RoadEvidence

RoadEvidence dashboard — a unified workspace for road inspections, image analysis, records, and road condition monitoring.

Inspiration

Road damage can be difficult to document and monitor through manual inspections. We wanted to build a solution that uses computer vision to help organize road inspection evidence and identify areas that may need attention.

What it does

RoadEvidence is a road inspection platform designed to analyze road images and organize inspection records. Its dashboard includes sections for new inspections, records, comparisons, and reports, helping users manage road condition evidence in one place.

How we built it

We built a web-based dashboard with a Python backend and OpenCV 5 for image processing. The intended workflow is to upload a road image, analyze its visual features, and display candidate damage regions for human review. Our project also explores AWS integration for cloud-based capabilities.

Challenges we ran into

Integrating the frontend and backend was a major challenge. We also considered how shadows, lighting, road markings, and surface textures can affect computer vision results. Making the interface clear and the inspection workflow reliable was another important part of development.

Accomplishments that we're proud of

We developed the RoadEvidence dashboard and brought together the main sections of a road inspection workflow in one interface. We are also proud of exploring how OpenCV-based analysis can support the organization and review of road inspection evidence.

What we learned

We learned about computer vision, web development, backend integration, and the importance of testing a complete application. We also learned that automated image analysis can produce false positives and should assist human inspectors rather than replace them.

What's next for RoadEvidence

Our next goals are to improve image-processing reliability, validate the complete upload-to-results workflow, strengthen repeat-inspection comparisons, and verify AWS integration. We aim to make RoadEvidence a more reliable tool for documenting and monitoring road conditions.

Limitations

Detected regions are candidates for review, not confirmed road damage. Results require human verification, and the system needs further testing across different road surfaces, lighting conditions, and environments before real-world deployment.