结合NSST和LSWT的光学/被动毫米波图像融合  

Fusion of visual/PMMW image based on NSST and LSWT

作　　者：柳菁[1] 聂建英[1] 

机构地区：[1]福州大学数学与计算机科学学院,福建福州350108

出　　处：《电视技术》2016年第12期12-16,共5页Video Engineering

基　　金：江苏省资源环境信息工程重点实验室基金项目(JS201104)

摘　　要：提出了一种结合NSST和LSWT的图像融合算法,并应用到安检中。该方法首先采用NSST对被动毫米波图和可见光图进行分解得到一组高低频系数;接着对分解后得到的低频系数进行LSWT变换,得到新的一组高低频系数,分别对其采用基于区域方差融合规则和CS算法进行压缩、融合、重构后进行LSWT逆变换;再对第一步的高频系数采用改进的拉普拉斯算子进行处理;最后NSST逆变换处理得到融合图像。实验仿真结果表明,所提出的方法融合效果较好,实现了对隐匿物品的快速识别。A method to deal with visual image and PMMW image for the security inspection is put forward, basing on Non-subsampled Shearlet Transform （NSST）and Lifting Stationary Wavelet Transform （LSWT）. Firstly, two images＇ low and high frequency components decomposed with NSST can be got. Next,the low-frequency components are decomposed by the LSWT to get the other componets which are fused based on regional variance and compressed, fused and reconstructed by CS. Then,do the lifiting stationary wavelet inverse transform. Thirdly, the first high-frequency components are dealed with Laplacian. Finally, the inverse NSST is used to get the final fused image. The experiment result indicates that the proposed method has a better result to realize speedy identification of the concealed.

关 键 词：光学图像 被动毫米波图像 NSST LSWT 图像融合 

分 类 号：TN911.73[电子电信—通信与信息系统]