The public dataset for water segmentation using Sentinel-2 multi-spectral satallite imagery ##The dataset can be accessed through Google Drive or Baidu Net Disk for Chinese users with code:t6uy
${Sentinel-2 water body detection dataset}
|-- 2018.04
`-- |-- rgb_nir
| |-- rgb_nir.tif
`-- |-- swir
| |-- swir.tif
|-- 2018.12
`-- |-- rgb_nir
| |-- rgb_nir.tif
`-- |-- swir
| |-- swir.tif
|-- 2019.02
`-- |-- rgb_nir
| |-- rgb_nir.tif
`-- |-- swir
| |-- swir.tif
|-- label.tif
The dataset included three batches of data of same area (Chengdu City) which are collected on April 2018, December 2018 and February 2019 respectively. The pixelwise label is annotated based on The April's data.
rgb_nir.tif(20982 x 20982 x 4 pixels) contains 4 bands (r,g,b,nir) raw raster pixel information and is of 10m pixel resolution.
swir.tif(10491 x 10491 x 1 pixels) contains swir band raw raster pixel information and is in 20m pixel resolution.
Please cite our work if it's helpful to your research.
@article{yuan2021deep,
title={Deep learning-based multi-spectral satellite image segmentation for water body detection},
author={Yuan, Kunhao and Zhuang, Xu and Schaefer, Gerald and Feng, Jianxin and Guan, Lin and Fang, Hui},
journal={IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing},
year={2021},
publisher={IEEE}
}