中国冬小麦10m分辨率的季中种植分布数据集WinterWheat-CN10m (2025)

作物早期准确监测对作物生长监测和产量预测具有重要意义。我们提出了一种利用 Sentinel-1/2 时间序列图像构建分类规则的方法，并成功开发了一种稳健且高效的自动化季内冬小麦制图方法（AIWM）。该研究利用光学和雷达遥感影像时间序列，协同不同指数之间的关系，构建了基于光学时序自动季内冬季作物指数(Automatic in-season Winter Crop Index，AIWC)和雷达时序作物结构差异指数(Crop structure difference index, CSDI)指数，成功应用于中国冬小麦种植区识别。在无样本条件下，实现了跨区域和跨年份的大尺度、时空连续的农作物制图，形成了中国冬小麦季中种植分布数据集。我们将在每年4月中旬前更新并发布该数据产品。本数据集提供了中国冬小麦在生长季中期的种植分布信息，空间分辨率为 10 米。数据基于自动化季内冬小麦制图方法（AIWM）生成，该方法综合利用了 Sentinel-1 和 Sentinel-2 的时序影像。数据以栅格像元形式提供，其中数值 0 表示其他作物或非作物地类，数值 1 表示冬小麦像元。需要注意的是，由于 2025 年四川、湖北、甘肃、陕西等部分地区的 Sentinel-1 影像缺失，导致相应区域的制图结果不完整。

WinterWheat-CN10m-2025-01.tif → 新疆

WinterWheat-CN10m-2025-02.tif → 山东、河北、山西、河南、江苏、安徽、北京、天津

WinterWheat-CN10m-2025-03.tif → 湖北、四川、甘肃、陕西（但影像部分缺失）

Accurate early crop monitoring is of great significance for crop growth monitoring and yield prediction. We propose a method for constructing classification rules using Sentinel-1/2 time-series imagery, and have successfully developed a robust and efficient automated in-season winter wheat mapping method (AIWM). This study utilizes optical and radar remote sensing time-series imagery, integrates the relationships between different indices, and develops two indices: the Automatic in-season Winter Crop Index (AIWC) based on optical time series and the Crop Structure Difference Index (CSDI) based on radar time series, which have been successfully applied to the identification of winter wheat planting areas in China. Under the zero-shot setting, large-scale, spatially and temporally continuous crop mapping across regions and years was achieved, forming the China Winter Wheat Mid-season Planting Distribution Dataset. We will update and release this data product by mid-April each year. This dataset provides information on the mid-growing-season planting distribution of winter wheat in China, with a spatial resolution of 10 meters. The data is generated based on the automated in-season winter wheat mapping method (AIWM), which integrates time-series imagery from both Sentinel-1 and Sentinel-2. The data is provided in the form of raster pixels, where the value 0 represents other crops or non-crop land types, and the value 1 represents winter wheat pixels. It should be noted that due to missing Sentinel-1 imagery in some regions such as Sichuan, Hubei, Gansu, and Shaanxi in 2025, the mapping results for the corresponding areas are incomplete.

WinterWheat-CN10m-2025-01.tif → Xinjiang
WinterWheat-CN10m-2025-02.tif → Shandong, Hebei, Shanxi, Henan, Jiangsu, Anhui, Beijing, Tianjin
WinterWheat-CN10m-2025-03.tif → Hubei, Sichuan, Gansu, Shaanxi (partial imagery missing)

搜集汇总

背景与挑战

背景概述

该数据集是中国2025年冬小麦季中种植分布数据，空间分辨率为10米，基于Sentinel-1和Sentinel-2时间序列影像，采用自动季内冬小麦制图方法（AIWM）生成，在无样本条件下实现大规模连续制图，总体精度超过93%。但2025年部分地区因影像缺失导致制图不完整。