EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Rotation-Equivariant Feature Learning on Polar BEV for Robust LiDAR Place Recognition

Zhenhuan Yuan, Youchun Xu, Zhichao Zhang, Yuan Zhu, Jianshi Li, Feng Lu, Le Wang, Jinsheng Chen, Wei Lei

Applied Sciences · 17.06.2026 · 16 (12) · 6155

DOI: 10.3390/app16126155 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Applied Sciences, online erschienen am 17.06.2026. Verfasst von Zhenhuan Yuan und 8 weiteren Autor:innen. Die Arbeit nennt 25 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

LiDAR-based place recognition is critical for long-term autonomous navigation in Global Navigation Satellite System (GNSS)-denied environments, yet existing methods struggle to balance accuracy and efficiency under substantial yaw rotations. This paper proposes a robust framework based on a multi-channel polar bird’s-eye-view (BEV) representation. Under yaw-dominated revisits, the polar BEV image transforms yaw rotation into cyclic column shifts, providing a useful structural prior for rotation-equivariant feature extraction. Raw point clouds are projected onto polar BEV grids encoding density, height, and intensity. A rotation-equivariant feature extractor comprising a Radial Compression Module and a rotation-equivariant Transformer module captures long-range azimuthal dependencies via Conditional Positional Encoding and Circular Relative-Position Bias. The equivariant features are aggregated by NetVLAD into a compact global descriptor, trained end-to-end with a hard-example mining triplet loss. Extensive experiments on the public KITTI and NCLT datasets, as well as our self-constructed LiDAR Place Recognition Revisit (LPRR) dataset, demonstrate competitive performance on KITTI and superior performance on NCLT and LPRR among the compared methods. The proposed framework achieves a favorable trade-off between performance and computational cost, and shows promising cross-dataset generalization on the evaluated NCLT and LPRR datasets without fine-tuning.

Zitieren

APA

Yuan, Z., Xu, Y., Zhang, Z., Zhu, Y., Li, J., Lu, F., Wang, L., Chen, J., & Lei, W. (2026). Rotation-Equivariant Feature Learning on Polar BEV for Robust LiDAR Place Recognition. Applied Sciences, 16(12), 6155. https://doi.org/10.3390/app16126155

Vancouver

Yuan Z, Xu Y, Zhang Z, Zhu Y, Li J, Lu F, et al. Rotation-Equivariant Feature Learning on Polar BEV for Robust LiDAR Place Recognition. Applied Sciences. 2026;16(12):6155. doi:10.3390/app16126155

BibTeX
@article{Yuan2026,
  author = {Yuan, Zhenhuan and Xu, Youchun and Zhang, Zhichao and Zhu, Yuan and Li, Jianshi and Lu, Feng and Wang, Le and Chen, Jinsheng and Lei, Wei},
  title = {Rotation-Equivariant Feature Learning on Polar BEV for Robust LiDAR Place Recognition},
  journal = {Applied Sciences},
  year = {2026},
  volume = {16},
  number = {12},
  pages = {6155},
  doi = {10.3390/app16126155}
}
Kostenlos starten →