REW-DETR: A Lightweight Real-Time Detection Transformer for Bridge Surface Defect Detection.
- 1College of Computer Science and Engineering, Guilin University of Technology, Guilin 541004, China.
- 2College of Information Engineering, Guilin College, Guilin 541006, China.
- 3Guangxi Key Laboratory of Embedded Technology and Intelligent System, Guilin 541004, China.
Abstract
Timely detection of bridge surface defects is important for structural safety assessment and inspection efficiency. To address defect scale variation, complex background interference, and the high computational cost of RT-DETR-R50, this study proposes REW-DETR, a lightweight bridge surface defect detector. REW-DETR adopts a coordinated lightweight design integrating efficient feature extraction, multi-level channel recalibration, and localization optimization. RepNCSPELAN4-ADown is employed to reconstruct the backbone and fusion stages, ECA is embedded at the P3-P5 feature levels to enhance channel-wise feature discrimination, and Wise-IoU v3 optimizes bounding-box regression without inference-time overhead. On the test set of the bridge surface defect dataset, REW-DETR achieved Precision, Recall, mAP₅₀, and mAP50:95 values of 87.0%, 70.4%, 73.7%, and 42.4%, respectively. These values exceeded those of RT-DETR-R50 by 3.8, 5.5, 4.0, and 4.5 percentage points, respectively. Meanwhile, the parameters and FLOPs were reduced by approximately 53.3% and 60.6%, respectively, while FPS increased from 51.2 to 70.8. After dataset-specific retraining on the bridge-defects-detection dataset from Roboflow Universe, REW-DETR achieved a Recall of 58.0% and an mAP₅₀ of 55.4% while retaining a lightweight configuration. These results demonstrate an improved accuracy-efficiency trade-off and motivate further evaluation of REW-DETR for UAV-assisted bridge inspection, while onboard inference performance remains to be validated.
Related
- FusionLite-DETR: A Lightweight Transformer-Based Model for UAV-Based Insulator Defect Detection
- LDA-YOLO11: a lightweight detail-attention detector for edge-deployable road surface defect detection
- CyberRoBERTa: A Lightweight Transformer-Based Language Model for IoT Intrusion Detection
- A lightweight photovoltaic panel defect detection model based on YOLO
- SA-DETR: Lightweight Structure-Aware Defect Detection Model for Sustainable Low-Carbon Steel Manufacturing Based on RT-DETR-L
BibTeX
@article{cheng2026rewdetr,
title = {{REW-DETR: A Lightweight Real-Time Detection Transformer for Bridge Surface Defect Detection.}},
author = {Xiaohui Cheng and Yonghong Pang and Yanjun Wang and Yun Deng and Qiu Lu},
journal = {Sensors (Basel, Switzerland)},
year = {2026},
month = oct,
doi = {10.3390/s26206401},
url = {https://europepmc.org/article/MED/42944263},
note = {CC BY},
}
RIS
TY - JOUR
TI - REW-DETR: A Lightweight Real-Time Detection Transformer for Bridge Surface Defect Detection.
AU - Xiaohui Cheng
AU - Yonghong Pang
AU - Yanjun Wang
AU - Yun Deng
AU - Qiu Lu
PY - 2026
DA - 2026/10/10
JO - Sensors (Basel, Switzerland)
DO - 10.3390/s26206401
UR - https://europepmc.org/article/MED/42944263
AB - Timely detection of bridge surface defects is important for structural safety assessment and inspection efficiency. To address defect scale variation, complex background interference, and the high computational cost of RT-DETR-R50, this study proposes REW-DETR, a lightweight bridge surface defect detector. REW-DETR adopts a coordinated lightweight design integrating efficient feature extraction, multi-level channel recalibration, and localization optimization. RepNCSPELAN4-ADown is employed to reconstruct the backbone and fusion stages, ECA is embedded at the P3-P5 feature levels to enhance channel-wise feature discrimination, and Wise-IoU v3 optimizes bounding-box regression without inference-time overhead. On the test set of the bridge surface defect dataset, REW-DETR achieved Precision, Recall, mAP₅₀, and mAP50:95 values of 87.0%, 70.4%, 73.7%, and 42.4%, respectively. These values exceeded those of RT-DETR-R50 by 3.8, 5.5, 4.0, and 4.5 percentage points, respectively. Meanwhile, the parameters and FLOPs were reduced by approximately 53.3% and 60.6%, respectively, while FPS increased from 51.2 to 70.8. After dataset-specific retraining on the bridge-defects-detection dataset from Roboflow Universe, REW-DETR achieved a Recall of 58.0% and an mAP₅₀ of 55.4% while retaining a lightweight configuration. These results demonstrate an improved accuracy-efficiency trade-off and motivate further evaluation of REW-DETR for UAV-assisted bridge inspection, while onboard inference performance remains to be validated.
SN - 1424-8220
ER -