测绘学报 ›› 2016, Vol. 45 ›› Issue (5): 574-580.doi: 10.11947/j.AGCS.2016.20140546

• 摄影测量学与遥感 • 上一篇    下一篇

干涉复小波复数域双变量滤波算法

何永红1,2, 朱建军1, 解清华1, 许兵1, 付海强1   

  1. 1. 中南大学地球科学与信息物理学院, 湖南 长沙 410083;
    2. 湖南科技学院土木与环境工程学院, 湖南 永州 425199
  • 收稿日期:2014-10-24 修回日期:2016-03-09 出版日期:2016-05-20 发布日期:2016-05-30
  • 作者简介:何永红(1978-),女,博士生,副教授,研究方向为遥感数据处理及应用。E-mail: heyonghong2004@163.com
  • 基金资助:
    国家自然科学基金(41274010;41474008);湖南省自然科学基金(14JJ2131)

An Bivariate Filtering Algorithm for Interfere Complex Wavelet Complex Domain

HE Yonghong1,2, ZHU Jianjun1, XIE Qinghua1, XU Bing1, FU Haiqiang1   

  1. 1. School of Geosciences and Info-Physics, Central South University, Changsha 410083, China;
    2. School of Civil and Environmental Engineering, Hunan University of Science and Engineering, Yongzhou 425199, ChinaAbstract
  • Received:2014-10-24 Revised:2016-03-09 Online:2016-05-20 Published:2016-05-30
  • Supported by:
    The National Natural Science Foundation of China(Nos.41274010;41474008);The National Natural Science Foundation of Hunan Province of China(No.14JJ2131)

摘要: 针对复小波双变量滤波模型仅考虑小波复系数实部,忽略了系数的虚部,导致信号相位噪声的增加而影响滤波效果的问题,提出基于复小波变换的复数域双变量模型干涉图滤波算法。该算法将双变量贝叶斯估计算法从实数域推广到了复数域,用噪声复系数概率密度函数刻画了小波复系数实部与虚部的相关性,根据小波分解复系数来估计噪声方差和信号方差,建立了复小波复数域双变量滤波模型,求得了干涉图复系数的贝叶斯估计。试验结果表明,本算法对干涉图噪声有较强的抑制能力,保留了干涉图的边缘及细节信息,滤波性能优于传统的实数域复小波双变量滤波、Goldstein滤波、单小波滤波和最优化融合滤波方法。

关键词: 干涉图, 复小波, 双变量, 滤波

Abstract: For complex wavelet bivariate filtering only considering real part, ignoring the imaginary part, it directly leads to the increase of signal phase noise,and influences the effect of filtering, An filtering algorithm based on complex wavelet and complex domain bivariate mode is proposed.Double variable Bayesian estimation algorithm was promoted from real field to complex field, the correlations between real and imaginary part were depicted using noise probability density function, the noise variance and the signal variance were estimated making use of the wavelet decomposition complex coefficient,complex domain double variable filtering model was established, Bayesian estimation of interferogram was obtained. Experimental results show that the algorithm has a stronger inhibitory noise ability, keep the edges and details of interferogram, the filtering performance is better than real domain complex wavelet bivariate filtering, Goldstein filtering, single wavelet filtering and optimal integration filtering methods.

Key words: interferogram, complex wavelet, bivariate, filter

中图分类号: