Monado SLAM dataset
收藏资源简介:
Monado SLAM数据集是一个由慕尼黑工业大学、慕尼黑机器学习中心和阿根廷Rosario的CIFASIS机构合作创建的数据集,用于评估视觉惯性里程计和同时定位与建图(VIO/SLAM)系统。该数据集包含来自三种不同的虚拟现实头戴设备的真实序列,包括Valve Index、HP Reverb G2和Samsung Odyssey+。数据集共包含64个非校准记录,总时长为5小时15分钟,涵盖了各种具有挑战性的场景,如高强度运动、动态遮挡、长时间跟踪、低纹理区域和不良光照条件等。数据集以CC BY 4.0许可证发布,旨在推动VIO/SLAM研究和开发。
The Monado SLAM Dataset is a collaborative dataset developed by Technische Universität München, the Munich Center for Machine Learning, and CIFASIS in Rosario, Argentina, aimed at evaluating Visual-Inertial Odometry and Simultaneous Localization and Mapping (VIO/SLAM) systems. This dataset contains real-world sequences collected from three distinct virtual reality (VR) head-mounted displays, namely Valve Index, HP Reverb G2 and Samsung Odyssey+. It consists of 64 uncalibrated recordings with a total duration of 5 hours and 15 minutes, covering a wide range of challenging scenarios including high-intensity motion, dynamic occlusions, long-duration tracking, low-texture regions and poor lighting conditions. The dataset is released under the CC BY 4.0 license, with the goal of advancing VIO/SLAM research and development.

- 1The Monado SLAM Dataset for Egocentric Visual-Inertial TrackingTechnical University of Munich, Munich Center for Machine Learning, Collabora Ltd., CIFASIS, CONICET-UNR · 2025年



