基于正交多小波的红外和可见光图像融合

The Fusion of Infrared and Visible Images with Orthogonal Multiwavelet

在应用中,预滤波器或平衡滤波器的设计影响着多小波的处理结果。针对OPTFR-多小波,设计了一种平衡滤波器,将其应用于红外和可见光图像融合。首先,对红外图像和可见光图像分别进行三层OPTFR-多小波变换,提取出两幅图像的低频分量和不同尺度的高频分量。由于图像的大部分能量集中在低频分量上,而不同尺度上的高频分量反映了图像在不同尺度上的细节信息,在对源图像融合时,对低频分量和高频分量分别采用能量和方差的加权平均进行融合,得到融合后的低频分量和不同尺度的高频分量,最后对这些融合后的分量进行重构,得到融合图像。仿真结果表明:和以往的融合算法相比较,该算法强化了源图像的细节信息,提高了融合的效果。

The design of pre - filter or balanced filter of multi - wavelets has an influence on processing results. The balanced multi - filter banks are designed for OPTFR - multi - wavelet and a new fusion method is presented. This method firstly extracts low frequency information and high frequency information at different scales from original infrared image and visible image using three levels OPTFR - multi - wavelet transform. Because most energy of images is concentrated on low frequency components, and high frequency components at different scales reflect detail information of images at different scales, so the fused low frequency and high frequency components can be obtained by weighted mean using energy and variance as weight respectively. Then the fusion image is reconstructed by opposite reconstruction process. The simulation shows that the approach retains the detail information of input images better than conventional fusion methods.