摘要
针对红外和可见光图像的自身特点,本文提出一种基于四元数小波变换(QWT)和自适应脉冲耦合神经网络(PCNN)模型相结合的红外图像与可见光图像融合的新算法。首先将红外图像与可见光图像分别进行四元数小波变换,分别得到低频子带和高频子带系数;其次,采用局部区域方差匹配的融合准则处理低频子带系数,并用自适应的PCNN模型处理高频子带系数,用一种改进的空间频率作为PCNN模型的刺激输入,且采用拉普拉斯算子调节PCNN模型的阈值;最后经过四元数小波逆变换实现图像的融合。将本文提出的新算法与经典的图像融合算法进行对比分析,实验结果说明,新方法取得了较好地视觉改进效果,并在客观标准上也达到一定的提高。
Aiming at the features of infrared and visible image,a new fusion algorithm which combines quaternion wavelet transform(QWT)with adaptive pulse coupled neural network(PCNN)is presented.In the proposed fusion process,the infrared image and visible image are decomposed into low-frequency sub-band and high-frequency sub-band coefficients respectively via the QWT at first step.Then the low-frequency sub-band coefficients are fused using local variance matching rule,the high-frequency sub-band coefficients are fused using adaptive PCNN model.An improved spatial frequency as the input of the PCNN is used,and the Laplace operator is used to adjust the threshold of PCNN model.Finally,the fused image is reconstructed based on inverse QWT.The experiment results show that compared to the traditional image fusion algorithms,this proposed algorithm achieves better subjective visual results and also improves the objective criteria.
作者
朱芳
刘卫
ZHU Fang;LIU Wei(College of General Education,Anhui xinhua University,Hefei 230088,China;Institute of Intelligent Machine,Chinese Academy of Science,Hefei 230031,China)
出处
《红外技术》
CSCD
北大核心
2018年第7期660-667,共8页
Infrared Technology
基金
安徽省质量工程<智慧课堂试点项目>(2017zhkt247)
安徽省高等学校自然科学研究重点项目(KJ2016A310)
安徽新华学院<概率论与数理统计A>教改课程项目(2015jgkcx11)
