YOLO-AVE: Improved YOLO11 for Reading of Digital Measuring Device Without Serial Communication Interface
IEEE Access · 2026 · 14 · 165052-165066
DOI: 10.1109/access.2026.3745233 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Access, 2026. Verfasst von Adindra Vickar Ega und 3 weiteren Autor:innen. Die Arbeit nennt 32 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Unit of Research, Innovation, and Community Engagement, Faculty of Engineering, Universitas Indonesia, for the Article Processing Charge.
Abstract (Originalsprache)
Automated digit recognition reading from digital measuring device (DMD) displays remains challenging due to environmental disturbances, camera angles, and tiny objects, such as decimal points. To address these limitations, this study proposes YOLO-AVE (Modified YOLO11 for Automated Visual Extraction of Digital Measuring Device Reading), an efficient, lightweight, optimized oriented object detection framework for DMD digit sequence recognition with tiny objects such as decimal points and negative signs. YOLO-AVE is built upon YOLO11s and integrates Oriented Bounding Box (OBB) detection heads, an additional P2 detection layer, GhostConv module, a novel Oriented SPPF (O-SPPF), and a novel Deformable C2PSA (D-C2PSA) module. Comprehensive ablation experiments demonstrate that YOLO-AVE with the A7 configuration achieves the highest Full Sequence Reading (FSR) accuracy of 97.9% with only 12 reading errors on a challenging unseen test set of 561 DMD display images. YOLO-AVE maintains an average FSR of 96.5% on an augmented version of the test set, including brightness, blur, and rotation separately. Furthermore, YOLO-AVE training with 5-fold cross-validation generates an average mAP50 of 99.5%, mAP ${}^{\mathrm {50-95}}$ of 96.1%, and an average FSR of 91.6%, confirming robust generalization with 33.8 FPS, demonstrating its potential application for real-time DMD reading in practical calibration environments without the need for a serial communication interface for data acquisition.
Förderung
- Unit of Research, Innovation, and Community Engagement, Faculty of Engineering, Universitas Indonesia, for the Article Processing Charge
Zitieren
APA
Ega, A. V., Gunawan, D., Halim, A., & Samodro, R. A. (2026). YOLO-AVE: Improved YOLO11 for Reading of Digital Measuring Device Without Serial Communication Interface. IEEE Access, 14, 165052-165066. https://doi.org/10.1109/access.2026.3745233
Vancouver
Ega AV, Gunawan D, Halim A, Samodro RA. YOLO-AVE: Improved YOLO11 for Reading of Digital Measuring Device Without Serial Communication Interface. IEEE Access. 2026;14:165052-165066. doi:10.1109/access.2026.3745233
BibTeX
@article{Ega2026,
author = {Ega, Adindra Vickar and Gunawan, Dadang and Halim, Abdul and Samodro, Rudi Anggoro},
title = {YOLO-AVE: Improved YOLO11 for Reading of Digital Measuring Device Without Serial Communication Interface},
journal = {IEEE Access},
year = {2026},
volume = {14},
pages = {165052--165066},
doi = {10.1109/access.2026.3745233}
}
EUVIMED