基于NSST变换的微光偏振图像融合算法  

Low-light polarization image fusion algorithm based on NSST transform

作　　者：沈薛晨 刘钧[1] 高明[1] SHEN Xuechen;LIU Jun;Gao Ming(Institute of Optoelectronic Engineering,Xi^an Technological University,Xl/an 710021,China)

机构地区：[1]西安工业大学光电工程学院

出　　处：《激光杂志》2020年第2期103-113,共11页Laser Journal

基　　金：微光夜视技术重点实验室基金(No.61424120503xx)

摘　　要：为进一步提高对微光偏振图像中目标的识别效率,提出了一种基于非下采样剪切波变换(NSST)的偏振图像融合算法。首先,通过融合得到偏振特征图像,将所含偏振信息量较多的偏振度图像和偏振角融合,通过基于区域方差与炳值加权相结合的融合算法得到偏振特征图像。最后,将偏振特征图像与亮度较强的强度图像融合,低频系数的融合采用PCNN与区域能量相结合的规则,高频系数的融合采用区域特性能量的规则。实验结果表明,在主观视觉上,图像观察舒适性较好;并且,通过选取方法的对比,融合后的图像在客观评价指标上,皆优于选取的方法。In order to further improve the recognition efficiency of targets in low-light polarization images,this paper proposes a polarization image fusion algorithm based on non-down-sampling shear wave transform(NSST).Firstly,the focus is to obtain polarization characteristic image by fusion.The polarization degree image and the polarization angle with more polarization information are fused,and the polarization characteristic image is obtained by fusion algorithm based on the combination of regional variance and weighting.Finally,the polarization characteristic image is fused with the intensity image with strong brightness.The fusion of low-frequency coefficient is based on the rule of the combination of PCNN and regional energy.And the fusion of high-frequency coefficient is based on the rule of regional characteristic energy.The experimental results show that the image observation is comfortable in subjective vision.In addition,by comparing the selection methods,the fused images are superior to the selection methods in terms of objective evaluation indexes.

关 键 词：微光偏振图像 图像融合 NSST变换 融合规则 

分 类 号：TN223[电子电信—物理电子学]