RADIal

该数据集名为RADIal，提供了约2小时的驾驶过程中同步的雷达、激光雷达和摄像头图像。该数据集重点关注雷达数据，并包含了车辆检测和自由空间分割的标签。雷达传感器是一款具有12个发射天线和16个接收天线的成像雷达，其视野范围达到180度，探测距离大约为100米。数据集包含了城市、乡村和高速公路等驾驶场景的平衡分布。规模上，该数据集包含了25,000个同步数据帧，其中大约8,300帧带有标签。任务方面，主要包括物体检测和自由空间分割。

This dataset, named RADIal, contains approximately 2 hours of synchronized radar, LiDAR, and camera imagery collected during driving. It prioritizes radar data and includes annotations for vehicle detection and free-space segmentation. The radar sensor used is an imaging radar with 12 transmitting antennas and 16 receiving antennas, boasting a 180-degree field of view and a detection range of approximately 100 meters. The dataset features a balanced distribution of driving scenarios including urban, rural, and highway environments. In terms of scale, the dataset consists of 25,000 synchronized data frames, of which approximately 8,300 are annotated. The primary downstream tasks supported by this dataset are object detection and free-space segmentation.

搜集汇总

背景与挑战

背景概述

RADIal是一个多传感器自动驾驶数据集，包含2小时的雷达、激光雷达和相机同步数据，覆盖多种驾驶环境，并提供8,252帧的车辆标注数据。数据集还包含GPS和车辆CAN总线信息，支持多任务学习研究。