基于蚁群算法的多传感器图像融合方法  

Multi-Sensor Image Fusion Method Based on Ant Colony Optimization Algorithm

作　　者：何同弟[1] 祝燎[1] 

机构地区：[1]河西学院物理与机电工程学院,甘肃张掖734000

出　　处：《测控技术》2017年第8期15-19,共5页Measurement & Control Technology

基　　金：河西学院校长基金(XZ2014-21);国家自然科学基金青年基金(41371338)

摘　　要：为了使融合后的多传感器图像获得更多的光谱信息、提高清晰度、降低数据冗余度,提出了一种基于蚁群算法的多传感器图像融合方法。对低分辨率图像上的蚁群以相位一致性作为启发信息,高分辨率图像中的蚁群以梯度强度作为启发信息,两个蚁群通过共享的信息素矩阵实现协作,根据信息素矩阵提取图像特征。算法采用区域能量的加权自适应融合规则确定低频系数,结合蚁群算法提取的边缘特征融合来指导高频系数融合。融合结果表明,该方法在不同分辨率上引入了多种启发信息,因而能够提取更加完整和有意义的图像特征,为多传感器图像融合提供了更智能、更细致、更全面的图像信息。A multi-sensor image fusion method based on the ant colony algorithms to obtain more spectral infor- mation, increase the resolution and reduce data redundancy of the fused multi-sensor image, is presented. With the usage of phase congruency as the inspiration information in the low-resolution image, and gradient strength as the inspiration information in the high-resolution image, the two ant colonies cooperate in the image by sha- ring the same pheromone matrix and extract the feature of the images according to the pheromone matrix threshold. The algorithm adopts the weighted adaptive fusion rules of regional energy to determine the low-fre- quency coefficients, and combines the edge feature fusion extracted by the ant colony optimization algorithm to guide the high frequency coeffcient fusion. Experiment results show that the method provides more intelligent, more detailed and more comprehensive image information for multi-sensor image fusion because it can extract more complete and meaningful image characteristics by introducing a variety of heuristic information in differ- ent resolution.

关 键 词：多传感器 蚁群算法 特征级融合 小波变换 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]