【7A文】High Resolution Imaging(高清晰度成像技术).ppt

《【7A文】High Resolution Imaging(高清晰度成像技术).ppt》由会员上传分享，免费在线阅读，更多相关内容在工程资料-天天文库。

1、0HighResolutionImaging1OutlineMotivationHistoricalEvolutionofIdeasImagingwithQuadraticRegularizationRegularizedLeastSquaresMigration(RLSM)PreconditionedLeastSquaresMigration(PLSM)SyntheticsCasestudy(WCSB)ImagingwithNon-quadraticRegularizationWheredowegofromhere?SummaryAcknowledgments2HighR

2、esolutionImaging:MotivationTogobeyondtheresolutionprovidedbythedata(apertureandband-width)byincorporatingquadraticandnon-quadraticregularizationtermsintomigration/inversionalgorithmsThisisnotanewidea…3LSDeconvolutionSparseSpikeDeconvolutionLSRadonTransformsHRRadonTransformsL2[Gaussian&Bori

3、ng]StableandFastAlgorithmsLowResolutionNon-quadraticNorms[Non-Gaussian]RequiresSophisticatedOptimizationEnhancedResolutionHistoricalEvolutionofConcepts:Deconvolution&RadonTransforms4MigrationwithAdjointOperators[Currenttechnology]LSMigration+QuadraticRegularization[Notinproductionyet]Migra

4、tion+Non-QuadraticRegularization[??]ResolutionEvolutionofConcepts:Migration5ASimpleAnalogy6Convolution/Cross-correlation/DeconvConvolutionCross-correlationDeconvolution7QuadraticandNon-quadraticRegularizationCostQuadraticNon-quadraticNon-quadraticnormsthathavebeenusedtoenhanceresolution:L1

5、-norm,CauchyNorm,HuberNorm89AdjLSHR10AnotherSimpleAnalogy11LSRadonhq12HRRadonqh13Priorinformationintheformofquadraticand/ornon-quadraticregularizationtermsplaysanimportantroleinthedeconvolutionandRadoninverseproblemsQ.IsthesamevalidforMigration/Inversion?Backtomigration14Priorinformationin

6、theformofquadraticand/ornon-quadraticregularizationtermsplayanimportantroleinthedeconvolutionandRadoninverseproblemsQ.IsthesamevalidforMigration/Inversion?A.Notsofast…firstweneedtodefinewhatweareinvertingfor..Backtomigration15MigrationasanInverseProblemForward(De-migration)AdjointData“Imag

7、e”orAngleDependentReflectivity16StructuralImaging:Thereisnogeneralagreementaboutthetypeofregularizationoneshouldusewheninvertingform=m(x,z)SparsesolutionsmightonlybeusefulforimaginginthescatteringregimeAVA/AVPImaging:Forangledependentimages,m(x,z,p),weattemptt