基于改进非下采样轮廓波的图像融合算法

Image Fusion Algorithm Based on Improved Nonsubsampled Contourlet

针对红外光和可见光传感器难以同时获取清晰的目标和场景问题,提出一种基于非下采样Contourlet系数压缩感知的可见光和红外图像融合方法。首先采用非下采样Contourlet变换对可见光图像和红外图像分别进行多尺度、多方向分解,得到各自的非下采样Contourlet系数;然后对两种图像的低频系数采用加权与平均相结合的系数融合方案直接融合,对两种图像的带通方向子带系数采用伪随机傅里叶矩阵进行观测后将观测值进行加权融合,再对融合后的带通方向子带观测值进行系数重构,最后通过非下采样Contourlet逆变换重构图像。数值实验显示,该融合算法计算复杂度低,融合效果好,能同时得到目标和场景均清晰的融合图像。

In order to solve the images simultaneously, a problem that the infrared light sensor and the visible light sensor can＇t get clear nonsubsampled Contourlet coefficient compressive sensing based on infrared and visible image fusion method was proposed. Firstly, the nonsubsampled Contourlet transformation method was used to do multiscale multi directional image decomposition for the infrared and visible image. Each nonsubsampled Contourlet coefficient of them was got. Then, the low-frequency coefficients of the infrared and visible image were fused by the weighted average fusion method. The band-pass sub-band coefficients by the pseudo-random Fourier matrix observations weights fusion method were used. After that, fused band-pass sub-band coefficients were reconstructed. Finally, image by the inverse nonsubsampled Contourlet transformation were reconstructed. The experiment results showed that this fusion algorithm can get image with clear object and background, and it has the low computational complexity and good fusion effect.