|
自动化学报 2005
基于期望值最大算法和离散小波框架的图像融合, PP. 699-704 Keywords: Imagefusion,maximizationalgorithm,discretewaveletframetransform Abstract: ?Thediscretewavelettransformhasbecomeanattractivetoolforfusingmultisensorimages.Thispaperinvestigatesthediscretewaveletframetransform.Amajoradvantageofthismethodoverdiscretewavelettransformisaliasingfreeandtranslationinvariant.Thediscretewaveletframe(DWF)transformisusedtodecomposetheregisteredimagesintomultiscalerepresentationwiththelowfrequencyandthehighfrequencybands.Thelowfrequencybandisnormalizedandfusedbyusingtheexpectationmaximization(EM)algorithm.Theinformativeimportancemeasureisappliedtothehighfrequencyband.Thefinalfusedimageisobtainedbytakingtheinversetransformonthecompositecoefficientrepresentations.Experimentsshowthattheproposedmethodismoreeffectivethanconventionalimagefusionmethods.
|