Research On Empirical Mode Decomposition Based Remote Sensing Images Fusion Method

Empirical Mode Decomposition(EMD) based remote sensing images fusionmethod is proposed in this paper.At the first of this paper, the background , current situation and significance ofimage fusion are brought into this text , and then, some traditional fusion methods atpixel level are introduced, later some assessment criteria about how to evaluate theresult of image fusion are being given, including some functions about objectivejudgment.Secondly, focus on the main points in this article——EMD, it can be used toanlysis the nonlinear and non-stationary signal, it also can decompose the signaladaptively into different frequency components. Bidimensional empirical modedecomposition is based on EMD, it can decompose images into different parts. It isvery useful in many image field. Multivariate empirical mode decomposition is a newEMD algorithms, it mainly focus on the problem of extreme value, pointing tomultivariate signals can be projected into uniform vectors, and then according the newsignals to extract the extream value, the experiment show the correction about thismethod.Finally, the above two EMD methods proceeds in image fusion at two spaces,first of all ,the images are decomposed by EMD , fusion is performed atdecomposition level and the fused IMFs are reconstructed to realize the fused image.In order to evaluate the result of image fusion ,this text uses the appropriate evaluatefunctions .The results of experiment show that these two methods can satisfy therequirement of remote sensing images fusion, that is at keeping image spectrum, it canenhance the spatial frequency of images.