TITLE:
A Conceptual Framework for a Lightweight System Integrated into Vehicles for Real-Time Road Surface Monitoring Using Vehicle-Mounted Vision Systems and Communication
AUTHORS:
Ngai Chi Yeung Ernest, Wong Boonyapat, Cheng Kin Tat Kinder, Lo Chiu Kit, Yan Kim
KEYWORDS:
Infrastructure Safety, Autonomous Road Inspection, Lightweight Embedded Systems, Data Sharing in Vehicular Networks, YOLO
JOURNAL NAME:
Journal of Computer and Communications,
Vol.13 No.11,
November
21,
2025
ABSTRACT: Road surface defects such as potholes and cracks pose significant risks to traffic safety and vehicle integrity. Traditional inspection methods, including manual surveys and citizen reports, are often inefficient, inconsistent, and lack real-time responsiveness. This paper proposes a conceptual framework for a lightweight, vehicle-integrated system that enables real-time road surface monitoring using a vision-based approach. The system utilizes consumer-grade hardwar-specifically Raspberry Pi and dashcams-combined with the YOLOv8 object detection model to identify road anomalies at high speeds. It incorporates GPS tagging and collaborative data sharing to alert nearby vehicles of detected hazards, enhancing driver awareness and safety. Additionally, the system integrates vehicle dynamics, such as suspension bounce, to improve detection accuracy and supports reinforcement learning through continuous data collection. Preliminary results, based on training with public datasets from Hong Kong and open-source repositories, demonstrate a recognition success rate exceeding 92%. This framework offers a scalable, cost-effective solution for intelligent road monitoring and lays the groundwork for future development through expanded data collection and deployment.