通过编辑真实场景的物体布局，生成逼真的合成激光雷达数据。
LiDAR-EDIT: LiDAR Data Generation by Editing the Object Layouts in Real-World Scenes
- 在真实激光雷达扫描上修改物体位置与类型，保持背景真实感。
- 生成数据保留正确投影几何，支持创建与原场景差异大的反事实场景。
- 适合自动驾驶仿真、数据增强及需要精确物体标注的任务。
我们提出 LiDAR-EDIT，一种面向自动驾驶的合成激光雷达数据生成新范式。该框架通过修改真实世界激光雷达扫描中的物体布局，在保持背景环境真实性的前提下生成新数据。相比从零生成点云的端到端方法，LiDAR-EDIT 允许用户完全控制物体数量、类型和姿态；相比新视角合成技术，本方法可构建与原始场景显著不同的反事实场景。通过球形体素化确保生成点云符合激光雷达投影几何。在物体移除与插入过程中，利用生成模型填补原始扫描中被遮挡的背景与物体部分。实验表明，该框架生成的激光雷达扫描具有高保真度，对下游任务具备实际应用价值。
原文摘要 · Abstract (English)
We present LiDAR-EDIT, a novel paradigm for generating synthetic LiDAR data for autonomous driving. Our framework edits real-world LiDAR scans by introducing new object layouts while preserving the realism of the background environment. Compared to end-to-end frameworks that generate LiDAR point clouds from scratch, LiDAR-EDIT offers users full control over the object layout, including the number, type, and pose of objects, while keeping most of the original real-world background. Our method also provides object labels for the generated data. Compared to novel view synthesis techniques, our framework allows for the creation of counterfactual scenarios with object layouts significantly different from the original real-world scene. LiDAR-EDIT uses spherical voxelization to enforce correct LiDAR projective geometry in the generated point clouds by construction. During object removal and insertion, generative models are employed to fill the unseen background and object parts that were occluded in the original real LiDAR scans. Experimental results demonstrate that our framework produces realistic LiDAR scans with practical value for downstream tasks.
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。