小波包多阈值去噪法及其在形变分析中的应用.pdf

《小波包多阈值去噪法及其在形变分析中的应用.pdf》由会员上传分享，免费在线阅读，更多相关内容在行业资料-天天文库。

1、第43卷第1期测绘学报VoI．43，No．12O14年1月ActaGeodaeticaetCartographicaSinicaJan．，2O14引文格式：ZHANGZhetao，ZHUJianjun，KUANGCuilin，eta1．Multi—thresholdWaveletPacketDe—noisingMethodandItsApplicationinDeformationAnalysis[J]．ActaGeodaeticaetCartographicaSinica，2014，43(1)：1320．(章

2、浙涛，朱建军，匡翠林，等小波包多阈值去噪法及其在形变分析中的应用EJ]．测绘学报，2Ol4，43(1)：13-2o．)DOI：10．13485／j．cnki．11-2089．2014．0003小波包多阈值去噪法及其在形变分析中的应用章浙涛，朱建军，匡翠林，周璀中南大学地球科学与信息物理学院，湖南长沙410083MuIti—threshoIdWaveletPacketDe—noisingMethodandltsAppIicationinDeformationAnaIysisZHANGZhetao，ZHUJiani

3、un，KUANGCuiIin，ZHOUCuiSchoolofGeosciencesandInfo—Physics，CentraISouthUniversity，Changsha410083，ChinaAbstract：Inthefieldofdeformationmenitoring，thetraditiona1waveletde—noisingmethodretainsonIYthelowfrequencyofusefuIinformation．Itiseasytogetridofintermediatefr

4、equencyandhighfrequencyusefuIinformation。ThewaveletpacketanaIysisisanewkindofwaveletanaIysismethoddevelopedinrecentyears．whichisamoresubtlede—noisingmethodforconsideringusefuIinformationofihevariousbands．ThekeyofthewaveIeipackeide—noisingis1oseIecttheappropr

5、iatethresholdcriteriaandtoprocessthewaveletpacketdecompositioncoefficientsbythethreshoId．buttheresearchesusingtraditiona1waveletpacketde—noisingmethodarenotsufficieni．ThisarlicleisfortheIackoftraditiona1waveIetandwaveIeipackelanaIysis．Accordingtothedistribut

6、ionofdifferentsignaIsandtheirnoise，waveletpacketdecompositioncoefficientsarearrangedbythefrequencyorder-andsegmentedinaccordancewithinformationtype，toselecttheappropriatethresholdcriteriaforeachbandandtoperformthresholdprocessing．Itisiheme{hodofwaveIe{packe1

7、de—noisingwithmuIti-threshoIdcriieriabasedonfrequencYorder．TheresuItsshowthatthismethodcaneffectiveIYremovethenoiseofeachbandthroughtheoreticaIanaIysisandpracticaIaPPIications．Thede—noisingabiIityofthismethodisbetterthantheothermethodssuchastraditiona1wavele

8、tde—noisingorwaveletpacketde—noising．SiudieshaveshownthatihismethodcanPreservetheusefuIinformationfromthede—noisingsignaIafterde—noisingwhenthesamplingfrequencyislow．Therefore，itcanbewidelYusedi