International Journal of Engineering and Manufacturing (IJEM)
IJEM Vol. 16, No. 3, 8 Jun. 2026
Cover page and Table of Contents: PDF (size: 2650KB)
Deep Learning-Based Pothole Detection Techniques under Multiple Weather Conditions
PDF (2650KB), PP.238-253
Views: 0 Downloads: 0
Author(s)
Henry Nii-Armah Mettle
1,*
Peter Appiahene
2
Michael Opoku
2
Index Terms
Pothole detection, YOLOv5, Single Shot Detector (SSD), Deep learning, Multi-Weather Dataset, Road Maintenance, Object Detection
Abstract
Potholes are a major concern for road infrastructure, traffic safety, and vehicle maintenance. Manual inspection methods for pothole detection are labor-intensive, time-consuming, and often inefficient for large road networks. This study evaluates and compares the performance of YOLOv5 and Single Shot Detector (SSD) models for automated pothole detection under diverse weather and lighting conditions. Using the Multi-Weather-Based Dataset (MWBD), images captured during daytime, twilight, and nighttime were annotated with bounding boxes and enhanced through data augmentation techniques such as shearing and flipping. Experimental results indicate that YOLOv5 achieves a precision of 92.2%, recall of 89.2%, F1-score of 90.7%, and mAP@0.5 of 90.0%, while SSD achieves a precision of 88.5%, recall of 92.0%, F1-score of 90.2%, and mAP@0.5 of 91.4%. The comparative analysis demonstrates that both models are effective in detecting potholes across varied road textures and environmental conditions, with trade-offs between precision and recall. This study highlights the suitability of deep learning-based object detection models for automated road inspection, reducing human effort, enhancing maintenance efficiency, and improving road safety. The novelty lies in the systematic comparison of YOLOv5 and SSD under multi-weather conditions, providing practical guidance for intelligent transportation systems.
Cite This Paper
Henry Nii-Armah Mettle, Peter Appiahene, Michael Opoku, "Deep Learning-Based Pothole Detection Techniques under Multiple Weather Conditions", International Journal of Engineering and Manufacturing (IJEM), Vol.16, No.3, pp.238-253, 2026. DOI:10.5815/ijem.2026.03.14
Reference
[1]X. Chen and H. Wang, “Hydro-mechanical analysis of water-induced pothole patch failure in asphalt pavement,” Constr. Build. Mater., vol. 413, p. 134767, 2024, doi: 10.1016/j.conbuildmat.2023.134767.
[2]T. Neumann, “Analysis of advanced driver-assistance systems for safe and comfortable driving of motor vehicles,” Sensors, vol. 24, no. 19, p. 6223, 2024, doi: 10.3390/s24196223.
[3]M. Hu and Q. Xia, “Pothole detection and recognition based on transfer learning,” arXiv preprint, 2025, doi: 2509.06750.
[4]B. Leelavathy, R. M. R. Kovvur, K. M. Z. Parvez, B. A. Goud, and R. K. Chenoori, “Real-time pothole detection system: A deep learning approach with SSD,” in Big Data and Artificial Intelligence, BDA 2024, A. Dasgupta et al., Eds., Lecture Notes in Computer Science, vol. 15526, pp. 1–13, Springer, 2025, doi: 10.1007/978-3-031-81821-9_27.
[5]S. Chen, D. F. Laefer, X. Zeng, L. Truong-Hong, and E. Mangina, “Volumetric pothole detection from UAV-based imagery,” J. Surv. Eng., vol. 150, no. 2, p. 05024001, 2024, doi: 10.1061/JSUED2.SUENG-1458.
[6]Y. Liao et al., “YOLO-MECD: Citrus detection algorithm based on YOLOv11,” Agronomy, vol. 15, no. 3, p. 687, 2025, doi: 10.3390/agronomy15030687.
[7]Y. Feng et al., “Spatial and momentum mapping modes for velocity map imaging spectrometer,” Appl. Sci., vol. 14, no. 5, p. 2190, 2024, doi: 10.3390/app14052190.
[8]J. Chen and M. J. Er, “Dynamic YOLO for small underwater object detection,” Artif. Intell. Rev., vol. 57, p. 165, 2024, doi: 10.1007/s10462-024-10788-1.
[9]M. G. Ragab et al., “A comprehensive systematic review of YOLO for medical object detection (2018 to 2023),” IEEE Access, vol. 12, pp. 57815–57836, 2024, doi: 10.1109/ACCESS.2024.3386826.
[10]G. Mahesh, R. S. Shankar, V. V. R. M. Rao, and S. N., “An object detection framework and deep learning models used to detect the potholes on the streets,” in 2024 Int. Conf. Advances in Modern Age Technol. Health Eng. Sci. (AMATHE), pp. 1–7, IEEE, 2024, doi: 10.1109/AMATHE61652.2024.10582091.
[11]P. Chandak, S. Chaurasia, and M. Patil, “Identification of potholes and speed bumps using SSD and YOLO,” in Intelligent Computing and Networking, IC-ICN 2024, G. Ghinea et al., Eds., Lecture Notes in Networks and Systems, vol. 1172, pp. 1–10, Springer, 2025, doi: 10.1007/978-981-97-8631-2_1.
[12]I. Almasri, D. Manasreh, and M. D. Nazzal, “AI meets ADAS: Intelligent pothole detection for safer AV navigation,” Vehicles, vol. 7, no. 4, p. 109, 2025, doi: 10.3390/vehicles7040109.
[13]J. Mehta, V. Mathur, D. Agarwal, A. Sharma, and K. Prakasha, “Pothole detection and analysis system (PoDAS) for real-time data using sensor networks,” arXiv preprint, 2025, doi: 2508.16626.
[14]H. Yang, Y. Song, Y. Liang, E. Tang, and D. Cao, “SDC-YOLOv8: An improved algorithm for road defect detection through attention-enhanced feature learning and adaptive feature reconstruction,” Sensors, vol. 26, no. 2, p. 609, 2026, doi: 10.3390/s26020609.
[15]W. D. Martins, F. Santos Osorio, and D. R. Bruno, “Real-time deep learning-based pothole detection on low-cost embedded devices for ADAS,” in Proc. Brazilian Conf. Robotics (CROS), pp. 1–6, 2025, doi: 10.1109/CROS66186.2025.11066083.
[16]D. V. Agrawal, V. Gupta, C. R. Krishna, M. Chopra, and A. Puri, “Label-efficient self-supervised representation learning-based road pothole detection from thermal images,” Eng. Sci., vol. 36, p. 1626, 2025, doi: 10.30919/es1626.
[17]Y. Duan et al., “Adversarial patches for fooling positive–negative obstacle detectors on unstructured roads,” IEEE Sens. J., vol. 25, no. 19, pp. 36466–36478, 2025, doi: 10.1109/JSEN.2025.3586858.
[18]V. G. Rao Alleshwaram, S. C. Kumar, C. K. Reddy, K. Butchi Raju, V. Anitha, and A. V. Rao, “Pothole detection using deep learning,” in Proc. 8th Int. Conf. Comput. Methodol. Commun. (ICCMC), pp. 2031–2037, 2025, doi: 10.1109/ICCMC65190.2025.11140736.
[19]S. Sivanraj, D. N. L. S. Uduwage, R. K. Shiwakoti, and M. Thripathi, “YOLOv11-s-based pothole detection model with integrated traffic data for maintenance decision support,” J. Adv. Coll. Eng. Manage., vol. 11, no. 1, 2025, doi: 10.3126/jacem.v11i1.84529.
[20]S. Parvin, F. Munsy, and K. Nur, “Multi-Weather Pothole Detection (MWPD) [Data set],” Mendeley Data, 2025, doi: 10.17632/s5hx9n2jc3.2.
[21]S. U. Ahmed et al., “Pothole detection using computer vision and Raspberry Pi,” J. Comput. Biomed. Inform., vol. 7, no. 01, pp. 690–695, 2024.
[22]A. Paramarthalingam, J. Sivaraman, P. Theerthagiri, B. Vijayakumar, and V. Baskaran, “A deep learning model to assist visually impaired in pothole detection using computer vision,” Decision Anal. J., vol. 12, p. 100507, 2024, doi: 10.1016/j.dajour.2024.100507.
[23]C. Ruseruka, J. Mwakalonge, G. Comert, S. Siuhi, F. Ngeni, and Q. Anderson, “Augmenting roadway safety with machine learning and deep learning: Pothole detection and dimension estimation using in-vehicle technologies,” Mach. Learn. Appl., vol. 16, p. 100547, 2024, doi: 10.1016/j.mlwa.2024.100547.
[24]S. Saleh, A. Jolfaei, and M. Tariq, “Real-time pothole detection with edge intelligence and digital twin in Internet of Vehicles,” IEEE Internet Things J., vol. 12, no. 5, pp. 4852–4859, 2025, doi: 10.1109/JIOT.2024.3517342.
[25]P. Chandak, S. Chaurasia, and M. Patil, “Identification of potholes and speed bumps using SSD and YOLO,” in Intelligent Computing and Networking, IC-ICN 2024, Lecture Notes in Networks and Systems, vol. 1172, G. Ghinea et al., Eds., pp. 1–10, Springer, 2025, doi: 10.1007/978-981-97-8631-2_1.
[26]M. Kucera, A. D. Krumins, K. A. Ozolins, and M. N. Avran, “Robust pothole detection using YOLOv5 and SSD: Accuracy, speed, and deployment trade-offs,” in 2025 9th Int. Symp. Innov. Approaches Smart Technol. (ISAS), pp. 1–8, IEEE, 2025, doi: 10.1109/ISAS66241.2025.11101955.
[27]M. Hu and Q. Xia, “Pothole detection and recognition based on transfer learning,” arXiv preprint, 2025, doi: 2509.06750.
[28]A. Insani, U. Khusni, T. Haryanto, Suherlan, Yasmin, and A. R. Hidayat, "Automatic detection of potholes using an enhanced Faster R-CNN for improved traffic safety," in Proc. 5th Int. Conf. Sci. Inf. Technol. Smart Admin. (ICSINTESA), Yogyakarta, Indonesia, 2025, pp. 646–651, doi: 10.1109/ICSINTESA68165.2025.11413655.
[29]A. Karukayil, C. Quail and F. Auat Cheein, "Deep Learning Enhanced Feature Extraction of Potholes Using Vision and LiDAR Data for Road Maintenance," in IEEE Access, vol. 12, pp. 184541-184549, 2024, doi: 10.1109/ACCESS.2024.3512783.