Research On Genetic Algorithm Based Remote Sensing Images Fusion Method

Genetic algorithm based remote sensing images fusion method is proposed in this paper.Based on the Genetic algorithm and MRA theory, the transformation and fusion rule of remote sensing images is discussed at first. Then utilizing the advantages of Contourlet in analyzing images and Genetic algorithm's good global search capacity, Contourlet and Genetic algorithm based remote sensing images fusion method also called weighted efficients optimized and calculated by genetic algorithm are proposed. The panchromatic image and the multispectral image are decomposed in high-pass subband and low-pass subband, after that, a series of weighted efficients were optimized and calculated by genetic algorithm. Then syncretize them by using the weighted efficients in different subbands. Finally, reconstruct the fused image by taking inverse transform of Contourlet.A remote sensing images fusion algorithm called fusion based on region structrue similarity and Genetic Algorithm is also proposed. In this method, Genetic Algorithm is used to optimize the weighting function of high frequency components of images after multi-scale decomposition. Then the subbands are syncretized by the weighting function and the the result image is reconstructed.At last, performance evaluation method of image fusion are researched, and the fusion results of these two algorithms are compared with the conventional approaches based on Contourlet domain and Genetic Algorithm tool. The experimental results show that the performance of the novel algorithms are better than the traditional contourlet fusion algorithms.