何永滔
(中山大學(xué)數(shù)學(xué)與計(jì)算科學(xué)學(xué)院,廣東 廣州 510275)
緊支撐正交的二維小波
何永滔
(中山大學(xué)數(shù)學(xué)與計(jì)算科學(xué)學(xué)院,廣東 廣州 510275)
基于Householder矩陣擴(kuò)充,構(gòu)造了緊支撐正交的二維小波,所構(gòu)造小波函數(shù)的支撐不超過(guò)尺度函數(shù)的支撐,并且給出了容易實(shí)施的顯式構(gòu)造算法.另外,還通過(guò)構(gòu)造反例說(shuō)明Riesz定理不適用于二元三角多項(xiàng)式.最后,構(gòu)造了算例.
多分辨分析;仿酉矩陣擴(kuò)充;二維正交小波;多相位分解;Riesz定理
眾所周知,小波因具有良好的時(shí)頻局部性而廣泛地應(yīng)用于信號(hào)分析,圖象處理,邊緣檢測(cè),數(shù)值計(jì)算等領(lǐng)域.構(gòu)造小波是小波分析的核心內(nèi)容,多分辨分析是構(gòu)造小波的重要工具.在實(shí)際應(yīng)用中,小波的緊支撐性能使快速小波變換是有限和,小波的對(duì)稱(chēng)性能使信號(hào)避免失真,小波的正交性能夠保持信號(hào)的能量,因此構(gòu)造具有緊支撐,對(duì)稱(chēng)性,正交性等良好性質(zhì)的小波就成為小波分析工作者關(guān)注的熱點(diǎn)之一.目前已有相當(dāng)多的文獻(xiàn)研究小波的構(gòu)造,其中一維小波的研究成果比較成熟.1988年文獻(xiàn)[1]給出了一維2帶緊支撐正交小波的構(gòu)造.但除Haar小波外,一維2帶緊支撐正交小波不可能具有對(duì)稱(chēng)性,這就限制了它在現(xiàn)實(shí)生活中的應(yīng)用.人們開(kāi)始關(guān)注其他尺度因子的情形,1995年文獻(xiàn)[2]構(gòu)造了一維3帶緊支撐正交對(duì)稱(chēng)與反對(duì)稱(chēng)的小波,研究了其精細(xì)結(jié)構(gòu).1998年文獻(xiàn) [3]構(gòu)造了一維4帶對(duì)稱(chēng)正交的尺度函數(shù)與小波.1999年文獻(xiàn)[4]構(gòu)造了一維M帶的緊支撐小波,研究了所構(gòu)造小波的性質(zhì).高維小波是處理高維信號(hào)的重要工具,但是有關(guān)高維小波構(gòu)造的文獻(xiàn)比較少.雖然文獻(xiàn)[5]利用Householder矩陣擴(kuò)充構(gòu)造了高維正交小波,構(gòu)造過(guò)程簡(jiǎn)單清晰且易于實(shí)施,但擴(kuò)充所得的矩陣元素中含有Laurent多項(xiàng)式分母.這就使得當(dāng)所給的尺度函數(shù)具有緊支撐性時(shí),所構(gòu)造的高維小波可能沒(méi)有緊支撐性.本文通過(guò)Householder矩陣擴(kuò)充構(gòu)造了緊支撐正交的二維小波,擴(kuò)充所得的矩陣元素分母中不含Laurent多項(xiàng)式,所構(gòu)造的緊支撐正交二維小波的支撐不超過(guò)尺度函數(shù)的支撐.另外,本文通過(guò)構(gòu)造反例說(shuō)明Riesz定理不適用于二元三角多項(xiàng)式.最后,給出了構(gòu)造算例.
為了敘述的方便,引入記號(hào):
致謝感謝程?hào)|升博士在編程方面有益的幫助.
[1]Daubechies I.Orthonormal bases of compactly supported wavelets[J].Comm.Pure Appl.Math.,1988, 41(7):909-996.
[2]Chui C K,Lian J A.Construction of compactly supported symmetric and antisymmetric orthonormal wavelets with scale=3[J].Appl.Comput.Harmon.Anal.,1995,2(1):21-51.
[3]Han B.Symmetric orthonormal scaling functions and wavelets with dilation factor 4[J].Adv.Comput. Math.,1998,8(3):221-247.
[4]Bi N,Dai X R,Sun Q Y.Construction of compactly supported M-band wavelets[J].Appl.Comput.Harmon. Anal.,1999,6(2):113-131.
[5]龍瑞麟.高維小波分析[M].北京:世界圖書(shū)出版公司,1995.
[6]Petukhov A.Construction of symmetric orthogonal bases of wavelets and tight wavelet frames with integer dilation factor[J].Appl.Comput.Harmon.Anal.,2004,17(2):198-210.
[7]Lawton W,Lee S L,Shen Z W.An algorithm for matrix extension and wavelets construction[J].Math. Comp.,1996,65(214):723-737.
[8]Chui C K,He W J.Compactly supported tight frames associated with re fi nable functions[J].Appl.Comput. Harmon.Anal.,2000,8(3):293-319.
[9]Han B,Jia R Q.Optimal interpolatory subdivision schemes in multidimensional spaces[J].SIAM J.Numer. Anal.,1999,36(1):105-124.
Compactly supported orthogonal bivariate wavelets
He Yongtao
(School of Mathematics and Computational Science,Sun Yat-sen University, Guangzhou 510275,China)
Based on the Householder matrix extension method,we construct compactly supported orthogonal bivariate wavelets.The supports of the constructed wavelets are not larger than that of scaling function,an explicit algorithm that can be easily applied is also presented.Furthermore,we prove that Riesz theorem can not be applied to bivariate trigonometrical polynomial.Finally,an example is given.
multiresolution analysis,paraunitary matrix extension,bivariate orthogonal wavelets, polyphase decomposition,riesz theorem
O174
A
1008-5513(2012)01-0008-09
2011-07-15.
國(guó)家自然科學(xué)基金(11071261,10911120394).
何永滔(1979-),博士生,研究方向:時(shí)頻分析與圖像處理.
2010 MSC:42C40,65T60