A 10-m resolution impervious surface area map for the greater Mekong subregion from multisource remote sensing images

Here, the 2016-2022 impervious surface area map exclusively for the greater Mekong subregion is prepared. We present a novel machine-learning framework implemented on the Google Earth Engine platform that merges Sentinel-1 Synthetic Aperture Radar images and Sentinel-2 Multispectral images to extract impervious surface area. Furthermore, we also introduce a training sample migration strategy that eliminates the need for collecting additional training samples and automates multi-temporal impervious surface area mapping. Finally, we perform a quantitative assessment with validation samples interpreted from Google Earth. Results show that the overall accuracy and kappa coefficient of the final impervious surface area maps range from 93.67% to 93.75% and 0.872 to 0.873, respectively. This dataset provides comprehensive measurements of impervious surface coverage and configuration that will help to inform urban ecological studies.

本研究制作了专属于大湄公河次区域（Greater Mekong Subregion）的2016-2022年不透水面（Impervious Surface Area）分布图。本研究提出了一种部署于谷歌地球引擎（Google Earth Engine）平台的新型机器学习框架，该框架融合哨兵1号（Sentinel-1）合成孔径雷达（Synthetic Aperture Radar）影像与哨兵2号（Sentinel-2）多光谱影像以提取不透水面信息。此外，本研究还引入了一种训练样本迁移策略，无需额外采集训练样本即可实现多时序不透水面分布的自动化制图。最后，本研究采用从谷歌地球（Google Earth）解译得到的验证样本开展定量评估。结果表明，最终不透水面分布图的总体精度与卡帕系数（Kappa Coefficient）分别介于93.67%至93.75%以及0.872至0.873之间。本数据集提供了不透水面覆盖度与空间格局的全面量化数据，可为城市生态学相关研究提供数据支撑。