Data Fusion Techniques For Multi-sources Remote Sensing Image And Experiments

In recent years, as the availability of remote sensing image of varying resolution has increased. Merging images of different spatial resolution has become a significant operation in the filed of digital image processing. What the multi-source remotely-sensed image fusion technique study is how to use the advantages of image originating from different kinds of sensors synthetically, and achieve a new image with more precise, more abundant information through a certain algorithm. With the development of information technologies, currently, remote-sensing image fusion becomes indispensable in image processing and understanding, as well as many important military and civil applications.Firstly, this paper presented the purpose of image fusion, the conception of remote sensing image fusion and the technical development situation at present at home and abroad. Aiming at the problem of original image that will affect the quality of the fusing image, this paper gives the image pretreated processes such as: geometry emendation, remove noise and image registration, and ortho-rectification the original image in the mountain area, avoid the influence of the topography. Secondly several classical methods of image fusion on the pixel level: Intensity-Hue-Saturation transform, multiplicative, Brovey transform, principle component analysis transform and wavelet transform have been summarized. Finally, three different groups of images was employed as demonstration.the subjective and objective evaluation criteria have been used to evaluate the performance of these three methods. The result shows that the best fusion method to improve the spatial resolution is IHS transform fusion; the best fusion method to keep the integrality of the spectrum characteristic is wavelet transform fusion; the fusion image quality depends on the fusion method, the images correlation coefficient, such as acquisition, spectral response scope, whether identical platform or not, the spatial resolution and type of image. ETM-8and ETM multi-spectral image which acquired at the same time and from Landsat-7 platform provide the best fusion result.