HeadsUp

• 

  HeadsUp

Inspiration

Have you ever wanted to keep looking at your phone even when crossing the street? Have you felt the pain of skipping a text just because you were trying to be safe? HeadsUp is for you! It tracks vehicles and traffic lights to tell you when it is time for you to look up.

What it does

It can recognize vehicles and see if the traffic light is a go or a stop. This is then relayed to the user using LEDs! It is smart enough to ignore red colored walkways and vehicles not in your direct field of view. And there's more! Its backed by a handmade wooden box!

How I built it

We used OpenCV for image recognition with focus on improving the efficiency of the code because Raspberry Pi 3, however fast it might be still needs optimized image recognition especially when dealing with multiple object image recognition.

Challenges I ran into

Raspberry Pi 3 was a little time consuming to set-up but we finally got it working with some help! The main challenge from software stand of view was making the code efficient. Initially, the code to recognize the color of traffic lights took ~4 seconds per photo on a quad-core laptop. Further, the learning curve of OpenCV was challenging too!

Accomplishments that I'm proud of

Learning OpenCV and optimizing it for a Raspberry Pi was the big challenge and most fun we had solving! It was a great learning opportunity and it is what we are proud of most!

What I learned

We learned OpenCV and image recognition on a Raspberry Pi. We also learned that Pi 3 is not made to be used without a display and it is quite tough to set up.

What's next for HeadsUp

We were planning on implementing image recognition for crosswalks but were not happy with the accuracy of the results. In the future, we want to make image recognition more efficient on a Pi and making HeadsUp more connected with a phone by implementing navigation instructions and push notifications.

Built With

• python