基于BEEMD分解的红外与可见光图像融合

针对二维经验模态分解（BEMD）算法在图像分解过程中存在模态混叠，提出了一种基于二维集合经验模态分解（Bi-dimensional Ensemble Empirical Mode Decomposition，BEEMD）算法的红外与可见光图像融合方法。为了抑制分解过程中存在的模态混叠现象，获得准确的特征分量和残差分量，使用BEEMD算法对图像进行分解。对获得的特征分量采用局部区域能量选择与加权的融合策略进行融合，而残差分量采用模糊逻辑进行融合。将融合后的特征分量和残差分量叠加得到最后的融合图像。实验结果表明，该方法能够很大程度上保留可见光图像的背景信息，同时突出红外图像的目标，具有较好的可视性，而且在平均梯度（AG）、标准差（SD）、信息熵（IE）等客观评价指标方面，也有明显的优势。

Infrared and Visible Image Fusion Based on BEEMD Decomposition

Aiming at the mode mixing existing in the image decomposition process of the Bi-dimensional Empirical Mode Decomposition（BEMD） algorithm, an infrared and visible images fusion method based on Bi-dimensional Ensemble Empirical Mode Decomposition（BEEMD） algorithm is proposed. First, in order to suppress the mode mixing phenomenon in the decomposition process and obtain accurate feature components and residue components, the BEEMD algorithm is used to decompose the image. Second, the extracted feature components are fused using a local area energy selection and weighted fusion strategy, while the residue components are fused using fuzzy logic. Finally, the fused feature components and residue components are directly superimposed to obtain the final fused image. The experimental results show that the method can largely retain the background information of the visible light image while highlighting the target of the infrared image, with better visibility. In addition, it has obvious advantages in objective evaluation indicators such as Average Gradient（AG）, Standard Deviation（SD）, and Information Entropy（IE）.