UAVPairs
收藏资源简介:
UAVPairs数据集是一个包含21,622张高分辨率图像的挑战性基准数据集,涵盖30个不同的场景,用于大规模无人机图像的匹配对检索。数据集通过基于SfM的3D重建技术自动标注,确保了图像对的真正匹配性。该数据集旨在解决现有数据集在分辨率和内容上与无人机图像存在显著差异的问题,并通过3D重建过滤掉了几乎所有的错误匹配图像。UAVPairs数据集为大规模无人机图像的匹配对检索提供了一个有效的解决方案,有助于提高图像检索模型的准确性和3D模型重建的质量。
The UAVPairs dataset is a challenging benchmark dataset comprising 21,622 high-resolution images spanning 30 distinct scenarios, tailored for large-scale UAV image matching pair retrieval. It is automatically annotated using SfM-based 3D reconstruction techniques, guaranteeing the authentic matching validity of the image pairs. This dataset addresses the significant discrepancy between existing datasets and actual UAV imagery in terms of resolution and content, and has filtered out nearly all mismatched image pairs through 3D reconstruction. The UAVPairs dataset provides an effective solution for large-scale UAV image matching pair retrieval, helping to improve the accuracy of image retrieval models and the quality of 3D model reconstruction.
UAVPairs数据集概述
数据集简介
- 名称:UAVPairs
- 类型:大型无人机图像匹配对检索基准数据集
- 特点:具有挑战性的基准测试集
数据用途
- 主要用于大型无人机图像匹配对检索任务
当前状态
- 开发中(Comming soon...)




