Unified Optimization of YOLOX for Robust Small Object Detection in Real-Time Potato Harvesting
Journal of Robotics and Mechatronics · 20.04.2026 · 38 (2) · 483-494
DOI: 10.20965/jrm.2026.p0483 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Robotics and Mechatronics, online erschienen am 20.04.2026. Verfasst von Joonam Kim und 3 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 53 Literaturangaben. Gefördert von: The Institute of Agricultural Machinery, National Agriculture and Food Research Organization.
Abstract (Originalsprache)
Automated potato harvesting requires vision systems meeting stringent operational constraints: near-zero potato misclassification (PMR 60%). Although YOLOX supports real-time processing, it exhibits critical performance limitations for small object detection. This paper introduces a unified optimization strategy combining training-level modifications (P3 feature enhancement, SimOTA parameter optimization, and size-aware loss weighting) with inference-level threshold optimization. Experimental results based on 10,000 images containing 232,000 annotations demonstrate that all four evaluated approaches achieved the defined operational constraints. SimOTA emerged as the optimal configuration, delivering the highest small object recall (18.18%) while maintaining PMR of 0.06% and IDR of 95.05%. P3 feature enhancement achieved the lowest PMR (0.05%), SALW provided balanced performance (PMR 0.06%, IDR 96.11%), and the P3+SimOTA combination failed critically (PMR 9.90%), revealing fundamental incompatibility between optimization components. All successful configurations exceeded real-time processing requirements (45–61 fps), confirming suitability for deployment in resource-constrained agricultural automation systems.
Förderung
- The Institute of Agricultural Machinery, National Agriculture and Food Research Organization
Zitieren
APA
Kim, J., Tokuda, K., Kim, G., & Yoshitoshi, R. (2026). Unified Optimization of YOLOX for Robust Small Object Detection in Real-Time Potato Harvesting. Journal of Robotics and Mechatronics, 38(2), 483-494. https://doi.org/10.20965/jrm.2026.p0483
Vancouver
Kim J, Tokuda K, Kim G, Yoshitoshi R. Unified Optimization of YOLOX for Robust Small Object Detection in Real-Time Potato Harvesting. Journal of Robotics and Mechatronics. 2026;38(2):483-494. doi:10.20965/jrm.2026.p0483
BibTeX
@article{Kim2026,
author = {Kim, Joonam and Tokuda, Kenichi and Kim, Giryeon and Yoshitoshi, Rena},
title = {Unified Optimization of YOLOX for Robust Small Object Detection in Real-Time Potato Harvesting},
journal = {Journal of Robotics and Mechatronics},
year = {2026},
volume = {38},
number = {2},
pages = {483--494},
doi = {10.20965/jrm.2026.p0483}
}
EUVIMED