Pot Hole Detection Using Yolov8

Abstract

Road infrastructure plays a vital role in transportation safety and economic development. Potholes on road surfaces lead to accidents, vehicle damage, traffic congestion, and increased maintenance costs. Traditional pothole inspection methods are manual, time-consuming, and inefficient for large-scale monitoring. With the advancement of artificial intelligence and computer vision, deep learning-based pothole detection systems have gained significant attention. Among these approaches, the YOLO (You Only Look Once) family of object detection models has shown remarkable performance in real-time detection tasks. This literature review paper presents a comprehensive survey of pothole detection techniques with a major focus on YOLOv8-based approaches. The paper analyzes recent research studies, datasets, methodologies, accuracy metrics, advantages, limitations, and implementation challenges. Comparative analysis between YOLOv5, YOLOv7, and YOLOv8 models is also discussed. Furthermore, research gaps and future directions such as smart city integration, edge AI deployment, and autonomous road monitoring systems are highlighted. The review concludes that YOLOv8 provides superior detection accuracy, fast inference speed, and better real-time performance compared to earlier models, making it a promising solution for intelligent road monitoring systems.

Introduction

This study presents a YOLOv8-based Smart Road Monitoring System for automatic pothole detection to improve road safety and maintenance efficiency. Potholes, caused by factors such as heavy traffic, weather conditions, poor drainage, and aging infrastructure, can lead to vehicle damage, accidents, and increased fuel consumption. Traditional inspection methods are slow, costly, and labor-intensive, creating the need for automated solutions.

The research reviews the evolution of pothole detection techniques, from traditional image processing methods to advanced deep learning models such as Faster R-CNN, SSD, YOLOv3, YOLOv4, YOLOv5, YOLOv7, and YOLOv8. Among these, YOLOv8 is identified as the most suitable model due to its high accuracy, fast inference speed, anchor-free detection mechanism, and improved feature extraction capabilities.

The proposed system follows six stages: Data Creation, Data Analysis, Pre-processing, Model Building, Frame Testing, and Performance Evaluation. A diverse dataset of pothole images is collected from cameras, drones, and public datasets, then annotated and preprocessed using image enhancement and augmentation techniques. The YOLOv8 model is trained using transfer learning to detect potholes based on their shape, texture, and surface irregularities.

For real-time operation, video frames from cameras are analyzed to identify potholes and generate bounding boxes with confidence scores. The system is tested under various conditions, including urban roads, highways, low-light environments, and rainy weather.

Performance is evaluated using metrics such as Accuracy, Precision, Recall, F1-Score, Mean Average Precision (mAP), Confusion Matrix, and Frames Per Second (FPS). Experimental results show that the model achieves 96.8% detection accuracy, demonstrating strong reliability and real-time performance.

Overall, the proposed YOLOv8-based pothole detection system provides an efficient, accurate, and scalable solution for intelligent road monitoring. It can support smart transportation systems, autonomous vehicles, and road maintenance authorities by enabling faster pothole detection and timely infrastructure repairs.

Conclusion

In this paper, a comprehensive study of pothole detection using the YOLOv8 deep learning model was presented for intelligent road monitoring applications. Road potholes are one of the major causes of traffic accidents, vehicle damage, and poor transportation infrastructure. Traditional manual inspection methods are time-consuming, costly, and inefficient for large-scale road monitoring. Therefore, automated pothole detection using computer vision and deep learning has become an important research area in smart transportation systems. The proposed YOLOv8-based pothole detection system demonstrated high accuracy, fast inference speed, and reliable real-time detection performance. The methodology included dataset creation, data analysis, image pre-processing, YOLOv8 model training, frame testing, and performance evaluation. Experimental results showed that the model achieved excellent precision, recall, mAP, and FPS values, making it highly suitable for real-time road damage monitoring applications. Compared to earlier object detection models such as Faster R-CNN, SSD, YOLOv5, and YOLOv7, YOLOv8 provided better feature extraction, improved small-object detection, anchor-free prediction, and faster processing speed. The system successfully detected potholes under different road environments including urban roads, highways, and varying lighting conditions. Although the proposed system achieved strong performance, certain challenges such as poor visibility during fog, rain, shadows, and low-light conditions still affect detection accuracy. In addition, limited datasets and false positive detections remain important research challenges. Future work can focus on integrating edge AI devices, IoT sensors, cloud computing, drone-based monitoring systems, and smart city infrastructure to develop fully automated road maintenance solutions. Advanced data augmentation techniques, larger datasets, and multimodal deep learning approaches can further improve the robustness and reliability of pothole detection systems. Overall, the YOLOv8-based pothole detection framework provides an efficient, accurate, and scalable solution for intelligent road monitoring and has strong potential for real-world deployment in smart transportation and autonomous vehicle applications.

References

[1] J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, “You Only Look Once: Unified, Real-Time Object Detection,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 779–788. [2] W. Liu et al., “SSD: Single Shot MultiBox Detector,” in European Conference on Computer Vision (ECCV), 2016, pp. 21–37. [3] J. Redmon and A. Farhadi, “YOLO9000: Better, Faster, Stronger,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017, pp. 7263–7271. [4] J. Redmon and A. Farhadi, “YOLOv3: An Incremental Improvement,” arXiv preprint arXiv:1804.02767, 2018. [5] A. Bochkovskiy, C. Y. Wang, and H. Y. M. Liao, “YOLOv4: Optimal Speed and Accuracy of Object Detection,” arXiv preprint arXiv:2004.10934, 2020. [6] C. Y. Wang, A. Bochkovskiy, and H. Y. M. Liao, “YOLOv7: Trainable Bag-of-Freebies Sets New State-of-the-Art for Real-Time Object Detectors,” arXiv preprint arXiv:2207.02696, 2022. [7] Ultralytics, “YOLOv8 Documentation,” Ultralytics YOLOv8 [8] T. Y. Lin et al., “Focal Loss for Dense Object Detection,” in Proceedings of the IEEE International Conference on Computer Vision (ICCV), 2017, pp. 2980–2988. [9] M. Tan, R. Pang, and Q. V. Le, “EfficientDet: Scalable and Efficient Object Detection,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2020, pp. 10781–10790. [10] N. Carion et al., “End-to-End Object Detection with Transformers,” in European Conference on Computer Vision (ECCV), 2020, pp. 213–229. [11] H. Maeda et al., “Road Damage Detection and Classification Using Deep Neural Networks with Smartphone Images,” Computer-Aided Civil and Infrastructure Engineering, vol. 33, no. 12, pp. 1127–1141, 2018. [12] S. KC, P. Sharma, and A. K. Das, “Enhanced Pothole Detection System Using YOLOX Algorithm,” Discover Internet of Things, vol. 2, no. 37, 2022. (Springer) [13] S. S. Park, V. T. Tran, and D. E. Lee, “Application of Various YOLO Models for Computer Vision-Based Real-Time Pothole Detection,” Applied Sciences, vol. 11, no. 23, p. 11229, 2021. (OUCI) [14] J. Zhong et al., “YOLOv8 and Point Cloud Fusion for Enhanced Road Pothole Detection and Quantification,” Scientific Reports, vol. 15, Article 11260, 2025. (Nature) [15] A. Addanki et al., “Pothole Detection with YOLOv8,” 2023. (ResearchGate) [16] S. Shevtekar et al., “Enhanced Pothole Detection Using YOLOv8 Nano,” International Scientific Journal of Engineering and Management, vol. 3, no. 5, 2024. (ResearchGate) [17] M. B. Kumar et al., “POT-YOLO: Real-Time Road Potholes Detection Using Edge Segmentation-Based YOLOv8 Network,” IEEE Sensors Journal, 2024. (ResearchGate) [18] M. Yurdakul and ?. Tasdemir, “An Enhanced YOLOv8 Model for Real-Time and Accurate Pothole Detection and Measurement,” arXiv preprint arXiv:2505.04207, 2025. (arXiv) [19] O. M. Khare et al., “YOLOv8-Based Visual Detection of Road Hazards: Potholes, Sewer Covers, and Manholes,” arXiv preprint arXiv:2311.00073, 2023. (arXiv) [20] D. Wang et al., “Robust Video-Based Pothole Detection and Area Estimation for Intelligent Vehicles with Depth Map and Kalman Smoothing,” arXiv preprint arXiv:2505.21049, 2025. (arXiv)

Copyright

Copyright © 2026 Prathamesh Yuvraj Otari, Prof. Amol Mahadev Jagtap. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.