Edge-AI Enabled Pothole Detection and Road Quality Mapping System for Indian Urban Roads

Objective: To design and implement an edge-computing based AI system for real-time pothole detection and road-quality mapping using smartphone sensors and embedded devices, with centralized data aggregation for civic authorities.

Potholes and poor road quality are persistent issues affecting urban and rural commuters, vehicle owners, and municipal bodies across India, leading to accidents, vehicle damage, and inefficient road maintenance cycles.

This project proposes an integrated engineering solution combining edge-AI algorithms deployed on smartphones or Raspberry Pi devices with cloud-based aggregation and visualization. Using computer vision on dashcam video and sensor fusion from smartphone IMU/GPS, the system detects potholes, classifies road quality, tags GPS location, and uploads results to a central map dashboard accessible by municipal authorities.

The deliverables include: (1) A mobile app or embedded device prototype for edge inference, (2) a backend with REST APIs and a web dashboard for live road-quality maps, (3) dataset collection and model training using Indian road imagery (e.g., Indian Driving Dataset, MIVIA), (4) cost analysis for scalable deployment, and (5) a live field-tested demonstration on local roads.

This solution enables proactive road maintenance, reduces manual surveys, improves commuter safety, and supports scalable adoption by smart city initiatives and civic agencies across India.