Inspiration

We found that the current price of smart doors on the market is incredibly expensive. We wanted to improve the current technology of smart doors at a fraction of the price. In addition, smart locks are not usually hands free, either requiring the press of a button or going on the User's phone. We wanted to make it as easy and fast as possible for User's to securely unlock their door while blocking intruders.

What it does

Our product acts as a smart door with two-factor authentication to allow entry. A camera cross-matches your face with an internal database and also uses voice recognition to confirm your identity. Furthermore, the smart door provides useful information for your departure such as weather, temperature and even control of the lights in your home. This way, you can decide how much to put on at the door even if you forgot to check, and you won't forget to turn off the lights when you leave the house.

How we built it

For the facial recognition portion, we used a Python script & OpenCV through the Qualcomm Dragonboard 410c, where we trained the algorithm to recognize correct and wrong individuals. For the user interaction, we used the Google Home to talk to the User and allow for the vocal confirmation as well as control over all other actions. We then used an Arduino to control a motor that would open and close the door.

Challenges we ran into

OpenCV was incredibly difficult to work with. We found that the setup on the Qualcomm board was not well documented and we ran into several errors.

Accomplishments that we're proud of

We are proud of getting OpenCV to work flawlessly and providing a seamless integration between the Google Home, the Qualcomm board and the Arduino. Each part was well designed to work on its own, and allowed for relatively easy integration together.

What we learned

We learned a lot about working with the Google Home and the Qualcomm board. More specifically, we learned about all the steps required to set up a Google Home, the processes needed to communicate with hardware, and many challenges when developing computer vision algorithms.

What's next for Eye Lock

We plan to market this product extensively and see it in stores in the future!

Built With

Share this project:
×

Updates