Skip to main content

Civic Eye turns an ordinary smartphone into a pothole survey tool. Mounted like a dashcam, it combines three signals: the camera (YOLOv8 object detection) to see the pothole, the motion sensor to detect the jolt, and GPS to pin the exact location.

A pothole is confirmed only when the camera and the sensor agree within about 1.5 seconds. Repeated jerks without a visible pothole are ignored as speed breakers or rough roads, which reduces false reports. Reports within 5 metres are merged into a single ticket.

Each ticket gets a severity level (Low / Medium / High), an estimated repair cost in rupees, and appears on a live OpenStreetMap map with a heatmap view. Citizens earn credits for confirmed reports, and a Government login lets officers verify tickets and mark them as fixed.

Built as a mobile-friendly web app that runs entirely in the browser, with an optional FastAPI backend for the YOLOv8 detection model.

Built with

pythoncsshtmljavascriptfastapiyolov8roboflowleaflet.jsopenstreetmapgithub pagesultralytics

Prize Categories Entered

  • Deepmind Google

    Best Use of Gemma 4

    Submitted
  • DigitalOcean

    Best Open-Source AI Project

    Submitted