Shopping online is a convenient way to get the stuff you want or need without ever having to leave home. More people shop online than buying physical goods directly from stores. Of course, this means those good are shipped straight to your door, which sends up a bright, shiny flag for porch pirates to begin their illegal shopping sprees. Unless your home, there’s nothing you can do to prevent theft from happening, but a decent security camera might be able to identify those thieves, and Bofu Chen from BarryNet has an easy to implement solution that won’t break the bank.
Chen’s DIY AI Security Camera was designed using a mobile phone or Pi Camera, a Raspberry Pi, and the open source BarryNet deep learning platform. The system also makes use of a laptop or desktop with Ubuntu installed to access the program’s analysis center and to check results of the deep learning application, as well as a Gmail account to send notifications to the user remotely.
All images and videos collected from an Android (only) phone or Pi Camera are analyzed on the Raspberry Pi, which hosts BarryNet locally, and then sends the information through email notifications. Users can train the AI platform to target different objects or people that enter the area being monitored, so the system doesn’t trip when a pet or family member enters the scene.
Chen has created a detailed walkthrough of the DIY AI Home Security Camera, complete with all the necessary code and software needed to get the system up and running on his BarryNet project page.