The Research On Satellite Images Fusion Algorithm Based On Multiscale Analysis

In this paper, infrared cloud image and visible cloud image collected by meteorological satellite are the primary data for images fusion. Aims are to improve the resolution of infrared cloud images by the multiscale analysis method, so as to lay a good foundation for the detection and early warning of strong convective clouds.This paper mainly focus on multi-wavelet transform and Contourlet transform and their application in cloud images fusion. Main work are as follows:This paper discusses the general theory of wavelet transform and multi-wavelet transform and provides the binary wavelet transform algorithm of image. This paper studies the images fusion method based on multi-wavelet transform and achieves the optical texture details information of visible image of high spatial resolution into the infrared cloud images of low spatial resolution in Visual C++ platform, in considering the physical meaning of cloud image data. Experiments shows that this method receives higher-resolution infrared cloud image and improves the physical information quantities and accuracy of infrared cloud images data.This paper emphatically focuses on discussing the principle of multiscale geometric analysis methods such as ridgelet transform, Curvelet transform and Contourlet transform. And also mainly studies the image fusion method based on Contourlet transform and tries to apply it to cloud image fusion research in the Matlab platform. Experimental results showed that the fusion method of Contourlet transform retains more infrared cloud image's information quantity than wavelet transform does and the merged infrared cloud image contains more texture information.