Inspiration

Our concept for our project is a secure/smart trash bin that educates and prevents people from throwing waste into the wrong category (trash, recycling, compost). Landfills such as the Pacific Garbage Patch are not getting smaller - becoming increasing environmental hazards to climate and ecosystems with each day. In addition, with increasingly strict requirements on recycling from foreign buyers of recycling such as China, most of the unbought recycling or recycling deemed “over-contaminated” are left to the landfill. It is imperative that we as humans do as much as we can to limit the negative impacts of our waste on the environment for the sake of our future generation.

What it does

Our trash bin works by having users hold their waste in front of a top mounted camera. The bin will detect the object and open the lid to the bin the user should dispose of their waste (trash, recycling, compost). The bin uses the Raspberry Pi microcontroller for object recognition and camera function using OpenCV. For the opening and closing of the trash bins, the Arduino Uno is utilized.

How I built it

We originally tried using machine learning with TensorFlow and YOLO v3 for object recognition but our restriction to 4 GB of memory on the Raspberry Pi left us inadequate space to install these necessary libraries for use on the microcontroller. Next, we tried image comparison using Google Cloud but ran into difficulties using the key for the Google API to run our program and get access to the database. Our last resort was finding an alternative to object recognition, stumbling across a machine learning algorithm using OpenCV. We were able to get the object recognition to work on a laptop but were unable to get the camera module to function on the Raspberry Pi before our 24 hour deadline was up.

Challenges I ran into

Next, we tried image comparison using Google Cloud but ran into difficulties using the key for the Google API to run our program and get access to the database. Our last resort was finding an alternative to object recognition, stumbling across a machine learning algorithm using OpenCV. We were able to get the object recognition to work on a laptop but were unable to get the camera module to function on the Raspberry Pi before our 24 hour deadline was up.

Accomplishments that I'm proud of

However, our working project is a functioning bin with 3 buttons that correspond to opening the trash, recycling, and compost bins. Even though this may not be as secure/smart as we originally hoped, this trash bin still helps society by forcing the user to make a conscious decision about what they are throwing away and how their disposal decisions impact the environment.

What I learned

The difficulty of learning new concepts in the short amount of time that is allocated and utilizing the resources that we are given.

What's next for The Green Machine

Built With

Share this project:
×

Updates