基于可能性信息质量合成的双模态红外图像融合算法选取

针对双模态红外图像在融合时异类差异特征两两合成出现信息冗余导致所选择的融合算法相互冲突,造成融合效果差甚至失效的问题,提出了一种基于可能性信息质量合成的双模态红外图像融合算法选取方法。首先计算双模态红外图像多融合算法下不同差异特征的融合有效度,利用可能性框架得到对应的可能性分布向量子集;其次计算向量子集的信息量和可信度,并对多个向量子集进行加权合成;然后构建基于信息质量的排序函数,得到每种融合算法下的非支配子集;最后构建多融合算法得分函数的联合分布对多种融合算法优化选择。实验结果表明,将基于质量来整合多个差异特征的方法运用于双模态红外图像融合算法选取中,所选出的融合算法在加权综合指标上高于其他算法均值55%以上,证明了本文方法的有效性和合理性;由多组实验算得本文方法平均耗时10.083 s,在时间效率上也符合实时图像融合应用的工程需求。

Dual-mode Infrared Image Fusion Algorithm Selection Based on Possibility Information Quality Synthesis

Aiming at the problem of information redundancy in the pairwise synthesis of heterogeneous difference features during the fusion of dual-mode infrared images,which leads to conflicts between the selected fusion algorithms,resulting in poor fusion or even failure,a method of dual-mode infrared image fusion algorithm selection based on possibility information quality synthesis is proposed.Firstly,the fusion effectiveness of different difference features under the dual-mode infrared image multi-fusion algorithms is calculated,and the possibility framework is used to obtain the corresponding possibility distribution vector subsets.Secondly,the information quantity and credibility of the vector subsets is calculated,and multiple vector subsets are weighted and synthesized.Then a ranking function based on information quality is constructed to obtain the non-dominated subsets under each fusion algorithm.Finally,the joint distribution of multiple fusion algorithms score function is constructed to optimize the selection of multiple fusion algorithms.The experimental results show that the method of integrating multiple difference features based on quality is used in the selection of dual-mode infrared image fusion algorithm,the selected fusion algorithm is more than 55%higher than the average of other algorithms in the weighted comprehensive index,which proves the effectiveness and feasibility of the method.The average time of the proposed method is 10.083 s after several experiments,which also meets the engineering requirements of real-time image fusion application in terms of time efficiency.