Inspiration

Shooting stars happen more frequently than most people think. Approximately five of them can be seen anywhere in the night sky per hour, but they're easy to miss in the blink of an eye. For space enthusiasts, Sky Eye is a novel tool that mitigates this problem and enhances the stargazing experience.

What it does

Sky Eye detects shooting stars from any user-input video or livestream footage. The input is sent from the sleek web-based frontend to the Python backend, which uses computer vision techniques to recognize and track moving celestial objects. Bright bounding boxes around shooting stars are overlaid onto the original video and shown to the user's device.

How we built it

We used React to build the frontend and Python/Flask on the backend. For video processing and motion detection, we implemented methods from OpenCV, a popular computer vision library.

Challenges we ran into

As mostly first-year students, we had to learn a lot of the techniques used in the project on-the-fly.

Accomplishments that we're proud of

We are proud of what we could accomplish in only 1 day. We had an idea and we went for it! Technical skill wise, we’re most proud of the shooting star detection and the connection between the frontend and backend.

What we learned

We learned about full stack development, motion detection techniques, team-based skills such as communication and management and version control.

What's next for Sky Eye

Our plans for Sky Eye is sky high! We plan to include features to improve the video processing and add additional features to aid in the stargazing experience. We hope to implement a feature to save time stamps when movement is detected, object classification to be able to differentiate a shooting star from other noise, background stabilization, connectivity to external cameras, phone compatibility, and much more is to come.

Share this project:

Updates