Trawler behavior discrimination method based on multi-step clustering

Abstract

The invention discloses a trawler behavior discrimination method based on multi-step clustering. According to the method, spatial and temporal data such as speed, angle, longitude and latitude, time and the like are comprehensively considered to establish a multi-dimensional distance model, and a time sequence neighbor calculation criterion is adopted to calculate a similarity distance matrix between track points, so that the matrix calculation time is greatly reduced, and an OPTICS algorithm and a [xi]-steep automatic recognition cluster method are used to realize track division based on non-global parameters, to obtain a track segment. Then a k-means algorithm is used to realize re-clustering of the track segments based on the similarity distance between the track segments, so that classification of track points is realized. Feature extraction is performed on the classification, a trawler behavior discrimination model is established, and rapid discrimination of trawler behaviors is realized. Experiments show that the method has the advantages of low parameter sensitivity, high precision, high universality, less time consumption and the like, and can be applied to rapid judgment of mass trawler behaviors.

Description

technical field [0001] The invention belongs to the field of fishery and data mining technology, and in particular relates to a multi-step clustering-based trawler behavior discrimination method. Background technique [0002] Fishing Vessel Monitoring System (VMS) is a comprehensive information service platform based on satellite navigation system, geographic information system, Internet, mobile communication network, etc. In China, with the application and popularization of the Beidou satellite positioning system, the Beidou satellite system can obtain a track data every 10 minutes that records the speed, course, time, position, bow direction, alarm and other information of fishing boats. Provide data support for application scenarios such as fishery production, maritime monitoring, and maritime rescue. Due to technical limitations, the VMS system cannot directly obtain the real-time behavior status of fishing boats, so using the time-series trajectory data of fishing boat...

AI Technical Summary

Problems solved by technology

The advantages of the MSC-FBI algorithm are simplicity and strong interpretability, but the method still has the following problems: (1) The quality of the classification directly depends on the selection of the global variable neighborhood radius and the minimum number in the neighborhood in the DBSCAN algorithm. (2) The algorithm has shortcomings such as high time consumption and poor versatility, and is not suitable for rapid classification of batches of ships.

[0015] Although these probabilistic conversion-based methods have strong robustness and are not affected by the distribution of trajectory point attributes such as speed, they still have the following two shortcomings: (1) training these models requires a large amount of prior data; ( 2) These models are only based on the state of the current track point, and the state of the next track point is judged after inputting the probability matrix, without considering the spatio-temporal locality of the trawler's track, which will lead to discrimination errors

Images