OpenSatMap 高分辨率卫星数据集

OpenSatMap 是一个高分辨率卫星数据集，专为大规模地图构建而设计。该数据集由中国科学院自动化研究所 (CASIA) 、中国科学院香港信息系统研究所人工智能与机器人研究中心、中国科学院大学 (UCAS) 、腾讯地图和北京邮电大学于 2024 年发布，相关论文成果为「OpenSatMap: A Fine-grained High-resolution Satellite Dataset for Large-scale Map Construction」，已被 NeurIPS 2024 接受。

OpenSatMap is a high-resolution satellite dataset specifically designed for large-scale map construction. This dataset was released in 2024 by the Institute of Automation, Chinese Academy of Sciences (CASIA), the Artificial Intelligence and Robotics Research Center of the Hong Kong Institute of Information Systems, Chinese Academy of Sciences, University of Chinese Academy of Sciences (UCAS), Tencent Maps, and Beijing University of Posts and Telecommunications. The associated paper titled "OpenSatMap: A Fine-grained High-resolution Satellite Dataset for Large-scale Map Construction" has been accepted by NeurIPS 2024.

搜集汇总

背景与挑战

背景概述

OpenSatMap是一个高分辨率卫星数据集，专为大规模地图构建而设计，具有细粒度的实例级注释和高达20级的分辨率（0.15米/像素），覆盖全球50多个城市和18个国家的多样地理位置。该数据集由中国科学院自动化研究所等机构联合发布，包含3,787张高分辨率卫星图像，适用于自动驾驶、图像识别和遥感等领域的研究。