Dur360BEV
收藏资源简介:
Dur360BEV是一个由杜伦大学计算机科学系和工程系创建的大规模真实世界自动驾驶数据集,包含360度球形RGB相机、高保真128通道3D LiDAR和RTK精炼的GNSS/INS系统。该数据集是首个配备球形相机模态的自动驾驶数据集,具有完全3D边界框注释。它旨在通过使用单个360度相机替代多相机系统来降低硬件复杂性,并针对自动驾驶中的鸟瞰图映射生成任务提供了基准架构。
Dur360BEV is a large-scale real-world autonomous driving dataset developed by the Departments of Computer Science and Engineering at Durham University. It is equipped with 360° spherical RGB cameras, high-fidelity 128-channel 3D LiDAR, and RTK-refined GNSS/INS systems. This is the first autonomous driving dataset with spherical camera modality, and it features complete 3D bounding box annotations. The dataset aims to reduce hardware complexity by replacing multi-camera systems with a single 360° camera, and provides a benchmark framework for the bird's-eye view mapping generation task in autonomous driving.




