According　to　the　color　and　geometric　properties　of　traffic　signs　in　China,　an　efficient　algorithm　of　the　segmentation　and　detection　of　yellow　traffic　signs　in　natural　scenes　is　proposed　in　this　paper.　Firstly,　the　image　segmentation　algorithm　based　on　RGB　model　is　improved　and　applied　on　segmenting　yellow　traffic　signs　in　natural　scenes.　Then,　the　detection　algorithm　based　on　the　shape　of　the　traffic　sign　is　used　to　locate　yellow　traffic　sign.　The　basis　of　the　detection　algorithm　is　to　calculate　component　differences　of　the　horizontal　and　vertical　projection.　The　experimental　results　demonstrate　that　this　efficient　algorithm　can　achieve　better　identify　results　for　yellow　traffic　signs　in　natural　scenes,　and　it　is　robust　and　has　broad　applicability.　©　2009　IEEE.