红外图像增强算法研究及其FPGA实现
Abstract
Infraredimagereflectsthespatialdistributionofinfraredradiationfromtargetsand
backgrounds.Theoriginalinfraredimageobtainedbyinfrareddetectorshavepoorvisual
effects,suchaspoorcontrast,lowsignal-to-noiseratio,andunclearimagedetail.Theoriginal
infraredimagemustbeenhancedtobesuitableforhumanobservation.FPGAhasthe
advantagesofparallelprocessingandpipelineprocessing.Therefore,theaimofthisarticleis
tostudyaninfraredimageenhancementalgorithmandtransplantittoFPGAfor
implementation,achievinghighdynamicinfraredimageenhancementprocessinganddisplay.
Therelevanttheoreticalbasisofinfraredimagewasfirstintroducedinthearticle,which
focusesonresearchingbasicinfraredimageenhancementalgorithmswithhighreal-time
performanceandeasyhardwareimplementation.Usinghistogram,contrast,signal-to-noise
ratioandimagedetailasdividingpoints,thispaperstudiedbasicinfraredimageenhancement
algorithms.
Secondly,inthestudyofinfraredimageindynamicscenes,animprovedinfraredimage
enhancementalgorithmbasedonimagelayeringwasproposed.Useguidedfilteringalgorithm
toperformlayeredprocessingoninfraredimage,obtainingbaselayerimagecontaininglarge
dynamictemperatureinformationanddetaillayerimagecontainingsmalldynamic
temperatureinformation;Thedetaillayerguidesthefilteringprocessagain,andtheimagehas
animportantsmallamountofgrayscaleinformation,whichcompensatesforthetarget
grayscaleinformationofthebaselayerimage.Useadaptiveplatformhistogramstoenhance
contrastandcompressdynamicrangeofcompensatedbaselayerimage;Usemedianfiltering
andadaptiveS-curvesuccessivelytoremovenoiseandenhanceimagedetailsindetaillayer
image;Performlinearweightedfusiononthebaselayeranddetaillayeroftheimage
proces