Aiming　at　the　problem　of　video　image　captured　by　the　monitoring　system　under　the　conditions　of　haze,　rain,　snow　and　uneven　illumination,　a　classification　method　of　different　illumination　is　proposed　in　this　paper.　Through　analyzing　the　characteristics　of　different　illumination　images,　the　features　of　different　illumination　images　can　be　extracted.　The　different　illumination　image　features　can　be　used　to　train　and　construct　the　classifier.　Finally,　the　different　illumination　images　are　classified　by　the　classifier.　The　experimental　results　show　that　the　support　vector　machine　(SVM)　algorithm,　BP　neural　network　algorithm　and　k-means　algorithm　all　can　achieve　the　classification　of　different　illumination　images,　and　SVM　algorithm　has　the　highest　classification　accuracy　and　shortest　running　time.