Visible-light　image　and　infrared　image　fusion　technology　can　be　used　to　obtain　an　image　that　has　both　the　detailed　texture　of　the　visible-light　image　and　the　characteristics　of　infrared　heat　radiation.　However,　most　existing　fusion　algorithms　inevitably　introduce　noise　from　the　source　image　during　fusion.　To　solve　this　problem,　an　image　fusion　algorithm　based　on　a　multiscale　transformation　framework　is　proposed　in　this　study.　This　algorithm　first　uses　multiscale　transformation　theory　to　decompose　the　source　images,　and　then　uses　a　method　based　on　the　standard　deviation　of　the　local　area　as　the　fusion　rule　of　the　low-frequency　components.　In　addition,　a　visual　saliency　detection　method　based　on　guided　filtering　is　used　to　extract　weight　mapping　of　the　high-frequency　components,　which　are　weighted　and　fused　according　to　the　distribution　of　the　weight　mapping.　Finally,　the　components　are　inversely　transformed　to　obtain　a　visible　light　-　infrared　fusion　image.　The　experiments　show　that　the　proposed　algorithm　can　effectively　maintain　the　detailed　texture　of　visible-light　images,　as　well　as　infrared　heat　radiation　information,　and　has　the　advantages　of　low　noise,　high　contrast,　and　a　better　visual　effect.