EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

SGE-Flow: 4D mmWave Radar 3D Object Detection via Spatiotemporal Geometric Enhancement and Inter-Frame Flow

Huajun Meng iD, Zijie Yu, Cheng Li, Chao Li iD, Xiaojun Liu

Sensors · 06.03.2026 · 26 (5) · 1679

DOI: 10.3390/s26051679 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 06.03.2026. Verfasst von Huajun Meng und 4 weiteren Autor:innen. Beteiligte Einrichtung: Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100190, China. Die Arbeit nennt 32 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China under Grant.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

4D millimeter-wave radar provides a promising solution for robust perception in adverse weather. Existing detectors still struggle with sparse and noisy point clouds, and maintaining real-time inference while achieving competitive accuracy remains challenging. We propose SGE-Flow, a streamlined PointPillars-based 4D radar 3D detector that embeds lightweight spatiotemporal geometric enhancements into the voxelization front-end. Velocity Displacement Compensation (VDC) leverages compensated radial velocity to align accumulated points in physical space and improve geometric consistency. Distribution-Aware Density (DAD) enables fast density feature extraction by estimating per-pillar density from simple statistical moments, which also restores vertical distribution cues lost during pillarization. To compensate for the absence of tangential velocity measurements, a Transformer-based Inter-frame Flow (IFF) module infers latent motion from frame-to-frame pillar occupancy changes. Evaluations on the View-of-Delft (VoD) dataset show that SGE-Flow achieves 53.23% 3D mean Average Precision (mAP) while running at 72 frames per second (FPS) on an NVIDIA RTX 3090. The proposed modules are plug-and-play and can also improve strong baselines such as MAFF-Net.

Schlagwörter

Stichwörter der Autor:innen

3D object detection4D millimeter-wave radarPointPillarsinter-frame flowspatiotemporal enhancement

Autor:innen und Institutionen

  • Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100190, China.
  • Key Laboratory of Electromagnetic Radiation and Sensing Technology, Chinese Academy of Sciences, Beijing 100190, China.
  • School of Electronic, Electrical and Communication Engineering, University of Chinese Academy of Sciences, Beijing 100049, China.

ORCID-Profile: Huajun Meng, Chao Li

Förderung

  • National Natural Science Foundation of China under Grant

Zitieren

APA

Meng, H., Yu, Z., Li, C., Li, C., & Liu, X. (2026). SGE-Flow: 4D mmWave Radar 3D Object Detection via Spatiotemporal Geometric Enhancement and Inter-Frame Flow. Sensors, 26(5), 1679. https://doi.org/10.3390/s26051679

Vancouver

Meng H, Yu Z, Li C, Li C, Liu X. SGE-Flow: 4D mmWave Radar 3D Object Detection via Spatiotemporal Geometric Enhancement and Inter-Frame Flow. Sensors. 2026;26(5):1679. doi:10.3390/s26051679

BibTeX
@article{Meng2026,
  author = {Meng, Huajun and Yu, Zijie and Li, Cheng and Li, Chao and Liu, Xiaojun},
  title = {SGE-Flow: 4D mmWave Radar 3D Object Detection via Spatiotemporal Geometric Enhancement and Inter-Frame Flow},
  journal = {Sensors},
  year = {2026},
  volume = {26},
  number = {5},
  pages = {1679},
  doi = {10.3390/s26051679}
}
Kostenlos starten →