Improved DETR-Based Detection of Icing and Snow on Aircraft Surfaces
Shi Yongsheng, Xiang Yuhang
2024
Abstract
Utilizing neural network models to detect icing and snow accumulation on aircraft surfaces can significantly reduce the workload of maintenance personnel, enhance operational efficiency, and lower aircraft operating costs. This proposal marks the first application of the transformer-based object detection model DETR to the detection of icing and snow on aircraft surfaces. To address the issue of significant boundary box prediction deviations in DETR, the RefineBox localization optimization network was employed for improvements. Performance was compared and analyzed on a custom dataset, revealing a 1.8% increase in the mAP metric for the enhanced model. Ground trials were conducted to validate the accuracy and feasibility of the improved model in detecting aircraft surface icing and snow. The results demonstrate that the enhanced model performs well, exhibits strong environmental adaptability, and can operate stably on mainstream devices.
DownloadPaper Citation
in Harvard Style
Yongsheng S. and Yuhang X. (2024). Improved DETR-Based Detection of Icing and Snow on Aircraft Surfaces. In Proceedings of the 1st International Conference on Data Mining, E-Learning, and Information Systems - Volume 1: DMEIS; ISBN 978-989-758-715-3, SciTePress, pages 58-63. DOI: 10.5220/0012877900004536
in Bibtex Style
@conference{dmeis24,
author={Shi Yongsheng and Xiang Yuhang},
title={Improved DETR-Based Detection of Icing and Snow on Aircraft Surfaces},
booktitle={Proceedings of the 1st International Conference on Data Mining, E-Learning, and Information Systems - Volume 1: DMEIS},
year={2024},
pages={58-63},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012877900004536},
isbn={978-989-758-715-3},
}
in EndNote Style
TY - CONF
JO - Proceedings of the 1st International Conference on Data Mining, E-Learning, and Information Systems - Volume 1: DMEIS
TI - Improved DETR-Based Detection of Icing and Snow on Aircraft Surfaces
SN - 978-989-758-715-3
AU - Yongsheng S.
AU - Yuhang X.
PY - 2024
SP - 58
EP - 63
DO - 10.5220/0012877900004536
PB - SciTePress