Robo-DETR: Robustness-Aware Depth-Guided Transformer for Monocular 3-D Object Detection Under Adverse Visual Conditions
IEEE Internet of Things Journal · 01.06.2026 · 13 (11) · 23793-23806
DOI: 10.1109/jiot.2026.3671809 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Internet of Things Journal, erschienen am 01.06.2026. Verfasst von Haoyu Li und 7 weiteren Autor:innen. Die Arbeit nennt 38 Literaturangaben. Gefördert von: Key-Area Research and Development Program of Guangdong Province, Shenzhen Science and Technology Program.
Abstract (Originalsprache)
Monocular 3-D object detection is attractive for autonomous driving due to its low cost, but its performance degrades severely under adverse visual conditions that corrupt appearance cues and destabilize depth-related representations. We propose Robo-DETR, a robustness-aware depth-guided transformer framework that combines minimal visual-conditioned input preconditioning with detector-internal depth reliability correction. Specifically, a lightweight condition recognition and enhancement block (CREB) stabilizes low-level visual cues, while a condition-aware depth-guided block (CDB) refines depth logits via foreground-aware supervision and structured dynamic/region-level adjustment to suppress visual-induced depth noise. A dual-branch transformer decoder further promotes depth-appearance consistency through modality-specific cross-attention and fusion. Experiments on KITTI-C and the real-world TJ4DRadSet demonstrate consistent improvements over competitive monocular detectors under diverse degradations, and ablations validate the contribution of each component.
Förderung
- Key-Area Research and Development Program of Guangdong Province
- Shenzhen Science and Technology Program
Zitieren
APA
Li, H., Zhang, X., Zhong, J., Chen, Z., Zhao, Y., Wang, B., Leng, J., & Sun, C. (2026). Robo-DETR: Robustness-Aware Depth-Guided Transformer for Monocular 3-D Object Detection Under Adverse Visual Conditions. IEEE Internet of Things Journal, 13(11), 23793-23806. https://doi.org/10.1109/jiot.2026.3671809
Vancouver
Li H, Zhang X, Zhong J, Chen Z, Zhao Y, Wang B, et al. Robo-DETR: Robustness-Aware Depth-Guided Transformer for Monocular 3-D Object Detection Under Adverse Visual Conditions. IEEE Internet of Things Journal. 2026;13(11):23793-23806. doi:10.1109/jiot.2026.3671809
BibTeX
@article{Li2026,
author = {Li, Haoyu and Zhang, Xinyang and Zhong, Jiaru and Chen, Zitong and Zhao, Yueran and Wang, Bo and Leng, Jianghao and Sun, Chao},
title = {Robo-DETR: Robustness-Aware Depth-Guided Transformer for Monocular 3-D Object Detection Under Adverse Visual Conditions},
journal = {IEEE Internet of Things Journal},
year = {2026},
volume = {13},
number = {11},
pages = {23793--23806},
doi = {10.1109/jiot.2026.3671809}
}
EUVIMED