用于酒店服务机器人训练的三维场景数据

本数据集专为训练在酒店环境中执行服务的机器人而设计，重点提升其在高标准、结构化且需保障客人隐私的空间中的任务执行、人机协作与环境适应性。数据集精细构建了标准酒店客房及其外部走廊的完整三维环境。场景内包含床（含床品）、床头柜、衣柜、书桌、座椅、窗帘、灯具、挂画、电视、迷你冰箱、卫生间入口等标准配置。我们特别模拟了酒店场景的高标准化布局（如床与床头柜的相对位置）、多种客房状态（如“已清洁”、“入住中”、“勿扰”）、以及动态出现的障碍物（如客人的行李箱、打开的衣柜门），以全方位复现真实酒店服务过程中机器人可能面临的各种情况本算法旨在处理三维模型，通过一系列步骤实现模型的分割、实例重组及格式转换，以生成新的实例模型，用于场景渲染和机器人训练等应用。

1.模型分割：本步骤接收任意初始三维模型作为输入，三维模型包括位置、尺寸、材质、顶点信息、法相信息、面片信息字段，运用拓扑连通性聚类算法将该组合模型拆分为多个面片组（face group），获取模型类型字段。此步骤有效提取模型的结构特征，有助于后续的实例重组。

2.模型实例重组：在此步骤中，对三维模型的位置、尺寸、材质、顶点信息、法相信息、面片信息字段进行分割，再利用Qwen-VL-Max和GroundingDino算法对分割后的部件进行组合，形成独立的模型实例，并获取其中的标签字段。标签字段能够使每个模型实例能够基于原模型的结构和信息进行识别和应用。

3.模型格式转换：本步骤将拆分获得的实例模型及其对应的材质信息转换为OpenUSD格式，并获取其中的碰撞体设置信息字段和动画约束信息字段，以使模型能够在场景中动起来。

通过以上步骤，将原本数据库中的模型进行重组，生成新的实例模型，并被组装成一个完整的场景，以满足场景渲染、机器人训练等多个应用需求。

This dataset is specifically designed for training service robots operating in hotel environments, with the goal of enhancing their task execution capabilities, human-robot collaboration, and environmental adaptability in spaces with high operational standards, structured layouts, and strict guest privacy protection requirements. It meticulously constructs a complete 3D environment of a standard hotel guest room and its external corridor. The scene includes standard hotel fixtures such as beds (with bedding), nightstands, wardrobes, desks, chairs, curtains, lamps, wall art, televisions, mini-fridges, and bathroom entrances. We have specially simulated the highly standardized layout of hotel scenarios (e.g., the relative positions of beds and nightstands), multiple guest room states (e.g., "cleaned", "occupied", "do not disturb"), and dynamically appearing obstacles (e.g., guests' suitcases, open wardrobe doors), to comprehensively replicate all possible scenarios that robots may encounter during real hotel service operations.

The accompanying 3D model processing pipeline is designed to generate new instance models through three core steps: model segmentation, instance recombination, and format conversion, which can be applied to scene rendering, robot training and other use cases.

1. Model Segmentation: This step takes any initial 3D model as input, which contains fields including position, size, material, vertex information, normal information, and face information. A topological connectivity clustering algorithm is employed to split the combined model into multiple face groups, and extract the model type field. This step effectively extracts the structural features of the model, laying a foundation for subsequent instance recombination.
2. Model Instance Recombination: In this step, the position, size, material, vertex information, normal information, and face information fields of the 3D model are first segmented. Then, the Qwen-VL-Max and GroundingDino algorithms are used to combine the segmented components into independent model instances, and the corresponding label field is obtained. The label field allows each model instance to be identified and applied based on the structure and information of the original model.
3. Model Format Conversion: This step converts the split instance models and their corresponding material information into the OpenUSD format, and extracts the collider setting information field and animation constraint information field, enabling the models to be animated within the scene.

Through the above steps, the original 3D models in the dataset are recombined to generate new instance models, which are then assembled into a complete scene to meet the requirements of multiple applications such as scene rendering and robot training.

搜集汇总

背景与挑战

背景概述

该数据集是一个专为酒店服务机器人训练设计的三维场景数据集合，包含90.45条数据，以zip格式提供，数据结构涵盖编号、标签、服务可及性、隐私等级、精准交互点坐标等12个字段，详细描述了酒店客房中的物体属性、状态和交互点。数据集模拟了标准酒店环境的布局、多种客房状态（如“已清洁”、“入住中”）以及动态障碍物，旨在提升机器人在结构化、高隐私要求空间中的任务执行能力和环境适应性。通过算法流程进行模型分割、实例重组和格式转换，生成OpenUSD格式的实例模型，支持场景渲染和机器人训练应用。