一种基于四元数小波变换的图像融合方法*

摘要

目的：基于四元数小波变换理论提出了一种新的图像融合方法。方法基本思想主要包括3步：首先，对图像进行四元数小波分解，得到各个尺度下的高频子带和低频子带。其次，对高频子带和低频子带分别采用区域最大值法和系数平均的融合准则，得到融合的四元数小波融合系数。最后，利用四元数小波逆变换得到融合图像。结果用所提方法得到了6组测试图像在主观和客观上的融合结果。结论实验表明，所提方法比离散小波变换、对偶数复小波变换和曲线波变换的图像融合方法更有效。%Purposes— To propose a novel image fusion scheme based on quaternion wavelet trans‐form .Methods—Our proposed QWT‐based image fusion scheme performs the three stages :1) per‐form a quaternion wavelet multi‐scale decomposition of each source image and obtain their low‐pass and high‐pass sub‐bands ;2) obtain the fused quaternion wavelet coefficients in the low‐pass bands with average method while in the high‐pass bands by fusing the largest absolution values ;3) recon‐struct the fused image by taking inverse quaternion wavelet transform .Results—The fused results are obtained on six groups testing images by the proposed fusion schemes from the subjective and objec‐tive aspects .Conclusion—The experimental results show that the proposed method is more effective than the discrete wavelet transform ,the dual‐tree complex wavelet transform and the image fusion schemes of curvelet transform .