Abstract: mage　feature 　extraction　and　matching　will　be　done　after 　color 　image　is 　transformed 　into 　gra ymay　cause 　color 　information 　loss 　and 　lead 　to　wrong　matching．in　order 　to 　solve　problem this　paperpresents 　a　method　color－based　scale－invariant 　feature 　transform　to 　realize 　feature 　image　matching．thenfeature 　points 　are 　extracted　by　using　improved　harris　al gorithm　ei ghbor 　information 　aroundthese　points 　is 　described　using　color 　invariant 　value　as 　input 　information．finall ythe　points 　betweentwo　images 　are　matched　using　correlation　coefficient．to　demonstrate 　the 　feasibilit y　of 　approachex－periments　with　groups　of 　street 　view 　imagescompared　with　the 　traditional 　sift　al gorithm．show　thatthe　proposed　al gorithm　can　not　onl y　intensive　match　pointshi gher　matching　correct 　rate 　speed　validity．