SS3DM
收藏资源简介:
SS3DM是由清华大学和Horizon Robotics合作创建的一个用于街道视图表面重建的合成3D网格数据集。该数据集包含从CARLA模拟器导出的精确街道视图3D网格模型,适用于评估表面位置和法线。数据集内容包括多视角RGB视频序列、LiDAR点云以及详细的地面真值3D网格,共计13535帧。创建过程中,虚拟车辆配备了六台RGB相机和五台LiDAR传感器,在多样化的户外场景中采集数据。SS3DM主要应用于数字娱乐和自动驾驶模拟等领域,旨在解决街道视图表面重建中的精确评估问题。
SS3DM is a synthetic 3D mesh dataset for street-view surface reconstruction co-developed by Tsinghua University and Horizon Robotics. This dataset contains precise street-view 3D mesh models exported from the CARLA simulator, which is designed for evaluating surface positions and normals. It includes multi-view RGB video sequences, LiDAR point clouds, and detailed ground-truth 3D meshes, totaling 13,535 frames. During the dataset creation, virtual vehicles equipped with six RGB cameras and five LiDAR sensors collected data in diverse outdoor scenarios. SS3DM is mainly applied in fields such as digital entertainment and autonomous driving simulation, aiming to solve the problem of accurate evaluation in street-view surface reconstruction.




