Road potholes are a major concern for transportation systems as they increase the risk of road accidents, vehicle damage, traffic congestion, and maintenance costs. Traditional road inspection methods rely on manual observation, which is time-consuming, labor-intensive, and often unable to detect potholes efficiently. To address these challenges, this project presents an intelligent pothole detection system based on the YOLOv8 deep learning model. The proposed system automatically identifies potholes from road images with high accuracy and real-time performance. A publicly available pothole dataset from Kaggle is used for training and testing the model, while Label Studio is utilized to annotate the images in YOLO format. The trained YOLOv8 model accurately detects potholes by generating bounding boxes around damaged road regions. The model achieved a mean Average Precision (mAP@0.5) of 0.779 and a mean Average Precision (mAP@0.5:0.95) of 0.512, demonstrating effective detection accuracy and reliable localization under varying road conditions. The model also provides fast inference speed, making it suitable for intelligent transportation systems and smart city infrastructure. The proposed system can assist government authorities, municipal corporations, and road maintenance departments in monitoring road conditions and scheduling timely repairs. Early pothole detection helps reduce vehicle damage, improve road safety, minimize maintenance costs, and enhance driving comfort. Furthermore, the lightweight architecture of YOLOv8 enables deployment on edge devices and real-time surveillance systems, making the solution cost-effective and scalable. Overall, this project demonstrates the effectiveness of deep learning and computer vision in developing an automated pothole detection system that supports safer roads, efficient maintenance planning, and improved transportation infrastructure.

Sarnitha D, Monisha EJ, Rithikaa B, Sharath K, Ms Dhanya K R,"AI-Powered Yolov8-Based Intelligent Pothole Detection and Road Condition Monitoring System" International Journal of Advanced Multidisciplinary Research and Educational Development, V2(4): Page(408-412) July - August 2026. ISSN: 3107-6513. www.ijamred.com. Published by Scientific and Academic Research Publishing.