five

DurLAR (A High-Fidelity 128-Channel LiDAR Dataset with Panoramic Ambient and Reflectivity Imagery)

收藏
OpenDataLab2026-05-31 更新2024-05-09 收录
下载链接:
https://opendatalab.org.cn/OpenDataLab/DurLAR
下载链接
链接失效反馈
官方服务:
资源简介:
DurLAR 是一个高保真 128 通道 3D LiDAR 数据集,具有全景环境(近红外)和反射率图像,适用于多模式自动驾驶应用。与现有的自动驾驶任务数据集相比,DurLAR 具有以下新特性: 128 通道的高垂直分辨率激光雷达,是现有数据集的两倍,全 360 度深度,20-50m 处的范围精度为 ±2 cm。环境照明(近红外)和反射率全景图像以 Mono16 格式(2048 × 128 分辨率)提供,这是唯一提供此功能的数据集。没有滚动快门效果,因为我们的闪光灯 LiDAR 同时捕获所有 128 个通道。环境照明数据通过板载照度计记录,这在以前的数据集中同样不可用。高保真 GNSS/INS 可通过以 100 Hz 运行的板载 OxTS 导航单元获得,除了 GPS 之外,还可以接收来自多个 GNSS 星座的位置和时间数据。采用 KITTI 数据格式作为事实上的数据集格式,以便可以使用 DurLAR 开发套件和现有的 KITTI 兼容工具对其进行解析。重复位置的多样性,因此数据集是在不同的环境和天气条件下在相同的驾驶路线上收集的,并且一天中的时间相对于环境条件有额外的变化。传感器放置 LiDAR:Ouster OS1-128 LiDAR 传感器,具有 128 通道垂直分辨率立体相机:Carnegie Robotics MultiSense S21 立体相机,具有灰度、彩色和红外增强成像器,2048x1088 @ 2MP 分辨率 GNSS/INS:OxTS RT3000v3 全球导航卫星和惯性导航系统,支持 GPS、GLONASS、北斗、伽利略、PPP 和 SBAS 星座的定位 照度计:Yocto Light V3,USB 环境光传感器(照度计),可测量高达 100,000 勒克斯的环境光

DurLAR is a high-fidelity 128-channel 3D LiDAR dataset equipped with panoramic ambient near-infrared and reflectance imagery, tailored for multimodal autonomous driving applications. Compared to existing autonomous driving task datasets, DurLAR features the following novel characteristics: 1. 128-channel LiDAR with ultra-high vertical resolution (twice that of existing datasets), full 360° depth coverage, and a ranging accuracy of ±2 cm within the 20–50 m range. 2. Panoramic ambient near-infrared and reflectance images provided in Mono16 format with a resolution of 2048 × 128, making DurLAR the only dataset offering this unique capability. 3. No rolling shutter artifacts, as the flash LiDAR captures all 128 channels simultaneously. 4. Ambient light data recorded via an onboard illuminometer, a feature that is also unavailable in prior datasets. 5. High-fidelity GNSS/INS data sourced from the onboard OxTS navigation unit operating at 100 Hz, which receives positioning and timing data from multiple GNSS constellations in addition to GPS. 6. Adopting the KITTI data format as the de facto dataset standard, enabling parsing using the DurLAR development kit and existing KITTI-compatible tools. 7. Diverse repeated traversals: the dataset was collected along the same driving routes under varying environments and weather conditions, with additional temporal variations in ambient conditions corresponding to different times of day. ### Sensor Specifications - **LiDAR**: Ouster OS1-128 LiDAR sensor with 128-channel vertical resolution - **Stereo Camera**: Carnegie Robotics MultiSense S21 stereo camera with grayscale, color, and infrared-enhanced imagers, featuring a resolution of 2048×1088 @ 2MP - **GNSS/INS**: OxTS RT3000v3 Global Navigation Satellite and Inertial Navigation System, supporting positioning for GPS, GLONASS, BeiDou, Galileo, PPP, and SBAS constellations - **Illuminometer**: Yocto Light V3 USB ambient light sensor (illuminometer) capable of measuring ambient light up to 100,000 lux
提供机构:
OpenDataLab
创建时间:
2022-08-16
搜集汇总
数据集介绍
main_image_url
背景与挑战
背景概述
DurLAR是一个高保真128通道3D LiDAR数据集,提供全景环境(近红外)和反射率图像,适用于多模式自动驾驶应用。其独特之处在于高垂直分辨率激光雷达、无滚动快门效果的环境照明数据以及在不同环境和天气条件下的重复位置多样性。
以上内容由遇见数据集搜集并总结生成
5,000+
优质数据集
54 个
任务类型
进入经典数据集
二维码
社区交流群

面向社区/商业的数据集话题

二维码
科研交流群

面向高校/科研机构的开源数据集话题

数据驱动未来

携手共赢发展

商业合作