UAVScenes
收藏资源简介:
UAVScenes 是一个大规模的多模态数据集,专为无人机(UAV)感知任务而设计,提供了包括相机图像和激光雷达点云在内的帧级语义标注,以及准确的六自由度(6-DoF)位姿和重建的3D地图。该数据集包含超过12万帧的语义标注数据,支持包括语义分割、深度估计、6-DoF定位、场景识别和新视图合成(NVS)在内的多种多模态感知任务。UAVScenes 数据集的创建旨在填补现有无人机数据集在帧级标注和多模态感知方面的空白,为无人机感知研究提供了一个全面的基准测试平台。
UAVScenes is a large-scale multimodal dataset specifically designed for unmanned aerial vehicle (UAV) perception tasks. It provides frame-level semantic annotations including camera images and LiDAR point clouds, as well as accurate 6-degree-of-freedom (6-DoF) poses and reconstructed 3D maps. This dataset contains over 120,000 frames of semantically annotated data, supporting a variety of multimodal perception tasks such as semantic segmentation, depth estimation, 6-DoF localization, scene recognition, and novel view synthesis (NVS). UAVScenes was developed to fill the gaps in existing UAV datasets regarding frame-level annotations and multimodal perception, offering a comprehensive benchmark platform for UAV perception research.
UAVScenes数据集概述
基本信息
- 数据集名称: UAVScenes
- 发布会议: ICCV 2025
- 论文链接: arXiv:2507.22412
- 数据集类型: 多模态无人机数据集
- 主要特点:
- 同时包含2D和3D模态数据
- 提供手动标注的语义标签(图像和LiDAR点云)
- 包含精确的6自由度位姿信息
数据内容
- 数据来源: 基于MARS-LVIG数据集扩展
- 标注工具:
- 2D标注: X-AnyLabeling
- 3D标注: CloudCompare
- 3D重建: DJI Terra
- 数据类型:
- 海康威视相机图像及标注
- Livox Avia LiDAR点云及标注
- 6-DoF位姿信息
- 重建的3D点云/网格地图
- 标注规模: 12万对图像和点云标注对
数据集结构
- 包含场景: 4个大型场景
- AMtown
- AMvalley
- HKairport
- HKisland
- 每个场景包含: 多个运行序列(如01,02,03等)
下载信息
- 版本选择:
- 完整数据集(interval=1)
- 关键帧数据集(interval=5,大小为1/5)
- 下载平台:
基准代码
- 正在准备中
引用格式
bibtex @article{Wang2025UAVScenes, title = {UAVScenes: A Multi‑Modal Dataset for UAVs}, author = {Wang, Sijie and Li, Siqi and Zhang, Yawei and Yu, Shangshu and Yuan, Shenghai and She, Rui and Guo, Quanjiang and Zheng, JinXuan and Howe, Ong Kang and Chandra, Leonrich and Srijeyan, Shrivarshann and Sivadas, Aditya and Aggarwal, Toshan and Liu, Heyuan and Zhang, Hongming and Chen, Chujie and Jiang, Junyu and Xie, Lihua and Tay, Wee Peng}, journal = {arXiv preprint arXiv:2507.22412}, year = {2025}, url = {https://arxiv.org/abs/2507.22412} }
许可信息
- 许可证类型: Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License
- 使用限制: 仅限学术用途

- 1UAVScenes: A Multi-Modal Dataset for UAVs南洋理工大学 · 2025年



