Ship prediction trajectory correction method based on sea and land constraints

Abstract

The invention discloses a ship prediction trajectory correction method based on sea and land constraints, and belongs to the technical field of sea moving target trend prediction. The invention provides a prediction result correction method based on a differentiated coastline buffer area in order to solve the problem that part of prediction results appear on land due to the fact that sea-land constraints are not fully considered in current various ship track prediction algorithms. The method comprises the following steps: firstly, carrying out differential buffer area calculation on common open sea areas, straits, islands, inland rivers and the like in an overwater passing environment, and generating a ship passable area under the condition of not changing an original topological relation; on the basis, analyzing a typical relation mode of a prediction track and a passable area edge in detail, designing a prediction track correction strategy for different modes, output results of various existing algorithms can be well corrected, and the method is an effective supplement for various existing ship track prediction methods.

Description

technical field [0001] The invention belongs to the technical field of forecasting the movement of a moving target at sea, in particular to a method for correcting a ship's predicted track based on land and sea constraints. Background technique [0002] Trajectory prediction is an important content in the problem of ship behavior law and trend prediction analysis. At present, the prediction methods of ship tracks are mainly divided into two categories: one is to use the navigation records of a large number of ships in history to summarize the typical navigation of ships under various typical space-time conditions through track space-time clustering and pattern extraction. track, based on which to predict the future track of the current ship; the other is to extract track spatial features, time features, ship motion features, etc., without using historical data, with the help of machine learning or deep learning methods, Completely start from the current track data to realiz...

AI Technical Summary

Problems solved by technology

Therefore, most of the above methods are currently applicable to track prediction problems in open sea areas or short-term predictions where the impact of sea-land relations is low, and it is often difficult to achieve better prediction results near rivers and islands.

Images