Data Fusion of Satellite Remotely Sensed Images and its Application in Agriculture

摘要

As the development of satellite remote sensing technology, it is possible to obtain various remotely sensed images from sensors with different spatial and spectral characters. Data fusion is a widely used technique to make full use of the different kinds of information so as to reach a more accurate and stable result. This paper investigates the extraction of Normalized Difference Vegetation Index (NDVI), an important parameter indicating the growth of crops in agriculture, from a SPOT panchromatic image and a TM multispectral image using 5 classical data fusion methods, they are Principal Component Spectral Sharpening (PCSS), Brovey Fusion, Gram-Schmidt Spectral Sharpening (GS), CN Spectral Sharping (CN) and wavelet fusion. Results show that the fused image by any of the methods contains more information content for NDVI extraction than before. Comparatively, GS has better effects in remaining both spectral information and brightness than other four methods.