Pothole Detection and Dimension Measuring

Potholes and driver drowsiness are two major causes of road accidents, leading to vehicle damage, traffic delays, and loss of human life. Traditional road inspection methods are manual, time-consuming, and inefficient, while existing driver monitoring systems often fail to provide timely warnings. To address these issues, this project presents a real-time intelligent road safety monitoring system that detects potholes and driver fatigue simultaneously. The system uses the YOLOv8 deep learning model for accurate pothole detection from images, videos, and live camera feed. Driver drowsiness and yawning detection is performed using facial landmark analysis by calculating Eye Aspect Ratio (EAR) and Mouth Aspect Ratio (MAR). The system generates instant visual and voice alerts through the in-vehicle display, and also provides an emergency SMS notification feature for critical situations. A web-based interface is implemented using Flask and OpenCV for monitoring, report generation, and data management. Experimental results show that the proposed system provides efficient realtime performance with reliable detection accuracy, making it suitable for smart transportation and road safety applications.