Effective waste management is one of the simplest, most convenient ways to lessen our impact on the environment within our own homes. However, waste contamination (where one category of waste is placed in another by the consumer) is still much more commonplace than it should be, resulting in negative environmental and economic consequences that could be easily preventable with the automation of waste management.
What it does
BitToBin, using machine learning and object recognition, sorts and deposits waste into compost, recycling and garbage bins automatically as each waste item is placed on the opening platform.
How we built it
A plywood-based framework to connect all hardware components was built. The Clarifai API was trained to recognize waste with image classes of various waste items. Open CV & webcam were set up to take pictures. Servos were programmed to rotate & open flaps with an Arduino. The abilities to recognize images of the webcam and to open flaps of the servos were integrated with one another using an Arduino.
Challenges we ran into
Integrating all the hardware components with one another at the end; Servo positioning & delay errors; Cable management
Accomplishments that we're proud of
Figuring out how to get the servos to move in time and to the right end position to have waste items easily drop into their proper bin
What we learned
When working in teams, effective communication can give one more of an advantage than technology.
What's next for BitToBin
Assigning different classes for each province's waste management regulations. Otherwise, more identifiable classes of waste items, better appearance, stronger material.
Check us out! link