— Potholes on road surfaces pose significant risks to vehicular safety and contribute to the deterioration of transportation infrastructure. Traditional manual inspection methods are inefficient, time-consuming, and fail to provide timely interventions. This research presents an automated pothole detection and mapping system utilizing the YOLOv4-tiny deep learning model for real-time object detection. The system processes images and video streams to accurately identify potholes and stores detection results in an SQLite database for structured data management. A web-based interface facilitates user interaction for location selection and visualization of detected potholes. The proposed solution emphasizes modularity, enabling easy updates, retraining, and deployment on resource-constrained devices. Experimental evaluations demonstrate the model's high accuracy and real-time performance, indicating the system’s potential for large-scale deployment in urban infrastructure maintenance programs.
📖 افتح في inklap 🔗 DOI 📮 اطلب بحثاً