Piece of Garbage

Have you ever been super confused and have thus poorly sorted your trash vs your recycling? Did you wish there was a better way to do things? we do that.

Comment 1

Inspiration

Sorting trash and recycling into their respective containers is hard. Making it easier to recycle things and doing so correctly only helps better make our future sustainable, as it makes the whole process as friction less as possible, hopefully increasing the rate of recycling and decreasing the rate of incorrect categorization.

What it does

This is a Piece of Garbage© that takes a picture of whatever garbage that you may be holding, and gives you a result as to where you should dispose of it. It also gives the users an incentive, as by scanning their id, they can be rewarded for correctly sorting between the trash and recycling, further encouraging them to use this system more, leading to a more sustainable future.

How we built it

We built our project using a raspberry pi 4 to host the frontend and back end in addition to the image capture and ML model. The frontend was made using html, css, and javascript while the backend was made using python and flask. The image capture is taken using opencv, and the ML model is a pretrained model from mobilenet. We also used an ultrasonic sensor to check for proximity and an rfid scanner to scan the id that is presented.

Challenges we ran into

Making the whole system work cohesively was a nightmare. The camera did not want to connect and interface with the raspberry pi for the longest time, in addition to tensorflow not working on any version of python other than 3.11, both of which was quite frustrating to deal with. After the whole system was built, taking the picture or reading the id didn't work sometimes, and debugging this also took a lot of time and effort.

Accomplishments that we're proud of

We are proud to have made a Piece of Garbage©, using various different software technologies and hardware, stringing all of them together to make it all work in the end. We also learned a lot, and had a lot of fun working on this project as a team, and are glad to have taken part in this.

What we learned

We learned that hardware is not easy to deal with. Demos usually fail, and it may take a long time to debug what's wrong, for the solution to be just it magically working again sometimes. We also learned various different technologies, like tensorflow, flask, html, and everything else we used to build the Piece of Garbage©, and how to integrate and connect all of these pieces together.

What's next for Piece of Garbage

Making the Piece of Garbage© more reliable, and possibly making an enclosure for the system to make it a more polished product. We can also improve our model, and possibly train our own that can better distinguish between trash and recycling based on the context of what the item is.

Share this project:

Updates