Smart Trashcan Conversion Kit

Back
Front

Inspiration

In certain countries such as Canada, garbage sorting is required. Recyclable must be separated from waste and thrown into a different garbage bin. This makes foreigners who are new to Canada and are unaware of its rules very uncomfortable and unsure of where to throw garbage. Our project aims to solve that problem.

What it does

The Smart Trashcan Conversion Kit is meant to latch on to the outside of a trash can frame and replace the lid. This smart conversion kit identifies garbage using a webcam, processes the garbage with a program built on a PyTorch neural network and tells an arduino which compartment to store the garbage in.

How we built it

A 3D model of the trash can frame was first constructed in order to plan out the layout of the arduino board and its components. Once the digital design was finalized the motor was secured in place and a cardboard platform was added. Lightweight and affordable materials were used for prototyping to avoid unnecessary spending.

A neural network was also trained on more than a thousand select images of trash in order to improve the accuracy in the identification of the types of trash.

Challenges we ran into

Near the end of the project we found out that the arduino board could not communicate with the python script. Unfortunately we ran out of time; the board and the script must remain separated for now.

Accomplishments that we're proud of

Training a neural network, and building the smart trashcan itself.

What we learned

The arduino board is a very capable electronics kit. If used correctly it can greatly benefit programmers looking to bring real world interaction to their programs.