简介：AccordingtotheprincipleofpolarizationimagingandtherelationbetweenStokesparametersandthedegreeoflinearpolarization,therearemuchredundantandcomplementaryinformationinpolarizedimages.Sinceman-madeobjectsandnaturalobjectscanbeeasilydistinguishedinimagesofdegreeoflinearpolarizationandimagesofStokesparameterscontainrichdetailedinformationofthescene,thecluttersintheimagescanberemovedefficientlywhilethedetailedinformationcanbemaintainedbycombiningtheseimages.Analgorithmofadaptivepolarizationimagefusionbasedonregionalenergydynamicweightedaverageisproposedinthispapertocombinetheseimages.Throughanexperimentandsimulations,mostcluttersareremovedbythisalgorithm.Thefusionmethodisusedfordifferentlightconditionsinsimulation,andtheinfluenceoflightingconditionsonthefusionresultsisanalyzed.