摘要
针对传统的红外与可见光图像融合算法所存在的边缘信息缺失等问题,提出了一种基于非下采样轮廓波变换和对比拉伸度的图像融合方法.首先,采用非下采样轮廓波对红外与可见光图像进行分解,得到高低频子带系数,采用“绝对值取大”和“窗口系数绝对值取大”的融合规则融合高频子带,低频子带采用改进的局部拉普拉斯能量的融合规则进行融合,经过NSCT逆变换得到初步融合图像.接下来,使用SUSAN算子对源图像进行边缘检测并进行对比度拉伸,应用上述NSCT算法融合拉伸后的边缘信息得到边缘融合图像.最后,重复对比度拉伸和NSCT算法对边缘融合图像和初步融合图像进行整合,得到最终的融合图像.实验结果表明,本文所提出的算法无论是在主观的视觉感知上还是在客观评价指标上均优于现有经典的融合算法,本文算法所得到的融合图像边缘信息更加丰富.
Aiming at the problem of poor edge information in existing infrared and visible image fusion algorithms,a fusion algorithm based on non-subsampled contourlet transform(NSCT)and image contrast stretching is proposed.First,the source images are decomposed into high and low frequency sub-band coefficients by the NSCT.The fusion rule of"absolute value maxminzation"and"absolute value of the window coefficient maxminzation"are adopted to fuse the high frequency coefficients.As for the low frequency coefficients,we adopt a fusion rule of the improved Laplace energy.Perform the inverse NSCT to get the initial fused image.Next,the Susan operator is used to detect the edge of the source image and image contrast stretching is carried out.We use the NSCT algorithm again to fuse the edge information to obtain the fused edge image.Finally,the NSCT algorithm is repeated to fuse the edge image and the initial image to obtain the final fusion image.Experimental results show that the proposed algorithm is superior to the existing classical fusion algorithm in both subjective visual perception and objective evaluation index,and the edge information obtained by the proposed algorithm is more plentiful.
作者
姜寒雪
郭立强
JIANG Han-xue;GUO Li-qiang(School of Computer Science and Technology,Huaiyin Normal University,Jiangsu Huaian 223300,China)
出处
《淮阴师范学院学报（自然科学版）》
CAS
2022年第1期17-23,共7页
Journal of Huaiyin Teachers College;Natural Science Edition
基金
国家自然科学基金资助项目(61203242)
江苏省大学生实践创新训练计划项目(202110323052Z)。
