RT-DETR-Pothole: lightweight real-time detection transformers for improved road pothole detection

Assessment of on-time road condition is crucial for ensuring the safety of the motorist. One of the recent approaches to detecting road potholes is to analyze images captured from an unmanned aerial vehicle (UAV). Although the traditional deep learning model could perform accurate detection durin...

Full description

Bibliographic Details
Main Author: Abd Rahman, Mohd Amiruddin
Format: Conference or Workshop Item
Language:English
Published: 2025
Online Access:http://psasir.upm.edu.my/id/eprint/115933/
http://psasir.upm.edu.my/id/eprint/115933/1/115933.pdf
_version_ 1848866893369180160
author Abd Rahman, Mohd Amiruddin
author_facet Abd Rahman, Mohd Amiruddin
author_sort Abd Rahman, Mohd Amiruddin
building UPM Institutional Repository
collection Online Access
description Assessment of on-time road condition is crucial for ensuring the safety of the motorist. One of the recent approaches to detecting road potholes is to analyze images captured from an unmanned aerial vehicle (UAV). Although the traditional deep learning model could perform accurate detection during offline analysis, there is still a limitation of the available algorithms that could perform real-time evaluation. Therefore, this study proposes a lightweight transformer algorithm, the real-time detection transformer (RT-DETR), for online evaluation of road pothole images. The models were tested in practical deployment scenarios and compared with several other object detection models, such as Faster RCNNSqueezeNet, YOLOv8x, YOLOv9e, YOLOv10x, and YOLO11x. The results show that the RT-DETRPothole outperformed all other models in detection accuracy, achieving the highest mAP0.50 (0.834) and mAP0.50-0.95 (0.565), along with a high F1-Score (0.809), indicating superior precision and recall, and at the same time it could maintain low inference time. Overall, RT-DETR-Pothole is the most suitable model for real-time pothole detection, especially for detecting smaller, less visible potholes, with a resonable inference time for pavement engineering applications.
first_indexed 2025-11-15T14:27:50Z
format Conference or Workshop Item
id upm-115933
institution Universiti Putra Malaysia
institution_category Local University
language English
last_indexed 2025-11-15T14:27:50Z
publishDate 2025
recordtype eprints
repository_type Digital Repository
spelling upm-1159332025-03-19T07:48:38Z http://psasir.upm.edu.my/id/eprint/115933/ RT-DETR-Pothole: lightweight real-time detection transformers for improved road pothole detection Abd Rahman, Mohd Amiruddin Assessment of on-time road condition is crucial for ensuring the safety of the motorist. One of the recent approaches to detecting road potholes is to analyze images captured from an unmanned aerial vehicle (UAV). Although the traditional deep learning model could perform accurate detection during offline analysis, there is still a limitation of the available algorithms that could perform real-time evaluation. Therefore, this study proposes a lightweight transformer algorithm, the real-time detection transformer (RT-DETR), for online evaluation of road pothole images. The models were tested in practical deployment scenarios and compared with several other object detection models, such as Faster RCNNSqueezeNet, YOLOv8x, YOLOv9e, YOLOv10x, and YOLO11x. The results show that the RT-DETRPothole outperformed all other models in detection accuracy, achieving the highest mAP0.50 (0.834) and mAP0.50-0.95 (0.565), along with a high F1-Score (0.809), indicating superior precision and recall, and at the same time it could maintain low inference time. Overall, RT-DETR-Pothole is the most suitable model for real-time pothole detection, especially for detecting smaller, less visible potholes, with a resonable inference time for pavement engineering applications. 2025 Conference or Workshop Item PeerReviewed text en http://psasir.upm.edu.my/id/eprint/115933/1/115933.pdf Abd Rahman, Mohd Amiruddin (2025) RT-DETR-Pothole: lightweight real-time detection transformers for improved road pothole detection. In: 2025 9th International Conference on Artificial Intelligence, Automation and Control Technologies (AIACT 2025), 17-21 Feb. 2025, Sapporo, Japan. (pp. 1-6). (Submitted)
spellingShingle Abd Rahman, Mohd Amiruddin
RT-DETR-Pothole: lightweight real-time detection transformers for improved road pothole detection
title RT-DETR-Pothole: lightweight real-time detection transformers for improved road pothole detection
title_full RT-DETR-Pothole: lightweight real-time detection transformers for improved road pothole detection
title_fullStr RT-DETR-Pothole: lightweight real-time detection transformers for improved road pothole detection
title_full_unstemmed RT-DETR-Pothole: lightweight real-time detection transformers for improved road pothole detection
title_short RT-DETR-Pothole: lightweight real-time detection transformers for improved road pothole detection
title_sort rt-detr-pothole: lightweight real-time detection transformers for improved road pothole detection
url http://psasir.upm.edu.my/id/eprint/115933/
http://psasir.upm.edu.my/id/eprint/115933/1/115933.pdf