Major-TOM/Core-S1RTC-SSL4EO

Core-S1RTC-SSL4EO数据集是基于Sentinel-1 RTC雷达数据的全球覆盖数据集，使用了SSL4EO-ResNet50-MOCO模型提取嵌入。数据集包含36,748,875个嵌入，总大小为332.5 GB。该数据集是CloudFerro和欧洲空间局（ESA）的Φ-lab合作的结果，旨在提供开放和免费的向量化扩展，并定义发布嵌入扩展的标准化方式。嵌入提取自常见的AI模型，使得浏览和导航大型数据集如Major TOM时，存储和计算需求降低。数据集的计算是在CloudFerro提供的GPU加速实例上进行的，使用了CREODIAS云服务平台。

The Core-S1RTC-SSL4EO dataset is an embedding dataset based on the Sentinel-1 radar dataset, containing 36,748,875 embedding vectors. These embedding vectors are extracted from the Major TOM Core-S1RTC dataset using the SSL4EO-ResNet50-MOCO model. Each sample in the dataset contains multiple fields such as unique ID, embedding vector, grid cell, product ID, timestamp, etc., to describe geospatial information. The purpose of the dataset is to provide an efficient way to browse and navigate large-scale Earth observation data, reducing storage and computational demands. The generation and release of the dataset were completed in collaboration between CloudFerro and ESA Φ-lab.