Estimation of maize planting area based on wavelet fusion of multi-resolution images

The integration of mid-coarse-resolution remote sensing images provides abundant information, and therefore tends to be a popular way in large scale crop planting area estimation. This research utilized conventional MODIS and TM records to present an instance in large area maize planting area estimation. The wavelet fusion was adopted for obtaining normalized difference vegetation index (NDVI) with a spatial resolution of 30 m from both MODIS and TM images. And the standard growing curves of main fall crops were then constructed with the NDVI time series, which indicating crops difference in phenology. Minimum distance classification was carried out with the NDVI time series for mapping maze sown area in a typical maize-planting county, Yuanyan, Henan province. The result was validated with the in-situ parcels, which showing a better gross and position accuracies (89% and 90%) than those with either MODIS or TM records. The research can provide an efficient way with abundance information from both mid and coarse resolution records, and thus improve the applicability of remote sensing in large area crop area estimation.