LOGOS: LiDAR-Only Gaussian Elevation Splatting for Unified Tiny Obstacle Segmentation

2026-06-19

AI Summary

Existing LiDAR-based methods struggle to accurately segment small obstacles—such as curbs, debris, and potholes—due to their height similarity to road undulations and significant performance degradation in degraded point clouds or unstructured off-road environments. This work proposes LOGOS, a LiDAR-only system that models the road surface as a continuous mixture of 2D Gaussian primitives and introduces, for the first time, a backpropagation-free LiDAR Gaussian splatting approach. By incorporating free-space-aware initialization, a normal-aware elevation splatting function, and smoothness-constrained pruning, LOGOS achieves unified and robust modeling across both flat and sloped terrains. Experiments demonstrate that LOGOS substantially outperforms state-of-the-art methods on diverse urban and off-road datasets, maintaining high accuracy and real-time performance even under sparse point clouds or complex topographies.

Abstract

Robust obstacle segmentation is essential for the safety of intelligent robots, where LiDAR-based perception systems play a fundamental role in the robot-environment interaction. While extensive LiDAR-based approaches have demonstrated high performance on common obstacles in urban scenarios, their results on tiny obstacles such as curbs, gravel, and potholes remain unsatisfactory due to the significant similarity between tiny obstacles and inherent road undulations. Moreover, their segmentation accuracy even deteriorates sharply when the LiDAR scans suffer from degradation in challenging off-road scenes. To overcome these bottlenecks, we propose LOGOS, a LiDAR-only unified tiny obstacle segmentation system, which models the road surface as a continuous mixture of 2D Gaussian primitives and distinguishes tiny obstacles via high-presicion elevation estimation. Unlike existing Gaussian splatting methods that rely on iterative RGB training, LOGOS is a backpropagation-free LiDAR-only approach. It directly estimates Gaussian parameters via a freespace-aware initialization by incrementally pruning non-road primitives using smoothness constraints. Subsequently, pointwise signed distances are computed via a novel normal-aware elevation splatting function, ensuring robustness to both flat and sloped terrains. We evaluate LOGOS on a highly heterogeneous benchmark of point cloud frames collected from urban mobility scenarios and mining haulage off-road environments. These data are practically acquired using different LiDAR sensors and exhibit large variations in point density, terrain roughness, and obstacle types. Experiments on the road and off-road scenes demonstrate that LOGOS significantly outperforms other state-of-the-art methods, particularly in degraded point cloud regions and challenging off-road scenarios, while maintaining real-time efficiency.

Problem

• tiny obstacle segmentation
• LiDAR-based perception
• road surface modeling
• off-road scenes
• point cloud degradation

Innovation

• LiDAR-only
• Gaussian splatting
• tiny obstacle segmentation
• elevation estimation
• normal-aware