Method and system for acquiring ship maneuvering behaviors based on ship AIS trajectory data

Abstract

The invention discloses a method and system for acquiring ship maneuvering behaviors based on ship AIS trajectory data. The method comprises retrieving single ship AIS original trajectory data corresponding to a to-be-detected ship in a ship AIS original database; preprocessing the single ship AIS original trajectory data; acquiring the uniformly-spaced single ship front section trajectory and the uniformly-spaced single ship back section trajectory corresponding to data of each ship sampling point position in a way of equal-time interval sampling; respectively calculating the vector velocity of the uniformly-spaced single ship front section trajectory and the vector velocity of the uniformly-spaced single ship back section trajectory corresponding to data of the current ship sampling point position; and confirming the ship maneuvering behavior corresponding to the data of the current ship sampling point position based on a preset ship maneuvering mode identification model library. According to the invention, the ship maneuvering behavior can be automatically identified, and a basis is provided for making a ship intelligent collision avoidance decision.

Description

technical field [0001] The invention relates to ship detection technology, in particular to a method and system for acquiring ship maneuvering behavior based on ship AIS track data. Background technique [0002] Ensuring the safety of ship navigation is an important link in ship inspection technology, so how to know the ship's maneuvering behavior in time is particularly important. At present, some foreign scholars have done some research on the recognition of ship maneuvering behavior. For example, Laxhammar (2009) used Gaussian Mixture Modal (GMM) to model the position and velocity characteristics of normal ship trajectory points; Ristic (2008) used kernel density Estimation methods model the eigenvalues ​​of normal ship trajectory points; some scholars use Bayesian networks to identify ship maneuvering behavior. [0003] However, the above methods all have defects to varying degrees. For example, although Laxhammar uses a mixed Gaussian model to model the position and ve...

AI Technical Summary

Problems solved by technology

[0003] However, the above methods all have defects to varying degrees. For example, although Laxhammar uses a mixed Gaussian model to model the position and velocity characteristics of normal ship trajectory points, and assumes that the normal moving ship track points obey the mixed Gaussian distribution, the estimated parameters of this method are too large. For example, Ristic uses the kernel density estimation method to model the eigenvalues ​​of normal ship trajectory points, but the calculation amount of this method is too large; for example, Bayesian network is used to identify ship maneuvering behavior, but this method uses conditional probability to Measuring ship maneuvering behavior requires corresponding prior knowledge, and the false alarm rate is high

Images