Density-based urban traffic flow hierarchical analysis method and device in GPS track

The invention relates to the field of GPS data processing, and provides a density-based urban traffic flow hierarchical analysis method and device in a GPS track. The problem of low clustering result precision caused by fixed parameters in a traditional clustering method is solved through a two-stage clustering method. According to the main scheme, the method comprises the following steps: defining a data structure of a GPS track, and carrying out preprocessing and track segmentation to obtain track fragments; constructing an HNSW structure according to the GPS track fragments, and defining a clustering search radius and a density threshold value in each layer; first-stage clustering ST-HDBC is implemented, first-stage clustering is implemented in the built hierarchical structure, and the GPS track fragments are clustered to obtain a primary clustering result; modeling is carried out according to the geographic features of the trajectory, and secondary combination is carried out on the GPS trajectory clusters