基于改进交叉双边滤波的红外与可见光图像融合算法

Infrared and Visible Image Fusion based on the Improved Version of Cross Bilateral Filtering

针对传统交叉双边滤波(CBF)算法在红外与可见光图像的融合过程中存在的噪声以及细节信息缺失等问题,提出一种基于非下采样剪切波变换(NSST)和特征提取来改进交叉双边滤波的图像融合算法.首先,应用NSST对两幅源图像进行处理,计算源图像与处理后图像的差值图像,并将差值图像叠加到源图像得到两幅预处理图像.其次,对两幅预处理图像进行交叉双边滤波,并与其计算差值融合得到细节图,对细节图和预处理图进行加权平均融合,得到初步融合图像,应用直方图均衡化和中值滤波对两幅源图像进行特征提取得到两个特征图,并采取绝对值取大融合特征图.最后,对融合后的特征图与初步融合图像取平均得到最终的融合图像.实验结果表明该算法在视觉效果和评价指标上均明显优于传统的融合算法.

Aiming at the problems of obvious noise and loss of details in the fusion of infrared and visible images by the traditional cross bilateral filtering algorithm,an image fusion algorithm based on the improved cross bilateral filtering.Firstly,two source images are processed by non-subsampled shearlet transform,the difference image between the source image and the processed image was calculated,and the difference image was added back to the source image to obtain two preprocessed images.Secondly,the two preprocessed images are crossed bilateral filtering,and the calculation difference between them was calculated into the detail image.The weighted average fusion of detail image and preprocessing image was carried out to obtain the preliminary fusion image.Then,histogram equalization and median filtering are used to extract the two source images to get the two feature images,and the feature images are fused according to the maximum of absolute values.Finally,the final fusion image was obtained by means of the fused feature image and the initial fusion image.Experimental results show that the proposed algorithm was obviously superior to the traditional fusion algorithm in terms of visual effects and objective evaluation indexes.