Inspiration
College students that let trash pile in their dorm. Leads to poor hygiene.
What it does
Alerts residents when the trash is full. Keeps track of who took the trash out. Also distinguishes between recyclable and non recyclable.
How I built it
We used an ultrasonic sensor connected to an Arduino(which we used for its analog signal capablities), that gets tripped when the trash reaches a certain level. We used socket.io and flask for the software. To keep track of who takes the trash out we used NKK switches. We had serial connections to raspberry pi to handle the data in python. We also tried to integrate computer vision into it by using Google Vision API to distinguish between recyclable and non recyclable materials.
Challenges I ran into
The computer vision part was quite hard. Didn't have enough time to train and use opencv so we used Google Cloud's Vision API instead. Ran out of time because of technical inexperience. Moved from plan b to plan c. We didn't give up though, so we were able to come out with a working prototype. Figuring out the working of the hardware in a real world environment was very challenging. It took a lot of trial and error to obtain the perfect data needed.
Accomplishments that I'm proud of
Completing a working sensor that connects to a website.
What I learned
The use of raspberry pi, socket.io, flask, serial connections, web frameworks, Google Vision API.
What's next for Trash out!
More Computer Vision! Including tracking faces, instead of using switches.
Log in or sign up for Devpost to join the conversation.