雷达与视觉摄像机外参自动标定与坐标对齐性能评价数据集

该数据集包含雷达-相机外参自动标定所需要的真值，包括雷达点云与图像像素的对应点、毫米波雷达点云数据、雷达轨迹数据、提取的候选车辆毫米波雷达轨迹、提取的视觉轨迹和一些中间文件。1、2、3文件名分别代表3次实验数据，点云（radarpcl.npy）是毫米波雷达点云数据，轨迹（radartrack.npy)是雷达轨迹数据，objpointset.npy是提取的候选车辆毫米波雷达轨迹，timestamps.txt是提取的视觉轨迹，JSON用于记录相机内参，以及保存标定结果，其余类型为产生的中间文件。This dataset contains the ground truth required for automatic calibration of radar-camera extrinsic parameters, including corresponding point pairs between radar point clouds and image pixels, millimeter-wave radar point cloud data, radar trajectory data, extracted candidate vehicle millimeter-wave radar trajectories, extracted visual trajectories, and several intermediate files. The filenames 1, 2 and 3 represent the data from three experiments respectively. Specifically, `radarpcl.npy` stores the millimeter-wave radar point cloud data, `radartrack.npy` stores the radar trajectory data, `objpointset.npy` is the extracted candidate vehicle millimeter-wave radar trajectory set, `timestamps.txt` contains the extracted visual trajectories, JSON files are used to record camera intrinsic parameters and save calibration results, and the remaining file types are generated intermediate files.

搜集汇总

背景与挑战

背景概述

该数据集旨在支持雷达与视觉摄像机的外参自动标定和坐标对齐性能评价，包含雷达点云与图像像素对应点、毫米波雷达点云和轨迹数据、提取的车辆雷达轨迹和视觉轨迹等真值文件。数据集由华中科技大学发布，数据量为27.46MB，包含21个文件，属于国家重点研发计划项目'雄安新区交通设施数字化建设示范应用'的一部分。