基于小波对比相关性及加权因子的图像融合算法  

Study on image fusion based on relativity of directional contrast and weight of wavelet coefficients

作　　者：陈焕平[1] 何明一[1] 

机构地区：[1]西北工业大学电子信息学院,西安710072

出　　处：《电子测量技术》2007年第10期33-37,共5页Electronic Measurement Technology

摘　　要：以小波方向对比度或图像局部特征作为判据的图像融合算法均有较好的效果。本文综合上述两个方面的因素,并定义了高频加权因子,提出了一种新的基于小波方向对比度相关性和高频加权因子的融合算法。该算法首先提取小波系数用以计算高频系数的方向对比度,而后计算相关度并构造加权系数,最后利用加权系数和高频加权因子重新得到融合图像的小波系数进行反变换,重构融合图像。经多聚焦图像的融合实验表明:该算法比本文提及的另外3种算法更为有效。Algorithms of image fusion based on directional contrasts of the high frequency wavelet coefficients or on area-based features are both have satisfactory effect. Based on the above two algorithm, a new algorithm of image fusion is proposed by adding a defined high frequency weighting factor. Firstly wavelet coefficients of the source images are used to calculate the contrasts of the highest frequency coefficients through wavelet transform. Then correlation is calculated and the weighting factor coefficient must be structured. Finally the fused image can be reconstructured by the inverse wavelet transferring with weighting coefficients and high frequency weighting factor. The experiments of the fusion of multi-focus images show that this algorithm is better than the other three algorithms.

关 键 词：图像融合 小波变换 对比度 局部相关性 加权因子 

分 类 号：TN911.73[电子电信—通信与信息系统]