电子学报 ›› 2012, Vol. 40 ›› Issue (10): 1984-1988.DOI: 10.3969/j.issn.0372-2112.2012.10.014

• 学术论文 • 上一篇    下一篇

基于相似性灰关联的Curvelet域多聚焦图像融合

马苗1,2, 万仁远1, 尹义龙2   

  1. 1. 陕西师范大学计算机科学学院, 陕西西安 710062;
    2. 山东大学计算机科学与技术学院, 山东济南 250101
  • 收稿日期:2012-03-21 修回日期:2012-06-06 出版日期:2012-10-25
    • 通讯作者:
    • 尹义龙男,1972年生于山东巨野,山东大学教授,博士生导师,主要研究方向为机器学习、图像处理等. E-mail:ylyin@sdu.edu.cn
    • 作者简介:
    • 马 苗 女,1977年生于山东聊城,2005年和2008年于西北工业大学分别获得工学博士学位和完成博士后研究工作,现为陕西师范大学副教授,硕士生导师.主要从事图像处理、灰色理论与信息隐藏等方面的研究. E-mail:mmthp@ snnu.edu.cn
    • 基金资助:
    • 国家自然科学基金 (No.61070097,No.10974130); 陕西省青年科技新星项目 (No.2011kjxx17); 陕西省自然科学基金 (No.2011JQ8009)

Multi-focus Image Fusion Based on Grey Relation of Similarity in Curvelet Domain

MA Miao1,2, WAN Ren-yuan1, YIN Yi-long2   

  1. 1. School of Computer Science, Shaanxi Normal University, Xi'an, Shaanxi 710062, China;
    2. School of Computer Science and Technology, Shandong University, Jinan, Shandong 250101, China
  • Received:2012-03-21 Revised:2012-06-06 Online:2012-10-25 Published:2012-10-25
    • Supported by:
    • National Natural Science Foundation of China (No.61070097, No.10974130); Shaanxi Provincial Young Science and Technology Star Scientific Research Project (No.2011kjxx17); Natural Science Foundation of Shaanxi Province,  China (No.2011JQ8009)

摘要: 针对多聚焦图像融合问题,提出一种基于相似性灰关联的Curvelet域可见光图像融合方法.该方法首先将待融合图像进行多级Curvelet分解,然后对各融合图像的高频系数进行分块,利用灰色理论中的灰色欧几里德关联度确定各子块间的相似性,并制定不同的高频系数融合策略,低频系数则采用算术平均法融合;最后,通过Curvelet逆变换重构融合图像.实验结果显示,该方法融合图像的信息熵、标准差和清晰度等指标优于金字塔融合法以及小波变换法等常见的多种融合方法.

关键词: 图像融合, Curvelet变换, 灰色关联分析, 相似性

Abstract: Focusing on multi-focus image fusion,this paper presents a Curvelet domain method based on grey relation of similarity for visible-light images.In this method,source images are respectively decomposed by multilevel discrete Curvelet transform first.Then,after high frequency coefficients were divided into small blocks,grey Euclid relational degrees of grey theory are used to compute the similarities of these blocks,on which the high-frequency coefficients are fused,while the arithmetic mean method is used to fuse the low-frequency coefficients.Finally,a fused image is reconstructed with the fused coefficients by performing the inverse Curvelet transform.Experimental results show that the proposed method is superior to pyramid-based methods and wavelet-transform-based methods in terms of entropy,standard deviation and clarity.

Key words: image fusion, Curvelet transform, grey relational analysis, similarity

中图分类号: