本文提出了一種基於二維經驗模態分解 (EMD) 和遺傳演算法的多聚焦 圖像融合方法。源圖像通過二維經驗模態分解得到本征模函數 (IMF) 分量， 根據T 檢驗對得到的本征模函數分量進行高低頻選擇。低頻係數融合使用改進 的視窗資訊熵最高準則進行，高頻係數採用視窗相關性計算準則，視窗相關性 係數融合根據閾值大小進行，閾值通過遺傳演算法確定。融合結果利用對融合 係數採用二維經驗模態分解逆變換得到。模擬結果表明該演算法顯著優於基於 圖元和基於視窗及基於小波的圖像融合方法。

A multi-focus image fusion method based on two-dimensional empirical mode decomposition and genetic algorithm is presented in this paper. First, a two-dimensional empirical mode decomposition is applied to the decomposition of source images. High and low frequency of intrinsic mode function component are classified by a T-test. Then low frequency coefficients are fused by improved maximum regional information entropy criterion whereas the high frequency coefficients are amalgamated in different threshold ranges of coefficients by regional correlation. The regional correlation threshold is selected by search of genetic algorithm. Finally, combined results are obtained by inverse two-dimensional empirical mode decomposition transform on fusion coefficients. Simulation results show that the proposed algorithm significantly outperforms traditional image fusion methods that are based on the pixel, region, and wavelet, respectively.