In　this　paper,　a　new　threshold　segmentation　algorithm　is　proposed　for　lane　line　identification.　First　of　all,　in　the　image　preprocessing　stage,　the　gray　image　is　binarized　by　threshold　algorithm　based　on　median　filter　and　Flood　Fill　algorithm.　Then,　the　binary　image　is　input　to　the　line　segment　detector　(LSD)　algorithm.　Finally,　Non-lane　markings　are　removed　through　geometric　characteristics　and　continuity　of　lane　line.　The　algorithm　has　good　robustness,　real-time　performance,　success　rate　and　accuracy.