Fusion Method Of Infrared And Visible Images

As an importantpart in image data fusion, the fusion of infrared image andvisible image has attracted widespread attention and deeply studied among theworld with the infrared technology has been catholically applied to biology,medicine,geography,militariesandother fields. Imagefusionis a kindof effectiveway to mergethe infraredimageinformationwithvisible imageinformationat thesame time, whichhasbeenwidelyused in targetdetection,targetidentificationandtarget tracking,etc., as well as in machinevision,astronomy,andexploration.Focus on the problems exist in the fusion image information contentinsufficient,block phenomenonserious, fusion method limitationstrong ect. Thispaperstudiestwomajoraspects:First,improvesthe fusionrulesbased on the existingfusionframe,solvestheinformation insufficient and "block phenomenon" of the fusion image; Second,proposes the image fusion algorithm based on Fractional Fourier transformbecause of the superior characteristics of handlingthe signal when applyingit tothe image processing.This is not only suits for the infrared and visible imagefusionbutalso can be foundmoreapplicabilityin the generalimagefusion.The main contributionof thispaperlisted asfollows:(1)Using the division mechanism to the infrared imagery,obtains the targetregion and the backgroundregion, then uses two generation of wavelets (NSTC)decompositionthe infraredand the visible light image.In the targetregionselectsthe big frequency coefficients and the background region uses based on thewindowpartialinformationentropyfusionmethodstrategy.(2)For the final fusion results obtained, proposing the "consistencyverification" rule,has avoided the black phenomenonwhich appearsin the fusionimage.(3)Consideringthe fractionalFouriertransformation(FRFT)of arbitraryangle rotationfeature,appliesit to infraredandvisibleimagesof the decomposition,thenthrough genetic algorithm to find the best value which contents the maximuminformation,thenutilizationthe effectivefusion rulesfused it.