Theories And Methods For Remote Sensing Image Fusion Based On Multi-resolution Analysis

Remote sensing image fusion is one important branch of image fusion, which studies how to fuse the multi-spectral images and the panchromatic image by using some certain fusion algorithm, making the fused images with a high spectral resolution as well as a high spatial resolution. The purpose of image fusion is simply taking advantage of complementary information among images, eliminating redundant information among images, highlighting the useful information and ultimately obtaining the image with more valuable information. Image fusion based on the multi-resolution analysis is a widely used and effective fusion algorithm.Remote sensing image fusion based on the multi-resolution analysis is to fuse the remote sensing images with different resolutions and different decomposition level by effective rules respectively. The key point of image fusion based on the multi-resolution analysis is the selection of fusion rules. In this paper, the author researches deeply into the criterion under the multi-resolution framework focusing on the fusion of the panchromatic image and the multi-spectral images. On the basis of the existing fusion theory, the author has adopted the more effective fusion rules. The main content of this paper is summarized as follows:1. The research backgrounds, the fusion significance, the development and status are introduced, and then the basic process and the level of fusion are summarized.2. The basic framework of image fusion and the common fusion algorithm of the pixel level are summarized. Subsequently, the evaluation standards of fusion are discussed, including the subjective evaluation and the objective evaluation.3. The basic theory of the multi-resolution analysis is introduced, and then the decomposition and reconstruction of the Wavelet transform and the Curvelet transform is described in detail. Finally, the Contourlet transform theory is simply introduced.4. The general framework of remote sensing image fusion based on the multi-resolution analysis is summarized, and then the process of remote sensing image fusion based on the Wavelet transform and the Curvelet transform is respectively elaborated. Subsequently, the influence factors of image fusion based on the multi-resolution analysis are simply introduced, and then the fusion algorithms are analyzed and compared through several groups of remote sensing experimental data.5. For the rules, which is the key point of image fusion based on the multi-resolution analysis, the rules based on visual contrast and mutual information are adopted on the basis of the existing fusion theory. Through the analysis and comparison of several groups of remote sensing experimental data, the effectiveness of the rules in this paper is proved.6. The research achievements and the disadvantages of this paper are summarized, and then the development prospect is looked forward to.