Method and system for measuring marine moving target trajectory similarity

Abstract

The invention provides a method and a system for measuring ocean moving target track similarity, and belongs to the technical field of ocean moving target track detection. The method comprises: receiving historical track data; preprocessing the historical track data to form first track data; randomly selecting one piece of unselected second track data from a preset database, the second track datacomprising n vectors; forming a plurality of sub-sequences according to the selected second track data; for each sub-sequence, calculating the Euclidean distance between the vector of the sub-sequenceand the vector of the first track data to form a cost matrix; calculating a cumulative cost matrix according to the cost matrix; obtaining DTW distances between corresponding sub-sequences of the first track data and the second track data according to a calculation result of the cumulative cost matrix; whether the DTW distance is smaller than a preset distance threshold value or not is judged; and under the condition that it is judged that the DTW distance is smaller than a distance threshold value, the calculated DTW distance and the corresponding cumulative cost matrix are associated and stored.

Description

technical field [0001] The invention relates to the technical field of trajectory detection of marine moving targets, in particular to a method and a system for measuring the similarity of the trajectory of marine moving targets. Background technique [0002] A time series is a series of observed data values ​​arranged in chronological order. At present, there have been a lot of studies on the similarity measurement of time series, among which the dynamic time warping (Dynamic Time Warping, DTW) algorithm is widely used. DTW is a nonlinear algorithm that combines distance measurement and time warping. Its main principle is as follows: first, a cost matrix is ​​constructed by calculating the distance between each component of two time series (usually using the Euclidean distance formula), and then using The dynamic programming algorithm finds an optimal path through several grid points in this matrix grid, so as to accurately calculate and measure the distance between two ti...

AI Technical Summary

Problems solved by technology

[0004] However, since the existing DTW method directly aligns the two trajectories from the starting time point for comparative analysis, in practice, the trajectories of moving targets in the ocean present diverse characteristics, which are easily affected by many factors. There will be different driving states in the segment, so the beginning of the time series corresponding to the two trajectories for similarity measurement is likely to be at different time points, and the moving target is also in different state stages. In this case, the starting time point Alignment for comparative analysis is difficult to obtain high-precision analysis results

[0005] On the other hand, due to reasons such as the moving target stops sailing or the obtained track point information is incomplete, the time series of the track to be matched may have multiple available segments and are not connected to each other. Each segment of the track presents different characteristics. May form new trajectory features

For this case, the method of calculating DTW distance in the prior art is difficult to calculate the result

Images