LaMAR

LaMAR数据集是由苏黎世联邦理工学院计算机科学系和微软混合现实与AI实验室合作创建的一个大规模数据集，专门用于增强现实中的定位和映射技术研究。该数据集通过使用增强现实设备（如HoloLens 2和iPhone）以及激光扫描仪捕获，覆盖了45,000平方米的区域，并记录了超过100小时的多传感器数据流。数据集不仅包含室内外图像，还包括光照和语义变化以及动态对象，旨在解决现有基准数据集在规模、多样性和真实性方面的不足。通过这个数据集，研究人员可以评估和改进AR设备在复杂环境中的定位和映射能力，推动AR技术的进一步发展。

The LaMAR dataset is a large-scale dataset co-created by the Department of Computer Science, ETH Zurich, and Microsoft Mixed Reality & AI Lab, specifically dedicated to research on localization and mapping technologies in augmented reality (AR). Captured using AR devices such as HoloLens 2 and iPhone, as well as laser scanners, it covers an area of 45,000 square meters and records over 100 hours of multi-sensor data streams. The dataset not only contains indoor and outdoor images, but also includes illumination and semantic variations as well as dynamic objects, aiming to address the shortcomings of existing benchmark datasets in terms of scale, diversity and authenticity. Through this dataset, researchers can evaluate and improve the localization and mapping capabilities of AR devices in complex environments, thereby promoting further development of AR technology.