DOFS
收藏资源简介:
DOFS数据集由香港中文大学机械与自动化工程系创建,是一个包含3D可变形物体(如弹塑性物体)的实验数据集,提供了完整的空间信息(顶部、侧面和底部信息)。数据集包含主动操作动作、多视角RGB-D图像、良好配准的点云、3D变形网格和3D占用信息。数据集通过一个低成本且创新的数据收集平台生成,该平台使用透明操作平面和6个RGB-D相机进行多视角数据采集。DOFS数据集主要用于机器人对3D可变形物体的动力学模型学习,旨在解决现实世界中机器人操作可变形物体的挑战。
The DOFS dataset was created by the Department of Mechanical and Automation Engineering of The Chinese University of Hong Kong. It is an experimental dataset containing 3D deformable objects such as elasto-plastic objects, providing complete spatial information including top, side and bottom spatial details. The dataset includes active manipulation actions, multi-view RGB-D images, well-registered point clouds, 3D deformation meshes and 3D occupancy information. It is generated via a low-cost and innovative data collection platform, which adopts a transparent operating plane and six RGB-D cameras for multi-view data acquisition. The DOFS dataset is mainly used for robotic learning of dynamic models of 3D deformable objects, aiming to address the challenges in real-world robotic manipulation of deformable objects.

- 1DOFS: A Real-world 3D Deformable Object Dataset with Full Spatial Information for Dynamics Model Learning香港中文大学机械与自动化工程系 · 2024年



