2016-2019年天宫二号多模态遥感场景分类数据集

天宫二号多模态遥感场景分类数据集是首个光学-SAR图像遥感场景分类数据集，包含城镇、荒漠、农田、湖泊、河流、山脉、沿海等7个场景类别的14000多张图像，具有多时、多空间分辨率和多模态等特点，场景呈现出高度的多样性和复杂的异质性，为国内外学者开展基于人工智能+遥感图像的域迁移学习的应用研究提供数据源。

The Tiangong-2 Multimodal Remote Sensing Scene Classification Dataset is the first optical-SAR image remote sensing scene classification dataset. It contains over 14,000 images covering seven scene categories including urban, desert, farmland, lake, river, mountain, and coastal areas. Characterized by multi-temporal, multi-spatial resolution and multimodal properties, the dataset's scenes exhibit high diversity and complex heterogeneity. It provides valuable data sources for domestic and international scholars to conduct applied research on domain transfer learning based on artificial intelligence and remote sensing images.

搜集汇总

背景与挑战

背景概述

该数据集是首个光学SAR图像遥感场景分类数据集，包含2016-2019年间超过14000张图像，覆盖城镇、沙漠、农田等7个场景类别，具有多时相、多分辨率和多模态特点，场景多样且异质性强，为基于人工智能的遥感图像领域迁移学习研究提供数据支持。