作　　者： (沈瑜）; (党建武）; (王阳萍）;

机构地区： 兰州交通大学电子与信息工程学院,甘肃兰州730070

出　　处： 《兰州交通大学学报》 2017年第4期114-120,共7页

摘　　要： 对于含噪声红外与可见光图像的融合,在滤除噪声的同时通常也会损失同为高频信息的边缘细节.为了解决这个问题,提出了一种基于Bilateral双边滤波器和非下采样Contourlet变换(nonsubsampled contourlet transform,NSCT)的含噪声图像融合算法.首先采用NSCT方法对图像进行多尺度分解,调整参数使得到的红外图像和可见光图像的高频图像中包含了大部分的边缘信息和噪声,为了保留边缘信息和滤除噪声,采用Bilateral双边滤波器分别对两幅高频图像进行处理,并得到红外高频图像的细节图像.为了将原始图像的边缘细节信息尽可能多地提取到融合图像中,同时增强边缘信息,将红外细节图像与可见光高频图像叠加到一起,提升融合图像的视觉效果.然后,采用局部区域标准方差系数方法作为红外与可见光图像低频图像的融合规则进行融合.最后,将融合后的高频图像和低频图像根据NSCT反向变换进行重构,得到融合后的图像.实验结果表明,该算法在滤除噪声的同时能够很好地保持边缘细节,对于含噪声图像的融合具有很好的效果. The fusion of noisy infrared image and visible light image loses edge details of high-fre- quency information while filtering noises. For this reason, a noisy image fusion algorithm is pro- posed based on Nonsubsampled Contourlet Transform （NSCT） and Bilateral filter. First of all, NSCT method is used to decompose the infrared image and visible light image to low-frequency component and high-frequency component. The high-frequency noises and edge contour are in the high-frequency domain, which are processed by Bilateral filter. Then the noises are filtered and the detailed image of high-frequency component of the infrared image is obtained. This detailed image is added to the high-frequency component of the visible light image to extract the edge information as much as possible from the original infrared and visible light image, and the edge information is enhanced. After that, the low-frequency coefficients fusion rule is adopted to fuse the infrared im- age and visible light image by the standard variance coefficient method in local area. Finally, the inverse NSCT is used to reconstruct the high-frequency and low-frequency coefficients so as toobtain the fused image. The experiment results show that the fused image saves most details of the original image while enhancing the edges and filtering the noise.

关 键 词： 图像融合 非下采样 变换 双边滤波器