At many public places, recycling is rarely a priority. Recyclables are disposed of incorrectly and thrown out like garbage. Even here at QHacks2017, we found lots of paper and cans in the garbage.

What it does

The Green Waste Bin is a waste bin that can sort the items that it is given. The current of the version of the bin can categorize the waste as garbage, plastics, or paper.

How we built it

The physical parts of the waste bin are the Lego, 2 stepper motors, a raspberry pi, and a webcam. The software of the Green Waste Bin was entirely python. The web app was done in html and javascript.

How it works

When garbage is placed on the bin, a picture of it is taken by the web cam. The picture is then sent to Indico and labeled based on a collection that we trained. The raspberry pi then controls the stepper motors to drop the garbage in the right spot. All of the images that were taken are stored in AWS buckets and displayed on a web app. On the web app, images can be relabeled and the Indico collection is retrained.

Challenges we ran into

AWS was a new experience and any mistakes were made. There were some challenges with adjusting hardware to the optimal positions.

Accomplishments that we're proud of

Able to implement machine learning and using the Indico api Able to implement AWS

What we learned

Indico - never done machine learning before AWS

What's next for Green Waste Bin

Bringing the project to a larger scale and handling more garbage at a time.

Share this project: