As　for　ASIFT,　ASIFT　has　been　proven　to　be　invariant　to　image　scaling　and　rotation.　Specially,　ASIFT　enables　matching　of　images　with　severe　viewpoint　change　and　outperforms　significantly　the　state-of-the-art　methods.　It　accomplished　this　by　simulating　several　views　of　the　original　images.　However,　we　found　that　the　simulated　parameters　are　continuous,　namely,　transformations　acquired　by　ASIFT　cant　express　the　real　relationship　between　reference　and　input　images.　Therefore,　a　particle　swarm　optimization　based　sample　strategy　is　presented　in　this　paper.　The　basic　idea　is　to　search　the　best　transform　in　continuous　parameter　space.　Experimental　results　show　that　the　proposed　PSO-ASIFT　algorithm　could　get　more　matches　compared　with　the　original　ASIFT　and　SIFT.　©　Springer-Verlag　Berlin　Heidelberg　2014.