DetectMe | Devpost

DetectMe

Inspiration

The beauty and mystery of the night sky have always inspired humanity to look upward. DetectMeteor was born from the desire to make meteor tracking more accessible to enthusiasts and amateurs alike. By combining computer vision with simple tools, we aimed to create a real-time system that could help users identify meteors effortlessly while encouraging exploration of the cosmos.

What it does

It's a system that uses computer vision to detect meteors in real-time. It identifies bright moving objects in video feeds, highlights them with bounding boxes, and logs timestamps for each detection. An accessible tool for sky enthusiasts to explore meteor tracking.

How we built it

Hardware: Webcam for live video capture. Software: Computer vision techniques for processing video feeds. Frame Processing: Converted video frames to grayscale and applied Gaussian blur to reduce noise. Object Detection: Used brightness thresholds and contour detection to isolate potential meteors. Logging and Visualization: Highlighted detected meteors with bounding boxes and logged timestamps for events.

Challenges we ran into

Noise and False Positives: Differentiating meteors from other bright objects like stars and planes was a key challenge. Parameter Tuning: Optimizing detection thresholds for varied lighting conditions took significant experimentation. Hardware Issues: Ensuring webcam compatibility and smooth frame capture required troubleshooting.

What we learned

Computer Vision Basics: Using techniques like thresholding, contour detection, and motion tracking to identify objects in video feeds. Optimization: Fine-tuning parameters to reduce noise and improve detection accuracy under various lighting conditions. Debugging: Troubleshooting hardware and software compatibility, especially when working with live video feeds. Resilience: Tackling challenges taught us patience and adaptability.