Inspiration
Our team has important values when it comes to designing and creating software. Our goal is to always help people who need it the most. We realized that many visually impaired people struggle with walking around or traveling by themselves, so this is the problem we decided to tackle.
What it does
SafeWalker is a safety-oriented application designed to enhance pedestrian awareness and safety for the visually impaired.
How we built it
This script utilizes YOLOv8 and OpenCV to detect traffic signs (pedestrian signals) and vehicles from a video feed. It analyzes the colors of pedestrian signals—red indicating "Do not cross the road" and green indicating "You may cross" — and provides voice alerts for both pedestrian signals and large vehicles passing in front of the client detected within the frame.
Challenges we ran into
Detecting the difference between a traffic light and a pedestrian walking light.
Accomplishments that we're proud of
Many successful test runs.
What we learned
Implementing pre-trained image detection models.
What's next for SafeWalker
Create a front end using React for SafeWalker, then implement Google Maps API to help visually impaired people get to where they need to be safely.
Log in or sign up for Devpost to join the conversation.