University-1652
收藏资源简介:
University-1652是一个多视角多源基准数据集,用于无人机地理定位研究。该数据集包含来自三个平台的数据:合成无人机、卫星和地面摄像机,涵盖全球72所大学的1652座建筑。数据集特点包括多视角图像、多源数据和大规模数据量,平均每个位置有71.64张图像。主要应用领域为无人机视角目标定位和无人机导航,旨在解决跨视角地理定位问题,通过无人机视角图像预测目标位置,以及根据卫星视角查询图像驱动无人机到达感兴趣区域。
University-1652 is a multi-view, multi-source benchmark dataset for drone geolocalization research. It contains data from three platforms: synthetic drone, satellite, and ground camera, covering 1652 buildings across 72 universities worldwide. The dataset features multi-view imagery, multi-source data, and large-scale data volume, with an average of 71.64 images per location. Its main application areas are drone-view target localization and drone navigation, which aim to solve cross-view geolocalization problems: predicting target locations through drone-view images, and driving drones to reach regions of interest using satellite-view query images.




