基于双边滤波和NSST的红外与可见光图像融合

Infrared and Visible Image Fusion Based on Bilateral Filters and NSST

摘要:

针对传统图像融合容易导致目标信息减弱、背景细节不清晰的问题，提出一种基于非下采样剪切波变换（Non-subsampled Shearlet Transform, NSST）和双边滤波的融合算法。首先，利用双边和高斯滤波器处理红外与可见光图像，得到包含红外目标的大尺度边缘图像；然后，采用NSST分解红外与可见光图像，得到相应的高频和低频子带系数，低频部分利用已得的大尺度边缘图像指导加权，高频部分采用绝对值取大的方法；最后将融合后的各频带系数经过NSST逆变换得到融合结果。实验结果显示，该方法既能有效突出红外目标，又充分保留了可见光图像中的背景信息，在信息熵、互信息和峰值信噪比等客观评价指标上也都优于传统的融合算法。

Abstract:

In traditional infrared and visible image fusion results, target is tend to be weakened and details of background are blurred. To solve these problems, a new fusion algorithm based on bilateral filters and non-subsampled shearlet transform (NSST) is proposed. Firstly, bilateral and Gaussian filters are applied in source images to obtain the image which includes large-scale edges and infrared target information. Then, the source images are decomposed into low and high frequency sub-bands by NSST. The image with large-scale edges is used to guide the fusion of the low frequency sub-band, and the rule of maximum absolute value selection is used for the fusion of high frequency sub-bands. Finally,Sperform the inverse transform in order to obtain the final fusion result. Experiments demonstrate that the proposed fusion method can obviously highlight the target area, preserve details of the background area, and the quality evaluation indexes including entropy, mutual information, and peak signal to noise ratio is increased compared with conventional algorithnm.