Abstract:Recognition of ship traffic patterns can provide insights into the rules of navigation, maneuvering, and collision avoidance for ships at sea. This is essential for ensuring safe navigation at sea and improving navigational efficiency. With the popularization of the Automatic Identification System (AIS), numerous studies utilized ship trajectories to identify maritime traffic patterns. However, the current research focuses on the spatiotemporal behavioral feature clustering of ship trajectory points or segments while lacking consideration for multiple factors that influence ship behavior, such as ship static and maritime geospatial features, resulting in insufficient precision in ship traffic pattern recognition. This study proposes a ship traffic pattern recognition method that considers multi-attribute trajectory similarity (STPMTS), which considers ship static feature, dynamic feature, port geospatial feature, as well as semantic relationships between these features. First, A ship trajectory reconstruction method based on grid compression was introduced to eliminate redundant data and enhance the efficiency of trajectory similarity measurements. Subsequently, to quantify the degree of similarity of ship trajectories, a trajectory similarity measurement method is proposed that combines ship static and dynamic information with port geospatial features. Furthermore, trajectory clustering with hierarchical methods was applied based on the trajectory similarity matrix for dividing trajectories into different clusters. The quality of the similarity measurement results was evaluated by quality criterion to recognize the optimal number of ship traffic patterns. Finally, the effectiveness of the proposed method was verified using actual port ship trajectory data from the Tianjin Port of China, ranging from September to November 2016. Compared with other methods, the proposed method exhibits significant advantages in identifying traffic patterns of ships entering and leaving the port in terms of geometric features, dynamic features, and adherence to navigation rules. This study could serve as an inspiration for a comprehensive exploration of maritime transportation knowledge from multiple perspectives.

Deep Learning and Spatial Analysis Based Port Detection

Ship collision risk analysis: Modeling, visualization and prediction

Docked Ship Detection Based on Edge Line Analysis and Aggregation Channel Features

Ship feature recognition methods for deep learning in complex marine environments

Port Sensing Computation Based on Maritime Big Data

A deep learning approach for port congestion estimation and prediction

Enhancing Inland Waterway Safety and Management through Machine Learning-Based Ship Detection

Graph deep learning recognition of port ship behavior patterns from a network approach

Research on A Ship Trajectory Classification Method Based on Deep Learning

Deep Learning-Based Automatic Detection of Ships: An Experimental Study Using Satellite Images

Bridging Optical, Infrared, and Satellite Insights

An Adaptive Deep Learning Framework for Shipping Container Code Localization and Recognition

A Multiscale Ship Detection Algorithm Based on Optical Remote Sensing Image

Ship detection with deep learning: a survey

Autonomous navigation and collision prediction of port channel based on computer vision and lidar

Ship recognition method combined with image segmentation and deep learning feature extraction in video surveillance

An Intelligent Ship Image/Video Detection and Classification Method with Improved Regressive Deep Convolutional Neural Network

An approach for traffic pattern recognition integration of ship AIS data and port geospatial features

A Novel Approach in Maritime Data Completion Using Deep Learning and NLP Techniques

Ship detection of optical remote sensing image in multiple scenes

High resolution remote sensing image ship target detection technology based on deep learning