Inspiration
Wanted to create something fun yet practical
What it does
Originally meant to be some sort of recycling bot that got out of hand. Very sus.
Robot wanders randomly and stops when it spots something sus such as an aluminum can (Pepsi or MountainDew can). When a sus object is detected, it alerts its user through a text and alerts the people around the bot through informative audio.
How we built it
Robot was built using a standard chassis robot kit, an L298 Motor driver, Lipo batteries, Stereovision Camera, and a Speaker. All written in Python and controlled through a wireless keyboard. The Computer Vision object detection was built using tensorflow and openCV. The image processing was built with openCV.
Challenges we ran into
Installing tensorflow was difficult, had to make sure the correct parallel processing models were compatible with our GPU and Processor. Many of the installs were very long and ate a lot of time and some had to be reinstalled. Cutting into development time. Multithreading with Python. Streaming video across a network.
Accomplishments that we're proud of
The ability to use Twilio and project sound. Live image processing and some object recognition capability.
What we learned
Patience
What's next for Sussy bot
Combining the Object Detection, sound bytes, and the Bot together.
Log in or sign up for Devpost to join the conversation.