预览加载中,请您耐心等待几秒...
1/3
2/3
3/3

在线预览结束,喜欢就下载吧,查找使用更方便

如果您无法下载资料,请参考说明:

1、部分资料下载需要金币,请确保您的账户上有足够的金币

2、已购买过的文档,再次下载不重复扣费

3、资料包下载后请先用软件解压,在使用对应软件打开

基于参数化模型的遥感图像匀光算法 Abstract: Imageuniformizationisanimportantpreprocessingstepinmanyremotesensingapplications.Traditionalmethodsoftenrelyonhandcraftedfeaturesorintensecomputation,whichlimitstheirefficiencyandpracticality.Parameterizedmodelsprovideanewsolutionforimageuniformizationbylearningthetransformfunctionfromdata.Inthispaper,weintroducetheconceptofparameterizedmodelsandtheirapplicationsinremotesensingimageuniformization.Wereviewseveralpopularparameterizedmodelsandanalyzetheiradvantagesandlimitations.Finally,wepresentsomeopenproblemsandpotentialresearchdirectionsinthisfield. Introduction: Remotesensingimageuniformizationisachallengingtasksincetheilluminationcondition,sensorresponse,andatmosphericeffectvaryacrossdifferentscenesandtime.Imageuniformizationseekstoremovethesevariationsandenhancethevisualqualityandinterpretabilityoftheimages.Traditionalmethodsoftenrelyonhandcraftedfeaturesorcomplexalgorithms,whicharesensitivetotheinputparametersanddomain-specificassumptions.Inrecentyears,parameterizedmodelshaveemergedasanewsolutionforimageuniformization.Parameterizedmodelslearnthetransformationfunctionfromdata-drivenapproaches,whichcanadapttodifferentimagingconditionsandlearnmoreeffectiverepresentations. Inthispaper,wepresentacomprehensivereviewoftheparameterizedmodelsforremotesensingimageuniformization.Wefirstintroducetheconceptofparameterizedmodelsandtheiradvantagesovertraditionalmethods.Then,wereviewseveralpopularparameterizedmodels,includingthehistogramequalization,adaptivehistogramequalization,contraststretch,andgammacorrection.Weanalyzethemathematicalfoundation,algorithmicimplementation,andempiricalperformanceofthesemodels.Wealsocomparethemwithtraditionalmethodsintermsofefficiency,accuracy,androbustness. Parameterizedmodels: Theparameterizedmodelsforremotesensingimageuniformizationcanbegroupedintotwocategories:globalandlocal.Globalmodelsaimtoequalizetheoverallimageintensitydistribution,whilelocalmodelsfocusonimprovingthelocalcontrastandremovingthenoise.Ingeneral,g