电子学报 ›› 2012, Vol. 40 ›› Issue (10): 1980-1983.DOI: 10.3969/j.issn.0372-2112.2012.10.013

• 学术论文 • 上一篇    下一篇

基于均值滤波矩阵组的二维信号快速经验模式分解方法

郭珈1,2, 王孝通2,3, 徐晓刚2,4, 张成堡5   

  1. 1. 海军91550部队, 辽宁大连 116023;
    2. 海军大连舰艇学院光电技术研究所, 辽宁大连 116018;
    3. 海军大连舰艇学院航海系, 辽宁大连 116018;
    4. 海军大连舰艇学院装备系统与自动化系, 辽宁大连 116018;
    5. 海军32132部队, 山东青岛 266405
  • 收稿日期:2011-07-18 修回日期:2012-08-25 出版日期:2012-10-25
    • 作者简介:
    • 郭 珈 女,1983年8月出生于辽宁大连市,现为海军91550部队工程师、博士.研究方向为图像视频处理. E-mail:guojia828@163.com 王孝通 男,1962年3月出生于浙江义乌,现为海军大连舰艇学院航海系教授、博士后、博士生导师.研究方向为信号分析、图像处理、电子海图等. E-mail:dljtxywxt@163.com
    • 基金资助:
    • 国家自然科学基金 (No.60975016); 青年科学基金 (No.61002052); 辽宁省自然科学基金 (No.20082176)

Fast 2D Empirical Mode Decomposition Based on Mean Value Matrix Filter Sets

GUO Jia1,2, WANG Xiao-tong2,3, XU Xiao-gang2,4, ZHANG Cheng-bao5   

  1. 1. Unit 91550, Dalian, Liaoning 116023, China;
    2. Institute of Photoelectric Technology, Dalian Naval Academy, Dalian, Liaoning 116018, China;
    3. Department of Navigation, Dalian Naval Academy, Dalian, Liaoning 116018, China;
    4. Department of Automatization, Dalian Naval Academy, Dalian, Liaoning 116018, China;
    5. Unit 92132, Qingdao, Shandong 266405, China
  • Received:2011-07-18 Revised:2012-08-25 Online:2012-10-25 Published:2012-10-25
    • Supported by:
    • National Natural Science Foundation of China (No.60975016); Youth Fund of National Natural Science Foundation of China (No.61002052); Natural Science Foundation of Liaoning Province,  China (No.20082176)

摘要: 根据矩阵变换理论,提出了一种基于滤波矩阵组的局部均值求取方法.该方法构造了均值滤波矩阵组,只需在待处理矩阵左右各乘一个均值滤波阵,即可得到待处理阵的加权均值,简化了包络均值求解过程.在上述局部均值滤波理论的基础上,提出了二维信号的快速经验模式分解算法.实验证明,算法运算时间比传统经验模式分解方法提高了两到三个数量级.

关键词: 经验模式分解(EMD)快速算法, 均值滤波矩阵, 图像融合

Abstract: A new local mean value computing algorithm is proposed based on the theory of the mean value matrix filter sets.The algorithm computes the weighted mean value of the matrix by multiplying a mean value filter matrix to each side of it.This special matrix can be calculated quickly and simplify the process of envelope computing.Based on the mean value matrix filter sets,a fast two-dimension EMD algorithm is proposed.Experiments show this algorithm works well on image fusion.It can also reduce the computation time by 1/100~1/1000 than that of the traditional EMD.

Key words: fast algorithm of empirical mode decomposition (EMD), mean value filter matrix, image fusion

中图分类号: