AUTHOR=Khan Malhar , Raza Muhammad Amir , Abbas Ghulam , Othmen Salwa , Yousef Amr , Jumani Touqeer Ahmed TITLE=Pothole detection for autonomous vehicles using deep learning: a robust and efficient solution JOURNAL=Frontiers in Built Environment VOLUME=9 YEAR=2024 URL=https://www.frontiersin.org/journals/built-environment/articles/10.3389/fbuil.2023.1323792 DOI=10.3389/fbuil.2023.1323792 ISSN=2297-3362 ABSTRACT=

Autonomous vehicles can transform the transportation sector by offering a safer and more effective means of travel. However, the success of self-driving cars depends on their ability to navigate complex road conditions, including the detection of potholes. Potholes pose a substantial risk to vehicles and passengers, leading to potential damage and safety hazards, making their detection a critical task for autonomous driving. In this work, we propose a robust and efficient solution for pothole detection using the “you look only once (YOLO) algorithm of version 8, the newest deep learning object detection algorithm.” Our proposed system employs a deep learning methodology to identify real-time potholes, enabling autonomous vehicles to avoid potential hazards and minimise accident risk. We assess the effectiveness of our system using publicly available datasets and show that it outperforms existing state-of-the-art approaches in terms of accuracy and efficiency. Additionally, we investigate different data augmentation methods to enhance the detection capabilities of our proposed system. Our results demonstrate that YOLO V8-based pothole detection is a promising solution for autonomous driving and can significantly improve the safety and reliability of self-driving vehicles on the road. The results of our study are also compared with the results of YOLO V5.