Abstract: Considering the problems of traffic sign automatic detection in natural scene images,an automatic algorithm for traffic signs detection based on a vehicle-borne mobile mapping system is proposed.The proposed algorithm synthesizes such techniques as self-adaptive image segmentation,gray-value projection,shape analysis,and stereo image matching to realize automatic detecting traffic signs and computing their geometric information such as spatial positions and sizes.The new algorithm is applied to actual images of natural scenes taken by the mobile photogrammetry system in Nanjing at different times.The experimental results show that the algorithm is robust and has higher detection speed and detection accuracy.