Fast Grid-Based Ground Segmentation of LiDAR Point Clouds Using Dual-Seed Expansion

Area of Science:

• Robotics and Autonomous Systems
• Computer Vision
• Geospatial Data Processing

Background:

• Ground segmentation is critical for LiDAR-based perception in autonomous driving.
• Accurate ground segmentation directly impacts object clustering and obstacle detection.
• Existing methods may struggle with sloped or sparse terrain, leading to missed detections.

Purpose of the Study:

• To propose a fast and accurate grid-based ground segmentation method for 3D LiDAR point clouds.
• To improve the robustness of ground segmentation in challenging environments.
• To provide an efficient preprocessing module for autonomous driving perception.

Main Methods:

• Projecting 3D LiDAR point clouds onto a 2D Cartesian grid with cell-wise height statistics.
• Utilizing a dual-seed expansion approach combining global and LiDAR-mounting-height seeds.
• Applying a local height-band criterion for point-level refinement to reduce misclassifications.

Main Results:

• Achieved average accuracy of 96.34%, F1-score of 97.11%, and ground Intersection over Union (IoU) of 94.40% on the SemanticKITTI dataset.
• Demonstrated a fast processing time of 12.15 ms per frame.
• Successfully reduced missed ground candidates in sloped or sparse regions.

Conclusions:

• The proposed dual-seed expansion method offers a fast and reliable solution for LiDAR ground segmentation.
• This method is well-suited as a preprocessing module for real-time autonomous driving perception.
• The approach effectively handles complex terrains, enhancing overall system performance.