Smart Pothole Detection

Project details

Smart Pothole Detection

End-to-End AI Road Monitoring Platform

A full-stack road-monitoring platform that runs a recall-tuned YOLOv8n model directly on Android for real-time pothole detection, then sends GPS-tagged evidence to a .NET and PostGIS backend. A web dashboard gives road authorities the tools to map, review, manage, and analyze reported defects.

Key features

• Real-time on-device YOLOv8n inference with TensorFlow Lite
• Native Kotlin and CameraX app with GPS evidence and offline sync
• ASP.NET Core 8 API with JWT, EF Core, and PostgreSQL/PostGIS
• React operations dashboard with maps, analytics, and report workflows
• Dockerized VPS deployment with Nginx, SSL, and S3-compatible storage

Tools used

KotlinTensorFlow LiteASP.NET CorePostgreSQLReactDockerGithub