计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (11): 171-174.

• 图形、图像、模式识别 • 上一篇    下一篇

基于NSCT与模糊逻辑的图像融合方法

郑义军1,2,任仙怡2,3,刘秀坚1,2,胡 涛2,3,张基宏1,2,3   

  1. 1.深圳大学 计算机科学与技术系,广东 深圳 518060
    2.深圳市可视媒体处理与传输重点实验室,广东 深圳 518029
    3.深圳信息职业技术学院,广东 深圳 518029
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-04-11 发布日期:2011-04-11

Image fusion based on NSCT and fuzzy logic

ZHENG Yijun1,2,REN Xianyi2,3,LIU Xiujian1,2,HU Tao2,3,ZHANG Jihong1,2,3   

  1. 1.Department of Computer Science and Technology,Shenzhen University,Shenzhen,Guangdong 518060,China
    2.Shenzhen Processing and Transmission of Visual Media Key Laboratory,Shenzhen,Guangdong 518029,China
    3.Institute of Information Technology,Shenzhen,Guangdong 518029,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-04-11 Published:2011-04-11

摘要: 提出一种基于非下采样Contourlet变换(NSCT)和模糊逻辑的图像融合方法。NSCT分解有利于更好地保持图像的边缘信息和轮廓结构,并增强图像的平移不变性。在对原图像进行多尺度几何分解后,针对图像融合过程中原图像融入信息程度的不确定性,采用基于模糊逻辑(Fuzzy logic)的融合规则指导图像的融合:根据高低频各自的频带特性和待融合图像的特点,对高频使用基于信息熵与模糊逻辑的融合方法,而对低频使用基于亮度、梯度、方差、信息熵等特性及模糊逻辑的规则得到融合系数,最终经过NSCT逆变换得到融合图像。实验结果表明,该算法的融合图像具有良好的视觉效果及客观评指标,指标与视觉效果也能够较好的统一。

关键词: 图像融合, 红外与可见光, 非下采样Contourlet变换, 模糊逻辑, 隶属度函数

Abstract: A method of image fusion based on Non-Subsampled Contourlet Transform(NSCT) and fuzzy logic is proposed in this paper.As NSCT which is fully shift-invariant can maintain the edges information and the contour of original image.The original images with NSCT are decomposed.And then,a fusion rule based on fuzzy logic is adopted to handle the uncertainty of the contribution of the source images to the fused image.According to the different features of high and low frequency-bands,the coefficients of high frequency-bands are fused by entropy and fuzzy logic,and the coefficients of low frequency-band are fused by intensity,gradient,variance,entropy and fuzzy logic.Finally,the inverse NSCT is performed to obtain the fused image.The experimental results show favorable visual effect and objective measures of the fused image,and the visual effect and the objective measures indexes have good consistency.

Key words: image fusion, infrared and visible light image, Non-Subsampled Contourlet Transform(NSCT), fuzzy logic, member-ship function