Crop rotation change detection in Huang-Huai-Hai plain, China

The situation of cropland use in China is very complicated. In many areas, the cropland is used in multi-cropped ways. Monitoring crop rotation changes is necessary for agricultural management and decision-making. The objective of this research is to detect the cropland areas and distributions, where the crop rotation changed from double-cropping to single-cropping, using high temporal resolution imagery. The study site is located in the Huang-Huai-Hai plain, the largest plain and main cultivated area in China. The data used in this study are 10-day composite SPOT-4 VEGETATION (VGT) Normalized Difference Vegetation Index (NDVI) over the period of 1999 and 2003. After the data preprocessing for removing the influence of cloud and other noise, the discrete Fourier transform was applied to the NDVI data set on a per pixel basis for the whole study area. The mean and the amplitudes and phases in the first three harmonics of two temporal profiles were used to identify the areas where changes occurred using decision tree classifier. The result of this study showed that more than 4 million hectare croplands were changed from double-cropping in 1999 to single-cropping in 2003 and the main changes occurred in the southeast of Hebei province and the north of Shandong province. © 2005 IEEE.