CrossEarth-SAR-200K
收藏资源简介:
CrossEarth-SAR-200K是由上海交通大学等机构联合构建的首个十亿级SAR视觉基础模型配套数据集,集成了公开与私有SAR影像资源,包含20万张弱监督与全监督标注图像,覆盖多传感器、多地域的复杂场景。该数据集通过统一标注标准整合了37万私有数据和126万公开数据,支持SAR图像的跨域语义分割任务,旨在解决SAR影像因成像机制差异导致的域偏移问题,为灾害监测、环境管理等应用提供通用化基础模型训练资源。
CrossEarth-SAR-200K is the first billion-scale supporting dataset for SAR visual foundation models, jointly constructed by Shanghai Jiao Tong University and other institutions. It integrates both public and private SAR image resources, contains 200,000 weakly-supervised and fully-supervised labeled images, and covers complex scenarios across multiple sensors and geographical regions. This dataset integrates 370,000 private data and 1.26 million public data via unified annotation standards, supports cross-domain semantic segmentation tasks for SAR images, aims to address the domain shift problem caused by differences in imaging mechanisms of SAR imagery, and provides general training resources for foundation models for applications such as disaster monitoring and environmental management.
CrossEarth-SAR-Benchmarks 数据集概述
基本信息
- 数据集名称: CrossEarth-SAR-Benchmarks
- 托管平台: Hugging Face Datasets
- 许可证: MIT License
数据描述
- 该数据集详情页面未提供关于数据内容、规模、格式、任务类型或具体用途的详细描述。
文件与访问
- 数据集页面地址:https://huggingface.co/datasets/conquer997/CrossEarth-SAR-Benchmarks
- 根据页面信息,未列出具体的文件结构或数据样本。




