__We are submitting for the Software Track__

Inspiration

We saw how much trash was in the ocean and wanted a way to fix it.

What it does

It is an AI model that identifies a person coming up and will have a light pointing at them and will let them deposit trash. The AI model will then look at the trash and if is indeed trash based on the AI model (a living thing won't count as trash) based on a dataset. The AI model is trained not to accept the item if it isn't trash.

How we built it

We used a dataset of different objects (bottles, animals, electronics, etc) and created a program that will differentiate whether or not the object qualifies as trash. We did the same type of program for our other AI model but used it to differentiate whether or not an object was a person (which we labeled as a "dolphin" after our mascot's pet in our logo).

Challenges we ran into

Initially, we wanted to make a 3D model but, the member we had in charge of the mechanical components of our project didn't come back the next day and the work he did didn't work so we are instead just doing a software-based project, and plan on doing the mechanical components in future.

Accomplishments that we're proud of

We are very proud of our 2 AI models. One of which is able to identify trash and the other which is able to identify a human by using computer vision. We trained our AIs using data models that we researched.

What we learned

We learned a lot about how to use AI and how to use computer vision. We also learned a lot about using CAD software, even though we weren't able to submit the final project. This was all of our first time using CAD software AI models so we are very proud of what we learned through the process.

What's next for UrTrash

In the Future We plan on making a physical trash can that will be programmed to open when it identifies a person. We also want to try and put Urtrash on some beaches on the East Coast to help keep our oceans clean!

Built With

Share this project:

Updates