it’s an unwanted intruder such as a raccoon, it remains shut.
On the opposite end another motion sensor just needs to be tripped and the flap will open automatically.
This is a really cool project with a lot of applications. With pets alone there’s flexibility with an example given of only allowing cats through and not dogs. This is useful if you’re doing some kind of training if you own both kinds of animal, or if you exclusively have one.
If your pet door is big enough for a large dog, it’s also a nice security measure to keep that point of entry into your home closed to criminals.
You can find everything you need to replicate this project over on Hackster.io. It was created to show off Windows IoT Core so it is built on top of that, and it also mentions that you can swap out the Raspberry Pi for a MinnowBoard.
Strangely, this isn’t the first maker project we’ve featured that used facial recognition on animals. SnappCat was another system we wrote about in April that would send you pictures of your cat when you were at work.