基于MODIS EVI时序数据的冬小麦种植面积提取

Extraction of Winter Wheat Planting Area Based on Data of MODIS EVI Time-series

基于MODIS EVI构建的时间序列谱,结合冬小麦生育期形态变化特征,提取了河南省冬小麦种植面积。结果表明,冬小麦在EVI特征空间中具有独特的序列谱相特征,返青期后冬小麦整体EVI表现为逐渐升高,EVI在开花期后下降,灌浆后期快速降低;采用决策树分类方法实现冬小麦分离提取,整个河南省冬小麦识别面积与官方统计面积仅相差482.00×103 hm2,精度达到90.88%;EVI时间序列谱反映的作物生长过程的生理意义明确,采用MODIS EVI时间序列谱的遥感分类方法可以较好地实现冬小麦的遥感分类提取,满足冬小麦的长势监测和遥感估产的需要。

Using the MODIS EVI series in integrated with the growth status of winter wheat,the growing area of winter wheat in Henan province was extracted.The results showed that in EVI′ s feature space,the winter wheat had its unique spectrum series trait.After green-up,the EVI of the winter wheat had an overall gradual increasing trend and then followed a decreasing trend after flowering.The decreasing rate became higher after grouting.The decision tree classifier（CART） was used to extract the winter wheat growing area.There was a minor 482 000 hm2 of difference between extracted number and the number officially publicized.The accuracy of extracted winter wheat growing area reached 90.88%.The EVI time series spectrum can clearly reflect the physical meanings of crop growth.Using the remote sensing classification method of the MODIS EVI time series spectrum can accurately extract the winter wheat growing area,and meet the needs of monitoring winter wheat growth and yield estimation by remote sensing.