UrbanIng-V2X
收藏资源简介:
UrbanIng-V2X 是一个大规模的多模态数据集,支持涉及车辆和基础设施传感器的协同感知。该数据集由三个位于德国英戈尔施塔特的交叉路口收集,包括12个车载RGB相机、2个车载LiDAR、17个基础设施热像仪和12个基础设施LiDAR。所有序列均以10Hz的频率进行注释,具有跨越13个对象类别的3D边界框,总计约712k个注释实例。该数据集旨在解决自动驾驶车辆在复杂城市交叉路口中的感知问题,通过多车协同感知来克服视野限制和遮挡问题。
UrbanIng-V2X is a large-scale multimodal dataset that supports collaborative perception involving vehicle and infrastructure sensors. The dataset was collected at three intersections in Ingolstadt, Germany, and includes 12 on-board RGB cameras, 2 on-board LiDARs, 17 infrastructure-mounted thermal cameras and 12 infrastructure-mounted LiDARs. All sequences are annotated at a frequency of 10 Hz, with 3D bounding boxes covering 13 object categories and a total of approximately 712,000 annotated instances. This dataset is designed to address the perception challenges of autonomous vehicles at complex urban intersections, overcoming field-of-view limitations and occlusion issues via multi-vehicle collaborative perception.

- 1通过英戈尔施塔特技术大学,慕尼黑工业大学 · 2025年



