cyberagent/VLG-Loc-Dataset

视觉语言全局定位（VLG-Loc）数据集用于评估机器在视觉和语言理解基础上进行全局定位的能力。数据集包含来自机器人不同摄像头的图像、元数据文件和点云数据。这些数据来源于多种环境，包括具有统一或多样几何和外观的室内环境，以及真实的零售商店环境。

The Vision-Language Global Localization (VLG-Loc) Dataset is designed for evaluating the robots ability to perform global localization based on visual and linguistic understanding. The dataset includes images from different robot cameras, metadata files, and point cloud data. These data are sourced from various environments, including indoor environments with uniform or diverse geometry and appearance, as well as real-world retail store environments.