利用NDVI指数时序特征监测秋收作物种植面积

Estimation of autumn harvest crop planting area based on NDVI sequential characteristics

摘要: 通过对湖北省江陵县各种地物MODIS数据NDVI时序特征的分析，选取夏秋作物轮作期（5月下旬）和NDVI均值为标准，采用分层方法区分秋收作物（中稻、晚稻、棉花）区与其它区，然后利用BP神经网络法对三种作物进行监督分类，得出了三种作物种植的空间分布。并将结果与ETM数据的神经网络监督分类值作为标准值进行了误差分析，得出其平均误差率为13.6%，能准确反映江陵县秋收作物分布状况，该分层模式加BP神经网络监督分类方法适用于多种作物种植区的作物分类。

Abstract: NDVI sequential characteristics of various ground objects was analyzed which was on MODIS data of Jiangling County, Hubei Province. By the standards of crop rotation period’s data and NDVI mean value, spatial distributions of three kinds of crops were obtained using BP neural network supervised classification based on the hierarchical method for crop regions and non-crop regions distinguishing. Compared with standard results calculated with ETM data, the average error ratio of the research is 13.6%. Results indicated that the method can accurately reflect various crop distributions in Jiangling county and be applied for crops classification.