压缩感知图像融合

目前图像融合的方法大多数都是基于小波变换的图像融合方法，通过对小渡变换之后的低频系数和高频系数分别采用不同的融合准则，来达到所需要的图像以进行下一步处理，这些方法需要知道原始图像，也就是对硬件要求较高。采用压缩感知图像融合，即，将压缩感知用于图像融合，使得只知道原始图像在某个变换下的投影值的情况下，通过对已知的投影值使用融合规则得到融合后的投影值，然后用重构算法重构出图像，大大降低了对硬件的要求。在此给出了压缩感知融合方法与基于小波变换的图像融合方法的实验结果，融合结果表明，在不降低融合效果和视觉效果的基础上，该方法能够极大地降低硬件成本。采用熵作为衡量融合效果的指标，并对用两种方法融合的结果图像做了对比，研究结果表明，CS融合方法要优于基于小渡变换的图像融合方法。

Image fusion based on compressive sensing

The compressive sensing(CS) technology is adopted for image fusion in this paper.In this method,the fusion rule is used to the known projection value to obtain the fused projection value and then reconstruct the images by reconfiguration algorithm in the case of only knowing the projection value of original image which is in a certain conversion.It reduce the requirement to the hardwares greatly.The results of experiments for methodes of the image fusion based on C/S and the image fusion based on wavelet transform are given.The fusion results show that the method of the image fusion based on CS can reduce hardware cost enormously without cutting down the vision effect.In this paper,entropy is taken as the index to evaluate the fusion effect,and the images achieved by two different fusion methods are compared.The research demonstrates that CS fusion method is better than the one based wavelet transform.