4Seasons

4Seasons数据集是由慕尼黑工业大学创建，专注于自动驾驶中的多天气同步定位与地图构建（SLAM）。该数据集包含超过350公里的记录，涵盖九种不同环境，如城市、乡村和高速公路，以及多种天气和光照条件。数据集通过融合直接立体视觉惯性里程计与实时运动学全球导航卫星系统（RTK-GNSS），提供高达厘米级的全球一致参考姿态。此数据集旨在解决自动驾驶中长期定位的挑战，支持视觉里程计、全局地点识别和基于地图的重定位跟踪等研究。

The 4Seasons Dataset was developed by the Technical University of Munich, focusing on multi-weather simultaneous localization and mapping (SLAM) for autonomous driving. The dataset contains over 350 kilometers of recorded data, covering nine distinct scenarios including urban, rural, and highway environments, as well as diverse weather and lighting conditions. It provides centimeter-level globally consistent reference poses by fusing direct stereo visual-inertial odometry and real-time kinematic global navigation satellite system (RTK-GNSS) data. This dataset aims to address the challenges of long-term localization in autonomous driving, supporting research in fields such as visual odometry, global place recognition, and map-based relocalization tracking.