A HACKATHON PROJECT THAT BUILT BY NHAN LE, COLIN ZEE, STEVEN TAT, PRAKHAR AGARWAL
Every year millions if not billions of tons of plastic bottles land up in landfills or oceans and its because people do not sort their plastic bottle waste in proper categories.
There are 2 parts to this - mobile app and the machine. The app lets you take the photograph of the object and it will send the photo to the database which call an API from clarifai and determine which category the bottle falls into and finally returns a result in form of text. The machine is made of a webcam, servo motors and arduino. After you place the bottle in the machine and press a button the webcam will take a picture of the bottle and the process is same as with the app but after the API returns a result the machine will perform a task based on the result- if the bottle falls in to recycle category a servo motor will rotate and put the bottle in recycle bin, if the bottle falls in to trash category another servo motor will rotate and put the bottle in trash bin if it falls in reuse the a led will light up.
CLARIFAI ML API Implemented CV cloud-based AI model Flutter/ Dart Developed UI for taking pictures and making API requests to classifier. Firebase Incorporated Flutter with Firestore to store users’ history images.
Raspberry malfunction. New technology (Flutter, Dart, GCP, CLARIFAI API). Bugs … Relative paths in Flutter, Raspberry PI, data-set labeling, learning curve. Time-constraint for training and labeling data. Collaborating, communication skills, dividing tasks.
Learned how to train a model using ML New technology (Flutter, Dart, GCP, Clarifai) UI/UX design Relative paths in Flutter, Raspberry PI, data-set labeling, learning curve Time-constraint for training and labeling data.
In future we plan to use raspberry pi instead of arduino. Also we wish to use optical sensor which can detect the microorganisms inside a bottle and really differentiate between a reusable bottle and recyclable bottle.
