用局部几何特征提升激光雷达检测模型在不同安装位置下的泛化能力
GBlobs: Local LiDAR Geometry for Improved Sensor Placement Generalization
- 采用GBlobs局部点云特征，避免依赖绝对坐标的位置偏差
- 在多种传感器布局下实现顶尖的3D目标检测性能
- 适合需要跨场景部署的自动驾驶感知系统
本技术报告介绍了RoboSense 2025竞赛第3赛道的最高排名解决方案，在不同传感器布置条件下实现了3D目标检测的最先进性能。该方案采用GBlobs——一种专为提升模型在多样激光雷达配置下泛化能力而设计的局部点云特征描述子。当前基于激光雷达的3D检测器在使用传统全局特征（即绝对笛卡尔坐标）训练时，容易产生“几何捷径”，导致模型过度依赖物体的绝对位置，而非形状与外观特征。尽管在域内数据上表现良好，但此类捷径严重限制了在不同点云分布（如传感器位置变化引起）下的泛化能力。通过将GBlobs作为网络输入特征，有效规避了这一几何捷径，促使网络学习到更鲁棒、以物体为中心的表征。该方法显著提升了模型泛化能力，充分体现了在本次挑战赛中的优异表现。
原文摘要 · Abstract (English)
This technical report outlines the top-ranking solution for RoboSense 2025: Track 3, achieving state-of-the-art performance on 3D object detection under various sensor placements. Our submission utilizes GBlobs, a local point cloud feature descriptor specifically designed to enhance model generalization across diverse LiDAR configurations. Current LiDAR-based 3D detectors often suffer from a \enquote{geometric shortcut} when trained on conventional global features (\ie, absolute Cartesian coordinates). This introduces a position bias that causes models to primarily rely on absolute object position rather than distinguishing shape and appearance characteristics. Although effective for in-domain data, this shortcut severely limits generalization when encountering different point distributions, such as those resulting from varying sensor placements. By using GBlobs as network input features, we effectively circumvent this geometric shortcut, compelling the network to learn robust, object-centric representations. This approach significantly enhances the model's ability to generalize, resulting in the exceptional performance demonstrated in this challenge.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。