Deep learning-based spectral demultiplexing for four Fabry–Pérot sensor systems
Optics Express · 16.11.2026 · 34 (23) · 43985
DOI: 10.1364/oe.618394 ↗Kurz gefasst
Fachartikel in der Zeitschrift Optics Express, online erschienen am 05.11.2026. Verfasst von Xiaoming Zhao und 6 weiteren Autor:innen. Die Arbeit nennt 30 Literaturangaben. Gefördert von: Science and Technology Department of Zhejiang Province, Taizhou Municipal Science and Technology Bureau.
Abstract (Originalsprache)
Multiplexed Fabry–Pérot (F–P) sensing suffers from spectral overlap and crosstalk caused by similar cavity lengths, making individual sensor response recovery challenging. Existing data-driven methods mainly predict physical quantities and lack direct reconstruction of sensor spectra. This study proposes a multi-sensor convolutional transformer (MSCT) to separate and reconstruct the complete interference spectra of four F–P sensors from a single multiplexed spectrum. By exploiting interference periods, local fringe patterns, and global spectral characteristics, MSCT effectively recovers individual sensor responses. Experimental results demonstrate excellent agreement with reference spectra, achieving an R 2 of 0.9995 and an RMSE of 0.04 AU, while preserving interference periods, intensity variations, and peak valley structures.
Förderung
- Science and Technology Department of Zhejiang Province
- Taizhou Municipal Science and Technology Bureau
Zitieren
APA
Zhao, X., Wu, K., He, W., Lou, L., Chen, H., Yu, Z., & Zhou, C. (2026). Deep learning-based spectral demultiplexing for four Fabry–Pérot sensor systems. Optics Express, 34(23), 43985. https://doi.org/10.1364/oe.618394
Vancouver
Zhao X, Wu K, He W, Lou L, Chen H, Yu Z, et al. Deep learning-based spectral demultiplexing for four Fabry–Pérot sensor systems. Optics Express. 2026;34(23):43985. doi:10.1364/oe.618394
BibTeX
@article{Zhao2026,
author = {Zhao, Xiaoming and Wu, Kaipeng and He, Wei and Lou, Liangliang and Chen, Haoran and Yu, Zufeng and Zhou, Cheng},
title = {Deep learning-based spectral demultiplexing for four Fabry–Pérot sensor systems},
journal = {Optics Express},
year = {2026},
volume = {34},
number = {23},
pages = {43985},
doi = {10.1364/oe.618394}
}
EUVIMED