Remote Sensing Fusion of Multi-Resolution Images for Estimating Maize Planting Area

摘要

The inconsistency on spatial resolution and estimating accuracy of remote sensing images are the primary limiting factor for the estimation of the large-scale planting area for maize. The main purpose of this paper is to develop a method of integrating mid- and low-resolution images for the estimation of large-scale crop planting areas. A mid-scale time-series normalized difference vegetation index (NDVI) dataset, which was derived from the fusion of the moderate-resolution imaging spectroradiometer (MODIS) and TM images based on the wavelet transform, was established. The maize planting area was estimated by using the minimum distance model and the accuracy was evaluated by contrasting the method of supervised classification with in-situ samples. The results show that the estimation of the maize planting area based on the time-series NDVI information of the fusing images reached high levels of gross and position accuracy, which were better than that of supervised classification. It indicated that this method could fully utilize the time-series information from the MODIS images and the spatial resolution of a TM image. The difference in phenophases among fall crops enables the effective classification of the spatial distribution of these crops. The fusion of time-series MODIS and TM images fusing information analysis will provide an available method for estimating maize planting area of large-scale.