简介:Studyofimageenhancementshowsthatthequalityofimageheavilyreliesonhumanvisualsystem.Inthispaper,weapplythisfacteffectivelytodesignanewimageenhancementmethodformedicalimagesthatimprovesthedetailregions.First,theeyeregionofinterest(ROI)issegmented;thentheUn-sharpMasking(USM)isusedtoenhancethedetailregions.Experimentsshowthattheproposedmethodcaneffectivelyimprovetheaccuracyofmedicalimageenhancementandhasasignificanteffect.
简介:Thisarticlepresentsagoodrobustandreal-timesystemschemeofthemobilerobotobstacledetectionandnavigation,whichprincipleofworkisbasedonthefeaturedescriptorSURF.Inthisscheme,firstly,theimageinformationofthemobilerobotpathwascapturedbythebinocularcamera;thenthefeaturepointswereextractedandcorrespondingmatchedusingSURFtothebinocularimagesastheundetectedobstacles;finallyfixedthepositionoftheobjectivebytheparallaxbetweenthematchingpointscombiningwiththebinocularvisioncalibrationmodel.Theoreticalderivationandexperimentalresultsshowthatthisschemeismoreaccurateforthedetectionandnavigationoftheinterestpoints.Ithasfastmatchingspeedandhighaccuracyandlowerror.So,ithascertainpracticaleffectandpopularizingvalueforthemobilerobotreal-timeobstacleavoidanceandnavigation.