GEE支持下的河南省冬小麦面积提取及长势监测

[目的]使用遥感技术对2017—2020年河南省冬小麦的空间分布信息进行高精度的提取,然后对2020年冬小麦的长势进行高频度的监测并结合气象条件进行分析.[方法]本文基于谷歌地球引擎(Google Earth Engine,GEE)云平台,对选取的Landsat 8影像数据根据NDVI最大值进行合成,然后进行特征构建,...

Area Extraction and Growth Monitoring of Winter Wheat in Henan Province Supported by Google Earth Engine

【Objective】 The aim of this study was to use remote sensing technology to extract the spatial distribution information of winter wheat in Henan province from 2017 to 2020, and then to monitor the growth of winter wheat in 2020 with high frequency and to analyze the meteorological conditions. 【Method】 Based on the cloud platform of Google Earth engine (GEE), the selected Landsat 8 image data were synthesized according to the maximum value of NDVI, and then the features were constructed to add terrain features, texture features, NDVI and a new feature NDVI amplification. Random forest classification method was used to train the sample data according to the constructed features to extract the winter wheat planting area in Henan province from 2017 to 2020. The accuracy of the extracted winter wheat sown area was verified by confusion matrix and Henan statistical yearbook data. After accuracy verification, a mask was generated for the extracted winter wheat planting area in Henan province in 2020. In the mask area (winter wheat planting area) combined with MODIS high time resolution image data, the NDVI synchronization difference method was used to monitor the winter wheat growth from February to April in 2020. 【Result】 The GEE cloud platform could be used to quickly map the spatial distribution information of winter wheat planting areas in Henan province. Using random forest method to add terrain feature, texture feature, NDVI and new feature NDVI could effectively improve the extraction accuracy of winter wheat and reduce the relative error with statistical data. Based on confusion matrix, the average overall classification accuracy was 95.2%, the average kappa coefficient was 0.909, and the average classification accuracy of winter wheat was 95.3%. Compared with the statistical yearbook data of Henan province, the relative errors of winter wheat sown area extracted by this method in Henan province from 2017 to 2019 were all less than 3%. The average relative error of winter wheat sown area in the main planting areas of winter wheat in Henan province was less than 6%. MODIS image data combined with NDVI difference model could be used to monitor the growth of winter wheat in Henan province in 2020. The growth of winter wheat in Henan province was better than that of previous years and 2019 during the return to green period. In the later growth stage of winter wheat, the growth of most areas was the same as that of previous years and 2019. On the whole, the growth of winter wheat in 2020 was better than that of previous years and 2019. 【Conclusion】 The method proposed in this paper could carry out high-precision extraction and high-frequency growth monitoring of winter wheat in Henan province, and could provide a scientific basis for local governments or some agricultural departments in arranging and guiding agricultural activities.