DurLAR
收藏资源简介:
DurLAR是由杜伦大学创建的高保真128通道3D LiDAR数据集,专为多模态自动驾驶应用设计。该数据集包含全景环境(近红外)和反射率图像,以及用于深度估计的基准任务样本。数据集通过装备有高分辨率128通道LiDAR、2MPix立体相机、lux计和GNSS/INS系统的驾驶平台收集。DurLAR不仅提供了高分辨率的LiDAR数据,还包括同步的环境光照信息,适用于各种天气和光照条件下的自动驾驶任务。数据集的创建旨在通过提供高分辨率、稀疏的真实场景深度信息,推动单目深度估计技术的发展,特别是在极端天气和光照变化下的性能评估。
DurLAR is a high-fidelity 128-channel 3D LiDAR dataset created by Durham University, designed specifically for multimodal autonomous driving applications. This dataset includes panoramic (near-infrared) and reflectance images, as well as benchmark task samples for depth estimation. It was collected via a driving platform equipped with a high-resolution 128-channel LiDAR, a 2MPix stereo camera, a lux meter, and a GNSS/INS system. DurLAR not only provides high-resolution LiDAR data, but also includes synchronized ambient lighting information, making it suitable for autonomous driving tasks under various weather and lighting conditions. The dataset was developed to advance monocular depth estimation technologies by offering high-resolution, sparse real-world scene depth information, particularly for performance evaluation under extreme weather and lighting variations.




