Inspiration
After walking around Southfield, Michigan we noticed lots of trash on the streets. We then thought of a fun way to combat this problem.
What it does
The app uses an AI camera detection to detect different pieces of garbage that need to be disposed of properly. It also uses a point system to keep track of how much you've cleaned up. Using machine learning, we created an AI to detect which type or recyclable the item is and awards points.
How we built it
We use Adobe XD to prototype our application. Once we had the design made, we worked on integrating the design with ReactJS and Android Studio with Java. Using Google Cloud technologies, we store the user information including galleries and points to be redeemed later. Using Python and OpenCV, we trained an AI to detect which type of material the recyclable item is and classifies it. The points are awarded on the material and the frequency you pick up trash. The more you recycle, the more points you earn!
Challenges we ran into
The main challenge we ran into was getting the camera to learn what each object is. Having the AI determine the difference between aluminum and paper materials. It was very difficult when a cup would be called an airplane. All that set aside we fixed our detection problem.
Accomplishments that we're proud of
Getting the camera to detect different objects in our environment was a big accomplishment for us because we have never worked with this kind of software.
What we learned
We learned a lot when it came to using the camera to detect objects, but we also learned more on how to create apps using android studios.
What's next for TrashScan
Implementing a cleaner UI along with better camera detection would be the next steps for TrashScan.
Log in or sign up for Devpost to join the conversation.