SDM-Citrus6D: dynamic 6D pose estimation of citrus based on semi-dense image matching
Journal of King Saud University Computer and Information Sciences · 2026-11 · 38 (9)
DOI: 10.1007/s44443-026-01399-7 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of King Saud University Computer and Information Sciences, online erschienen am 12.11.2026. Verfasst von Xufeng Xu und 9 weiteren Autor:innen. Die Arbeit nennt 50 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: Key R&D Program of Zhejiang Province, State Key Laboratory of Agricultural Equipment Technology.
Abstract (Originalsprache)
6D pose perception facilitates the acquisition of precise position and orientation in three-dimensional space, thereby establishing a reliable foundation for intelligent sorting, robotic grasping control and automated packaging. However, existing 6D pose estimation methods encounter significant challenges in agricultural sorting lines. The reliance on instance-specific CAD models in conventional methods limits adaptability to natural variations in fruit scale and curvature, while the weak textures of smooth peel and dynamic inter-frame motion degrade feature extraction accuracy. Therefore, this study introduces a semi-dense image matching-driven 6D pose estimation framework, named SDM-Citrus6D, specifically designed for dynamic citrus pose tracking during conveyor inspection. The method first employs a segmentation model to achieve accurate detection and target mask extraction. A semi-dense image matching model, CitrusMatch, is then constructed, incorporating the Wavelet Feature Enhancement Module (WFEM) and the Adaptive Fuzzy Region Matching (AFRM) mechanism to enhance feature discriminability on weak-texture surfaces and suppress motion-blur interference. Based on the image matching results and depth information, a combination of the Kabsch and Iterative Closest Point (ICP) algorithms is applied to solve the 6D pose. A robotic experimental platform for fruit pose estimation was established to validate the proposed method. The results show that the CitrusMatch achieves a matching accuracy of Acc@5 px = 97.09%. Ablation analysis reveals that WFEM and AFRM play complementary roles: WFEM enhances low-frequency context in weak-texture regions to boost pose estimation, while AFRM refines sub-pixel localization accuracy. In validation experiments, SDM-Citrus6D achieved an ADD@5 of 92.50% for citrus 6D pose estimation. Compared with relevant studies, SDMCitrus6D exhibits enhanced robustness in dynamic environments and improved estimation precision. In summary, the proposed SDM-Citrus6D realizes accurate 6D pose estimation based on high-precision image matching, offering technical support for intelligent perception and automation in postharvest citrus commercialization.
Förderung
- Key R&D Program of Zhejiang Province
- State Key Laboratory of Agricultural Equipment Technology
Zitieren
APA
Xu, X., Guo, K., Jiang, H., Xu, T., Wei, Z., Guo, R., Li, Z., Yu, T., Chen, J., & Rao, X. (2026). SDM-Citrus6D: dynamic 6D pose estimation of citrus based on semi-dense image matching. Journal of King Saud University Computer and Information Sciences, 38(9). https://doi.org/10.1007/s44443-026-01399-7
Vancouver
Xu X, Guo K, Jiang H, Xu T, Wei Z, Guo R, et al. SDM-Citrus6D: dynamic 6D pose estimation of citrus based on semi-dense image matching. Journal of King Saud University Computer and Information Sciences. 2026;38(9). doi:10.1007/s44443-026-01399-7
BibTeX
@article{Xu2026,
author = {Xu, Xufeng and Guo, Kai and Jiang, Huanyu and Xu, Tao and Wei, Zichao and Guo, Ruokai and Li, Zetong and Yu, Tiancheng and Chen, Jin and Rao, Xiuqin},
title = {SDM-Citrus6D: dynamic 6D pose estimation of citrus based on semi-dense image matching},
journal = {Journal of King Saud University Computer and Information Sciences},
year = {2026},
volume = {38},
number = {9},
doi = {10.1007/s44443-026-01399-7}
}
EUVIMED