一种基于NSCT和像素相关性的图像融合算法

由于小波变换在处理二维图像融合中存在缺陷,提出一种更优越的基于NSCT变换的图像融合算法,分别介绍了低频系数和高频系数融合规则。低频系数融合用基于能量方差决策值最大法,可以有效提高图像融合质量;高频系数融合规则用区域对比度与区域能量相结合的方法,可以减少噪声干扰。得到低频和高频融合系数后,通过NSCT反变换重构图像,得到融合图像。最后选用4种常用的图像融合方法与此文算法进行对比仿真实验。通过主观观察和客观指标的比较,充分证明此文融合算法的优越性。

An image fusion algorithm based on NSCT and pixels correlation

Considering this disadvantage of wavelet transform and its application in image fusion, a novel algorithm of image fusion was preposed based on nonsubsampled contourlet transform ( nsct).low-frequency and high-frequency coefficients of fusion rules are introduced respectively.the low-frequency coefficients use an fusion rule based on local energy and variance.the rule can effectively improve the fused image. the high-frequency coefficients use an fusion rule based on the combinations of local energy and contrast.the rule can reduce noise disturbance.the fused the low-frequency and high-frequency coefficients are reconstructed to obtain fusion image through NSCT inverse transform. four kinds of common image fusion algorithms are compared with this algorithm by computer simulation.objective indicators and the subjective visual impression show that this algorithm is more effective and superior than the four common algorithms.