国土资源遥感
國土資源遙感
국토자원요감
REMOTE SENSING FOR LAND & RESOURCES
2014年
3期
61-66
,共6页
经验模式分解( EMD)%加权滤波经验模式分解( WFEMD)%图像融合%内涵模式分量( IMF)
經驗模式分解( EMD)%加權濾波經驗模式分解( WFEMD)%圖像融閤%內涵模式分量( IMF)
경험모식분해( EMD)%가권려파경험모식분해( WFEMD)%도상융합%내함모식분량( IMF)
empirical model decomposition(EMD)%weighted filter empirical model decomposition(WFEMD)%image fusion%intrinsic model functions(IMF)
加权滤波经验模式分解(weighted filter empirical mode decomposition,WFEMD)作为一种新的多尺度、多分辨率分析方法,与小波、超小波和现有二维经验模式分解方法相比,更加适合于二维图像中的细节特征分析。该方法运用自适应加权滤波器直接求取均值面,解决了传统二维经验模式分解( empirical mode decomposition,EMD)方法的固有缺陷；将WFEMD方法引入遥感图像融合,能够更好地提取原始图像的特征,为图像融合提供更多的信息。鉴于此,提出了一种基于WFEMD变换的图像融合方法。首先,利用WFEMD的自适应性、多尺度性和高频细节信息的强获取能力,将待融合的图像分别进行WFEMD分解,对不同图像的内涵模式分量( intrinsic mode functions, IMF)按照该文提出的细节/背景原则进行融合,剩余分量按照平均原则进行融合。最后,将融合后的内涵模式分量重构,获取融合图像。实验证明,该方法的融合效果优于其他图像融合方法。
加權濾波經驗模式分解(weighted filter empirical mode decomposition,WFEMD)作為一種新的多呎度、多分辨率分析方法,與小波、超小波和現有二維經驗模式分解方法相比,更加適閤于二維圖像中的細節特徵分析。該方法運用自適應加權濾波器直接求取均值麵,解決瞭傳統二維經驗模式分解( empirical mode decomposition,EMD)方法的固有缺陷；將WFEMD方法引入遙感圖像融閤,能夠更好地提取原始圖像的特徵,為圖像融閤提供更多的信息。鑒于此,提齣瞭一種基于WFEMD變換的圖像融閤方法。首先,利用WFEMD的自適應性、多呎度性和高頻細節信息的彊穫取能力,將待融閤的圖像分彆進行WFEMD分解,對不同圖像的內涵模式分量( intrinsic mode functions, IMF)按照該文提齣的細節/揹景原則進行融閤,剩餘分量按照平均原則進行融閤。最後,將融閤後的內涵模式分量重構,穫取融閤圖像。實驗證明,該方法的融閤效果優于其他圖像融閤方法。
가권려파경험모식분해(weighted filter empirical mode decomposition,WFEMD)작위일충신적다척도、다분변솔분석방법,여소파、초소파화현유이유경험모식분해방법상비,경가괄합우이유도상중적세절특정분석。해방법운용자괄응가권려파기직접구취균치면,해결료전통이유경험모식분해( empirical mode decomposition,EMD)방법적고유결함；장WFEMD방법인입요감도상융합,능구경호지제취원시도상적특정,위도상융합제공경다적신식。감우차,제출료일충기우WFEMD변환적도상융합방법。수선,이용WFEMD적자괄응성、다척도성화고빈세절신식적강획취능력,장대융합적도상분별진행WFEMD분해,대불동도상적내함모식분량( intrinsic mode functions, IMF)안조해문제출적세절/배경원칙진행융합,잉여분량안조평균원칙진행융합。최후,장융합후적내함모식분량중구,획취융합도상。실험증명,해방법적융합효과우우기타도상융합방법。
Weighted filter empirical mode decomposition( WFEMD) , as a new multi-scale and multi-resolution analysis algorithm, is more appropriate for the analysis of the image details than wavelet, super wavelet and dimensional empirical mode decomposition, and can solve the inherent defects of the traditional two-dimensional empirical mode decomposition( EMD) . The main reason is that it directly computes the mean envelope by adaptive weighted mean filter. When WFEMD is introduced to the remote sensing image fusion, the characteristics of original images can be better extracted, and more information for fusion can be obtained. Firstly, the source images are decomposed by using WFEMD with the capability of acquirement of the high frequency data, the adaptability for some intrinsic mode functions ( IMF) and the residual component, and then the IMFs and the residual component are fused with the details/background and average fusion regularity respectively at the corresponding scales. Finally, the fused IMFs and the residual component are reconstructed to obtain fusion results. Experiments have shown that the proposed algorithm is efficient in image fusion and is better than other current algorithms.
