Multispectral and Panchromatic Image Fusion Based on Genetic Algorithm and Data Assimilation

Most of fusion algorithms for multispectral and panchromatic image such as: principal component analysis, contrast pyramid decomposition, IHS method, Brovey method, PCA method, wavelet transformation, Gaussian-Laplace pyramid, and so on, their fusion rules could not be adjusted adaptively according to the purpose of the fusion image. In order to solve this problem, data assimilation conception in meteorological field is introduced.lt means that observation data and numerical simulation data are integrated to obtain more nature objective analysis results. The framework of fusion based on data assimilation and genetic algorithm for multispectral and panchromatic image was presented. In the framework, Weights of indices of the various attributes were determined according to their important degree in the following processing; the objective function was composed of weights sum of various evaluation indices of image, then the objective function was optimized using genetic algorithm to obtain suitable images. Finally, the experiment about the fusions of panchromatic and multispectral image (Spot5, QuickBird), prove the validity of this framework.