A deep learning solution for compressed semantic segmentation of LiDAR point cloud maps

Authors

  • Bui Thi Thanh Tam (Corresponding Author) Institute of Information Technology and Electronics, Academy of Military Science and Technology
  • Cao Van Toan Institute of Information Technology and Electronics, Academy of Military Science and Technology
  • Phan Huy Anh Institute of Information Technology and Electronics, Academy of Military Science and Technology
  • Pham Van Quoc HUS High School for Gifted Students of Science, VNU University of Science
  • Pham Dang Duong HUS High School for Gifted Students of Science, VNU University of Science

DOI:

https://doi.org/10.54939/1859-1043.j.mst.IITE.2025.131-138

Keywords:

Deep learning, Localization and navigation; Point cloud; LiDAR; Semantic segmentation.

Abstract

Navigating Unmanned Aerial Vehicles (UAVs) in Global Navigation Satellite System (GNSS)-denied environments often relies on pre-built Light Detection and Ranging (LiDAR) maps. However, the large memory footprint and high computational cost of these point cloud maps pose significant challenges for resource-constrained UAVs. This paper proposes a deep learning solution using a lightweight, modified RandLA-Net architecture to efficiently compress and semantically segment these maps. Our results demonstrate a significant reduction in model size and memory usage while maintaining competitive segmentation accuracy, presenting a viable solution for real-time, on-board processing on embedded systems.

References

[1]. Chang, Y., “A Review of UAV Autonomous Navigation in GPS-Denied Environments,” Robotics and Autonomous Systems, Vol. 170, p. 104533 (2023).

[2]. Meng, H.-Y., “VV-Net: Voxel VAE Net with Group Convolutions,” in IEEE/CVF International Conference on Computer Vision (ICCV) (2019).

[3]. Guibas, H. T., “KPConv: Flexible and Deformable Convolution for Point Clouds,” in IEEE/CVF International Conference on Computer Vision (ICCV), Seoul (2019).

[4]. Hu, Q., “RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds,” in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, WA, USA (2020).

[5]. Yang, B., “PIXOR: Real-Time 3D Object Detection from Point Clouds,” in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2018).

[6]. Qi, C. R., “PointNet: Deep Learning on Point Sets for 3D Classification,” in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2017).

[7]. Qi, C. R., “PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space,” in Advances in Neural Information Processing Systems (NeurIPS) (2017).

[8]. Landrieu, L., “Large-Scale Point Cloud Semantic Segmentation with Superpoint Graphs,” in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2018).

[9]. Armeni, I., “Joint 2D-3D-Semantic Data for Indoor Scene Understanding,” in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2017).

Downloads

Published

30-10-2025

How to Cite

[1]
Bui Thi Thanh Tam, Cao Van Toan, Phan Huy Anh, Pham Van Quoc, and Pham Dang Duong, “A deep learning solution for compressed semantic segmentation of LiDAR point cloud maps”, JMST, no. IITE, pp. 131–138, Oct. 2025.

Issue

Section

Electronic Engineering

Most read articles by the same author(s)

1 2 > >>