Sentinel-2 dataset for ship detection

This database is a rework of the database built for the first phase of CORTEX project (https://esacortexproject.agenium-space.com/) funded by ESA. The creation process of the DB was redesigned, and new Sentinel2 acquisitions were included into the dataset. Like the previous database, it was generated using Sentinel-2 images distributed through the Copernicus open access hub (https://www.copernicus.eu/en, https://scihub.copernicus.eu/). These Sentinel-2 images are L1C products acquired in Danish sovereign waters in 2019, 2021 and 2022. More specifically, 31 10980x10980 L1C sentinel-2 products were selected (25 for the train dataset, 6 for the test dataset), each of them with a cloud coverage below 10% according to the cloud mask products. The database gathers small (80x80) tiles extracted from Sentinel-2 products. The DB description is given in AIS_Sentinel2_ships_db_CCN-CORTEX-TN-AS-011_1_0.pdf document attached to the zip file. This work is funded by a contract in the framework of the EO SCIENCE FOR SOCIETY PERMANENTLY OPEN CALL FOR PROPOSALS EOEP-5 BLOCK 4 issued by the European Space Agency.

本数据库是对欧洲空间局（ESA）资助的CORTEX项目第一阶段（https://esacortexproject.agenium-space.com/）所构建数据库的重构版本。数据库的创建流程已重新设计，并将新增的哨兵2号（Sentinel-2）影像数据纳入数据集。与前代数据库一致，本数据集所用的哨兵2号影像均通过哥白尼开放获取中心（Copernicus Open Access Hub，https://www.copernicus.eu/en, https://scihub.copernicus.eu/）分发获取。该批哨兵2号影像为L1C级产品，采集于2019、2021及2022年的丹麦主权海域。具体而言，本次共遴选31幅分辨率为10980×10980的哨兵2号L1C级产品（训练集25幅，测试集6幅），根据云掩膜产品测算，所有影像的云覆盖率均低于10%。本数据库包含从上述哨兵2号产品中裁切得到的80×80小尺寸瓦片。该数据库的详细说明可参阅压缩包附带的《AIS_Sentinel2_ships_db_CCN-CORTEX-TN-AS-011_1_0.pdf》文档。本研究依托欧洲空间局发布的"面向社会的地球观测科学永久开放提案征集EOEP-5 BLOCK 4"框架下的合同资助完成。

搜集汇总

背景与挑战

背景概述

该数据集是基于Sentinel-2 L1C图像的船舶检测数据集，包含31个高分辨率图像（25训练，6测试），云覆盖率低于10%，适用于深度学习和小卫星能力提升研究。数据集由AGENIUM Space创建，ESA资助，遵循非商业共享许可。