2019-2020年度东南亚30m分辨率水田数据集

依托Google Earth Engine，采用种渐进式的精细分辨率水稻提取方法，提取2020年东南亚地区的10米分辨率水稻分布数据。首先，基于2000-2020年的500m分辨率的MOD09数据集，提取了NDVI、EVI和LSWI时间序列数据，然后对2000-2020年移栽-淹水期的NDVI、EVI和LSWI的值生成水田分布的潜在区；在此基础上，采用随机森林分类器（树数=300，叶数=5），以Sentinel-1的逐月VH特征和Sentinel-2的百分位特征作为输入，按照37个分区逐个开展监督分类，在潜在水稻空间分布的基础上，获取10米分辨率的水稻空间分布。

Based on Google Earth Engine, we produced 10-meter resolution rice distribution data for Southeast Asia in 2020 using a progressive high-resolution rice extraction method. First, we extracted time-series datasets of Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), and Land Surface Water Index (LSWI) based on the 500-meter resolution MOD09 dataset from 2000 to 2020. Next, we generated potential rice paddy distribution areas using the NDVI, EVI, and LSWI values during the transplanting-flooding period from 2000 to 2020. Subsequently, we conducted supervised classification individually for each of the 37 zoning regions using a random forest classifier (with 300 decision trees and 5 leaf nodes per tree), taking monthly VH features from Sentinel-1 and percentile features from Sentinel-2 as input. Finally, we obtained the 10-meter resolution spatial distribution of rice based on the acquired potential rice spatial distribution.

搜集汇总

背景与挑战

背景概述

该数据集是2019-2020年度东南亚地区的10米分辨率水田分布数据，采用渐进式精细分辨率水稻提取方法，结合多源遥感数据和随机森林分类器生成。数据由中国科学院空天信息创新研究院制作，存储容量为2.11 GB，包含952个GeoTiff格式文件。