Research Of Bus Lines And Sites Extraction Based On GPS Trajectory And Its Application

With the development of intelligent public transport and positioning technology matures,GPS positioning technology has been widely used in the transport sector.A large number of trace data collected by the onboard GPS contains a large amount of informations,which has very important research value.In this thesis,we go on bus line extraction and bus stop mining based on bus track data.Trajectory compression algorithms are used for line geometry key point extraction and clustering algorithms related are used for mining sites in this thesis,which have achieved good results.The main research work are as follows:(1)To reduce the amount of data and improve processing speed,we process the drift trajectory data and remove invalid track data.Using the angle limits law and vertical distance limits law to clean up redundant data.(2)Use Douglas-Peucker algorithm to compress track data,use vertical distance and level distance ratio to remove jitter on line.By comparison with road graph,it demonstrates that the result is good.(3)Based on mesh density clustering algorithm,we distinguish cluster of trajectory data.Combined with speed factor,more refined cluster can be got.Besides,we choose less prone to congestion periods combined with speed bound go on trajectory cluster statistical analysis.Get site clustering and cluster representative point as the bus station.Experimental results show that this way can find bus station effectively.