EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

LapDINO: A DINOv3 and Laplacian Pyramid-Based Approach for Outdoor Terrain Segmentation

Shiquan Ling, Xingchen Qin iD, Wenkang Xu, Mingmin Fu, Hao Huang, Shijie Ma, Zhenyu Liu

Sensors · 17.06.2026 · 26 (12) · 3843

DOI: 10.3390/s26123843 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 17.06.2026. Verfasst von Shiquan Ling und 6 weiteren Autor:innen. Beteiligte Einrichtung: School of Mechanical Engineering, Zhejiang University, Hangzhou 310030, China. Die Arbeit nennt 40 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

As autonomous driving technology expands from structured urban roads to unstructured outdoor environments, precise understanding of complex terrain has become a critical requirement for ensuring safe vehicle navigation. However, outdoor environments are characterized by high dynamics, drastic illumination variations, ambiguous category boundaries, and prohibitive annotation costs, making traditional supervised learning methods that rely on large amounts of pixel-level annotations difficult to generalize. In this paper, we propose a novel dual-path bidirectional interactive encoder, termed LapDINO, that effectively combines the strong semantic generalization capability of the self-supervised foundation model DINOv3 with the multi-scale frequency analysis capacity of the Laplacian pyramid. Specifically, we leverage DINOv3 to extract global semantic features as a “semantic map”, while simultaneously obtaining multi-scale high-frequency details through Laplacian pyramid decomposition as “structural contours”. Building upon this, we design a bidirectional cross-attention fusion mechanism that enables dynamic interaction and mutual refinement between semantic information and geometric details. Furthermore, we introduce a multi-branch attention enhancement module that extracts pyramid features from three complementary perspectives. To address domain shift, we design lightweight visual adapters that enable efficient fine-tuning of the frozen DINOv3 backbone. Finally, we construct two off-road terrain segmentation datasets, VOTD and VOCD, to facilitate research in this domain. Experimental results demonstrate that the proposed method achieves state-of-the-art performance, striking an optimal balance between accuracy and computational efficiency, thereby providing a robust and efficient engineering solution for terrain perception in off-road environments.

Schlagwörter

Stichwörter der Autor:innen

DINOv3bidirectional cross-attentionlaplacian pyramidmulti-branch attentionoff-road terrain segmentationvisual adapter

Autor:innen und Institutionen

  • School of Mechanical Engineering, Zhejiang University, Hangzhou 310030, China.
  • Geely Auto, Hangzhou 310051, China.

ORCID-Profile: Xingchen Qin

Zitieren

APA

Ling, S., Qin, X., Xu, W., Fu, M., Huang, H., Ma, S., & Liu, Z. (2026). LapDINO: A DINOv3 and Laplacian Pyramid-Based Approach for Outdoor Terrain Segmentation. Sensors, 26(12), 3843. https://doi.org/10.3390/s26123843

Vancouver

Ling S, Qin X, Xu W, Fu M, Huang H, Ma S, et al. LapDINO: A DINOv3 and Laplacian Pyramid-Based Approach for Outdoor Terrain Segmentation. Sensors. 2026;26(12):3843. doi:10.3390/s26123843

BibTeX
@article{Ling2026,
  author = {Ling, Shiquan and Qin, Xingchen and Xu, Wenkang and Fu, Mingmin and Huang, Hao and Ma, Shijie and Liu, Zhenyu},
  title = {LapDINO: A DINOv3 and Laplacian Pyramid-Based Approach for Outdoor Terrain Segmentation},
  journal = {Sensors},
  year = {2026},
  volume = {26},
  number = {12},
  pages = {3843},
  doi = {10.3390/s26123843}
}
Kostenlos starten →