OpenRobotLab/GRScenes

GRScenes数据集是一个大规模、动态、真实的环境数据集，包含超过10万高质量的场景原型，共99个环境（69个家庭场景和30个商业场景）。每个场景都经过精心设计和精确建模以确保真实性。数据集中的对象具有完整的内部建模，支持机器人进行开门、拉抽屉等真实动作。此外，数据集为每个场景中的对象设置了对象级别的语义标签，支持局部导航和对象搜索等任务。

GRScenes dataset is a large-scale, dynamic, and realistic environment dataset, built from over 100k high-quality scene prototypes and now includes 99 environments (69 home scenes and 30 commercial scenes). Each scene has been carefully designed and accurately modeled for realism. The objects within the dataset have full internal modeling, supporting robots in performing realistic actions such as opening doors or sliding drawers. Additionally, the dataset sets object-level semantic labels in each scene, supporting tasks like local navigation and object searching.