Fusion of Infrared and Visible Image Based on Genetic Algorithm and Data Assimilation

The available fusion methods for infrared and visible image (such as: principal component analysis, contrast pyramid decomposition, wavelet transform, and so on), their fusion rules could not be adjusted according to the purpose of the following use of images. Reference of the data assimilation ideas in meteorology, it 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 infrared and visible image was presented. 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 to be optimized by using genetic algorithm to obtain suitable images. Some experiments validate this framework of high effectiveness and accuracy.