Remote Sensing Image Fusion Algorithm Based On Contourlet Transform Research

At present, the methods of getting remote sensing data is more and more, and multiple source of information is the biggest characteristic of remote sensing data. How to use effectively these data and realize the complementary advantages of multi-source data, remote sensing image fusion technology makes these problems came into being, and become the research hotspot in image processing.Single remote image looks like a description of the objective reality of inaccurate and incomplete. Through remote sensing image fusion technology based on certain standards, reasonable disposal and complementary resources observation information, superior fusion image can be obtained with many images than any of the source image quality. It can more accurately reflect the objective reality.This article first discusses the research background and development prospect of remote sensing image fusion technology, then describes remote sensing image fusion processing technology’s basic concept, fusion hierarchy classification, domestic and foreign research dynamic and specific application. The remote sensing image preprocessing and registration algorithms are studied systematically, using the basic principle of remote sensing data fusion algorithm and fusion performance evaluation index of the image, after some experiments proved that the proposed new method has more superior performance, and has an important significance on the later research work.land, water, and earth’s atmosphere are very complex, and the current sensors are limited in time, space, spectral resolution and radiation, so that these sensors don’t record complete data. Therefore, the error will not only reduce the quality of the remote sensing data, but also directly affect the validity of the fusion results. The most common errors of remote sensing data are geometric errors and radiation errors. basing on the formation principles of these errors, this paper make corrective algorithm respectively, and put forward the combination of mutual information image registration based on Contourlet transform method.The focus of this article is to introduce detailed several kinds of remote sensing image fusion methods that commonly used. After analyzing their respective advantages and disadvantages of these algorithms and applicable scope, the Contourlet transform is mainly studied. Regarding the traditional Contourlet transform as the fundamental starting point, combined with other image processing algorithms, the paper proposed remote sensing image fusion algorithm based on Contourlet transform binary contrast pyramid. This method integrated use the advantages of the existing algorithms about multi-scale analysis. Through analyzing and comparing the experiment of different fusion methods for the fusion results, for different imaging characteristics of sensors and fusion application purposes, this paper make relatively suitable fusion rules and fusion operators, and at last discussing remote sensing image fusion system.The comprehensive evaluation on the fused images is also indispensable, this paper summarizes, sorted out and classified the existing objective evaluation index of remote sensing image fusion, introducing a kind of high real-time performance evaluation method based on singular value decomposition.Finally, the paper summarizes the research achievements and the deficiencies, and discussed the future development direction.