What it does
It sorts round and slightly larger-than-average candy based on color.
How we built it
With tender love and care ♥ We used resources from the Robotics Team in NJIT and from the Hardware Lab in the hackathon to build the machine. It was heavily jerry-rigged with melted plastic, glue, and lots of tape. The motors were attached to our Arduino controller through an assortment of wires attached to a bread board. The programs were written in Arduino, which is based off of C++. The camera is done with a Raspberry Pi that has Raspbian loaded into it. It takes an image and uploads it to Google Cloud for computing, and receives it again. Google Cloud's Compute Engine is used for the Tensorflow calculations, done in Tensorflow with Python. After training the model, the program waits on data to be put in and will return a string of the color of the selected object.
Challenges we ran into
Not having some parts. Getting locked out of the robotics room for the first six hours. A severe lack of SD cards. Nonexistent background knowledge. Internet issues with raspberry pi
Accomplishments that we're proud of
What we learned
Tensorflow. Google Compute Engine. Arduino. Raspbian. Breadboards.
What's next for Chad
probably sleep for me Chad might "steal yo' girl"