Abstract:
The maritime channel is of great significance to maritime situation awareness. However, the existing channel extraction methods have some problems, for example, the channel waypoints are easily to be lost, and the extraction accuracy is not high. According to the principle that the number of ship's waypoints equals the number of islands, ports and other geographical areas, this study proposed an adaptive extraction method of maritime lanes based on a combination of geographic information and AIS(Automatic Identification System) data. Firstly, constructing the AIS track data set of round-trip ships between ports. Each AIS track in the data set was segmented at equal intervals, and the average speed of each track was calculated. The initial threshold of speed difference was set, and the number of times that the average speed difference of adjacent segments exceeding the threshold was calculated as the initial number of turns of ships. Secondly, counting the number of AIS tracks passing through islands, ports and other geographical areas. If the number of turns was not equal to the number of turns extracted by AIS tracks, it should adaptively adjust the speed difference threshold according to the difference between the two, re-extract the ship's waypoints so as to the number of AIS tracks was equal to the number of geographical areas, and then take the connection point of the track segment where the average speed difference exceeds the final threshold as the ship's waypoint. Then, the density-based spatial clustering of applications with noise was used to cluster the waypoints of all ships into a waypoint set, and the centroid of the waypoint set was calculated as the waypoint of the channel. Finally, taking the route waypoints as the vertices and the connecting lines between the waypoints as the edges, the digraph was constructed, The final route extraction result was obtained after deleting the connecting lines on the land. The AIS data of 20 ships traveling between the Weifang port and the Lianyungang port was used to carry out the maritime channel extraction experiment. The results showed that compared with the channel extraction method based on traffic path anomaly detection, the average false detection rate of ship's waypoints obtained by this method is reduced by 9.1%, and the average missed detection rate is reduced by 16.7%, which significantly improve the accuracy of channel waypoint location extraction.