Lane-level steering information detection method and system based on unsupervised learning

The invention discloses a lane-level steering information detection method and system based on unsupervised learning. Comprising the following steps: a data preprocessing step: carrying out road network topology processing, track data cleaning and map matching to realize connection between track points and attribution road sections; and an intersection lane information identification step: carrying out clustering analysis on the matching tracks in a certain range of the adjacent end points of the road section intersection based on a Gaussian mixture model, and further carrying out road section intersection lane information detection. And a lane-level steering information mining step: considering different lane steering information, and designing different lane steering information identification rules based on an unsupervised classification method. According to the method, exploratory research is carried out on lane-level steering recognition in the unsupervised classification direction, and a certain technic