V2X-Real
收藏资源简介:
V2X-Real是由加州大学洛杉矶分校开发的用于车辆与万物协同感知的大型数据集。该数据集通过两辆自动驾驶车辆和两个智能基础设施收集,配备多种传感器如激光雷达和多视角摄像头。数据集包含33,000个激光雷达帧和171,000张相机图像,以及超过120万个标注的边界框,涵盖10个类别,主要用于解决城市环境中复杂的感知问题。数据集根据协同模式和自我视角分为四个子数据集,分别支持车辆中心、基础设施中心、车辆间和基础设施间的协同感知。
V2X-Real is a large-scale dataset developed by the University of California, Los Angeles (UCLA) for vehicle-to-everything (V2X) collaborative perception. This dataset is collected using two autonomous vehicles and two intelligent infrastructure nodes equipped with various sensors including LiDARs and multi-view cameras. The dataset contains 33,000 LiDAR frames, 171,000 camera images, and over 1.2 million annotated bounding boxes covering 10 categories, and is primarily designed to address complex perception tasks in urban environments. The dataset is divided into four subsets based on collaboration modes and ego perspectives, which respectively support vehicle-centric, infrastructure-centric, inter-vehicle, and inter-infrastructure collaborative perception.

- 1V2X-Real: a Largs-Scale Dataset for Vehicle-to-Everything Cooperative Perception加州大学洛杉矶分校 · 2024年



