Model Detects as Pepsi
Non recyclable can
Model Detects as Garbage
Why do people not recycle even though the garbage cans always have clear labels? Because it takes time to read it. Every time I stop by a garbage can, I spend a long time deciding which type of garbage my bubble tea cup is. Is it waste or can it be recycled? We decided to solve this problem by allowing people to leave the garbage and just forget about it.
What it does
Our CanSorter sorts the types of waste using image recognition, which means that all you have to do is leave the garbage in the garbage can, and the garbage will go into the appropriate section automatically!
How we built it
Challenges we ran into
One of the main challenges we ran in to is integrating the different parts of the device. We spent a lot of time deciding which interface to use, iOS, Android, etc. But we ended up using webcam because it is the simplest to use given our time constraint. Also, receiving the signal via serial from the cellular IoT device was also difficult as we had no prior experience using the device. Writing the code to send characters and receive and process them was also a difficult task that we could not have done without help as we were not extremely familiar with serial communication and using Cloudant.
Accomplishments that we're proud of
We were able to get the machine learning model up and running and it actually detects cans. We completed the circuit to push the can to the appropriate section and we learned a lot about using cloud services and IoT technology.
What's next for CanSorter
We plan to sort even more categories of garbage in the future as garbage comes in many types, not just two, and also implement the same sorting technology in other fields.