Inspiration
We were inspired by the everyday chaos of busy streets and public spaces and the lack of real-time tools to make sense of it. We wanted to build something that bridges the gap between AI and practical, real-world impact by turning a simple camera feed into actionable insight about how people and vehicles move through the world.
What it does
It analyzes real-time webcam footage using computer vision to detect and count people and vehicles as they move through a scene. The processed data is then presented through interactive heatmaps and visualizations, giving a glance at picture of crowd and traffic density over time.
How we built it
We used Python as the backbone of the project, leveraging YOLO (You Only Look Once) for real time object detection on live webcam feeds. YOLO allowed us to accurately identify and track people and transport in each frame with impressive speed. We then used Claude's assistance to build a clean, user-friendly interface in HTML that displays the detection data as heatmaps and visual overlays.
Challenges we ran into
One team member didn't have their laptop available, so they shifted focus to ideation and support rather than hands-on coding. We also ran into difficulties sourcing different test footage and getting various video inputs to work reliably with our pipeline.
Accomplishments that we're proud of
We collaborated really effectively as a team despite having only just met for the first time. We're also proud of how well the final product came together, the real-time detection and visualization worked better than we expected given the time constraints.
What we learned
We learned how to integrate YOLO into a live video pipeline, and gained hands-on experience translating raw detection data into meaningful visual outputs like heatmaps. Beyond the technical side, this hackathon showed us how to collaborate efficiently with people we'd never met before, such as dividing responsibilities on the fly and adapting when things didn't go to plan. It also reinforced how powerful modern AI tools can be when applied to real-world problems. Something which once would have required expensive infrastructure can now be prototyped in a single day with a laptop and the right model.
Log in or sign up for Devpost to join the conversation.