天宫二号多模态遥感场景分类数据集

多模态遥感场景分类数据集（MRSSC）是基于天宫二号宽波段成像仪和三维成像微波高度计获取的高质量对地观测数据构建的光学-SAR场景分类数据集。该数据集包含城市、农田、山区、沙漠、海岸、湖泊和河流共7个典型场景的12167幅图像。MRSSC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，该数据集将激发多模态遥感场景分类的创新研究思路和算法研究。

The Multimodal Remote Sensing Scene Classification Dataset (MRSSC) is an optical-SAR scene classification dataset constructed from high-quality Earth observation data collected by the Tiangong-2 Wide Band Imager and 3D Imaging Microwave Altimeter. This dataset contains 12,167 images covering 7 typical scenes: urban areas, farmland, mountainous regions, deserts, coastlines, lakes, and rivers. MRSSC provides a valuable data source for domestic and international scholars to conduct applied research on artificial intelligence-based domain transfer learning for remote sensing images, and will stimulate innovative research ideas and algorithmic studies for multimodal remote sensing scene classification.

搜集汇总

背景与挑战

背景概述

天宫二号多模态遥感场景分类数据集是基于天宫二号宽波段成像仪和三维成像微波高度计获取的高质量对地观测数据构建的光学-SAR场景分类数据集，包含7个典型场景的12167幅图像，旨在支持基于人工智能的遥感图像域迁移学习研究。