电子学报 ›› 2016, Vol. 44 ›› Issue (4): 761-766.DOI: 10.3969/j.issn.0372-2112.2016.04.002

• 学术论文 • 上一篇    下一篇

NSCT域内结合边缘特征和自适应PCNN的红外与可见光图像融合

闫利1, 向天烛1,2   

  1. 1. 武汉大学测绘学院, 湖北武汉 430079;
    2. 武汉大学测绘遥感信息工程国家重点实验室, 湖北武汉 430079
  • 收稿日期:2014-09-25 修回日期:2015-04-30 出版日期:2016-04-25
    • 作者简介:
    • 闫 利 男,1966年8月出生于山西山阴,1999年在武汉大学获得大地测量与测量工程博士,现为武汉大学教授、博士生导师,主要从事摄影测量与遥感、地面三维激光扫描的研究. E-mail:lyan@sgg.whu.edu.cn;向天烛 男,1988年12月出生于湖北宜昌,现为武汉大学摄影测量与遥感专业博士生,主要从事摄影测量、遥感图像处理的研究. Email:skylabs@whu.edu.cn
    • 基金资助:
    • 国家自然科学基金 (No.41271456)

Fusion of Infrared and Visible Images Based on Edge Feature and Adaptive PCNN in NSCT Domain

YAN Li1, XIANG Tian-zhu1,2   

  1. 1. School of Geodesy and Geomatics, Wuhan University, Wuhan, Hubei 430079, China;
    2. State key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan, Hubei 430079, China
  • Received:2014-09-25 Revised:2015-04-30 Online:2016-04-25 Published:2016-04-25
    • Supported by:
    • National Natural Science Foundation of China (No.41271456)

摘要:

针对传统的基于多尺度变换的红外与可见光图像融合,对比度不高,边缘等细节信息保留不充分等问题,结合NSCT变换的多分辨率、多方向特性和PCNN全局耦合、脉冲同步激发等优点,提出一种基于NSCT变换结合边缘特征和自适应PCNN红外与可见光图像融合算法.对于低频子带,采用一种基于边缘的融合方法;对于高频方向子带,采用方向信息自适应调节PCNN的链接强度,使用改进的空间频率特征作为PCNN的外部激励,根据脉冲点火幅度融合子带系数.实验结果验证了该算法的有效性.

关键词: 图像融合, 非下采样Contourlet变换, 脉冲耦合神经网络, 边缘特征, 空间频率, 红外图像

Abstract:

To improve the contrast and preserve more image details in the fusion of infrared and visible images,a fusion method for infrared and visible images based on nonsubsampled contourlet transform (NSCT) combined with image edge feature and adaptive pulse coupled neural network (PCNN) is proposed.For the low frequency subband,the fusion is based on edges of images.For the high frequency subbands,the orientation information of each pixel in images is utilized as the linking strength,and a modified spatial frequency is adopted as the input to motivate the adaptive PCNN,and the fire amplitude is employed to determine the coefficients selection.Experimental results indicate the effectiveness of the proposed algorithm.

Key words: image fusion, nonsubsampled contourlet transform, pulse coupled neural network, edge feature, spatial frequency, infrared image

中图分类号: