FCE-YOLO-Based Visual Detection and Localization in Near-Field Narrow-Space Environments
Sensors · 14.11.2026 · 26 (22) · 7200
DOI: 10.3390/s26227200 ↗Kurz gefasst
Fachartikel in der Zeitschrift Sensors, online erschienen am 14.11.2026. Verfasst von Weilin Cheng und 6 weiteren Autor:innen. Die Arbeit nennt 38 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China.
Abstract (Originalsprache)
Automated disassembly and assembly of inertial confinement fusion (ICF) diagnostic heads demand visual detection and localization in near-field narrow-space environments under simulated sensor noise and a shallow depth of field, where general object detectors are noise-sensitive and imprecise at high intersection-over-union (IoU) thresholds. We propose FCE-YOLO, a YOLOv11-based detection network. A weighted bidirectional feature pyramid network (BiFPN) replaces concatenation fusion; the proposed bidirectional coordinate cross-attention module sharpens bounding-box edge alignment through directional aggregation; and the distance-sensitive Wise-IoU (WIoU) loss constrains bounding-box regression. A binocular localization front end and a hierarchical control strategy with far-field approach and near-field fine positioning complete the pipeline; a teach-in work-point offset absorbs the constant systematic bias, while iterative residual elimination prevents error accumulation within the positioning loop. FCE-YOLO improves the mean average precision (mAP) over IoU 0.50–0.95 from 77.08% to 84.53% under a random image-level split, with slightly fewer parameters and less computation. Under a stricter frame-grouped re-partition, it attains 82.50 ± 0.56% on the held-out frame groups, against 76.74 ± 1.26% for the baseline. Field validation across six measurement sessions yields a single-measurement depth repeatability of 0.220 mm and a tool-center-point repeat-positioning accuracy below 0.1 mm on all axes.
Förderung
- National Natural Science Foundation of China
Zitieren
APA
Cheng, W., Li, H., Chen, L., Yang, Z., Chen, L., Yang, P., & Huang, Y. (2026). FCE-YOLO-Based Visual Detection and Localization in Near-Field Narrow-Space Environments. Sensors, 26(22), 7200. https://doi.org/10.3390/s26227200
Vancouver
Cheng W, Li H, Chen L, Yang Z, Chen L, Yang P, et al. FCE-YOLO-Based Visual Detection and Localization in Near-Field Narrow-Space Environments. Sensors. 2026;26(22):7200. doi:10.3390/s26227200
BibTeX
@article{Cheng2026,
author = {Cheng, Weilin and Li, Haiyan and Chen, Li and Yang, Zhenghua and Chen, Lin and Yang, Pin and Huang, Yunbao},
title = {FCE-YOLO-Based Visual Detection and Localization in Near-Field Narrow-Space Environments},
journal = {Sensors},
year = {2026},
volume = {26},
number = {22},
pages = {7200},
doi = {10.3390/s26227200}
}
EUVIMED