Real-Time
Road Defect Detection
Live video-based identification of potholes and surface damage
An AI-powered computer vision system that detects road furniture damage and potholes from live video feeds with high precision, enabling smart city monitoring and automated infrastructure maintenance.
Live video-based identification of potholes and surface damage
Deep learning-based object detection pipeline
Supports CCTV, dash cams, and drone inputs
Our solution demonstrates a real-time computer vision system capable of detecting road furniture (traffic signs, signals, street lights, barriers) and road surface defects such as potholes from live video feeds. The system is designed to support smart city monitoring, road safety audits, and automated infrastructure maintenance workflows.
Manual road inspection is slow, expensive, and inconsistent. Undetected potholes and damaged road furniture increase accident risk, vehicle damage, and maintenance delays. Traditional monitoring lacks real-time automation and scalability.
Developed a deep learning-based object detection pipeline (YOLO-based) that identifies multiple road furniture classes and potholes from video streams. The system performs real-time detection, severity estimation, and structured reporting for scalable infrastructure monitoring.
