To Obtain the Planting Area of Corn Crop Based on MODIS Satellite Data

Abstract

The MODIS satellite data was used to extract corn planting area in Northeast China (including Heilongjiang, Jilin and Liaoning Provinces) in this paper. The support vector machine (SVM) method was adopted for the classification to extract the distribution of corn in the sampling area through the HJ satellite imageries, the MODIS-EVI data stacked by different time series was masked by the classification results. Then the EVI time series curves of the corn in the sampling area were used to analyze the curve of corn in the sampling area and get the standard curve of corn in Northeast China. The absolute mean distance between MODIS-EVI time series and the standard curve deriving from sampling area was calculated, and the maximum possible corn distribution was obtained considering different thresholds with different regions. Finally, we compared the distribution of corn in Northeast China according to MODIS data results and actual corn distribution in Heilongjiang Province. The result showed that the obtaining precision of corn planting area is about 74%. It is concluded that the MODIS image can be utilized for the extraction of the crop planting area.