Multimodal Recognition Methods for Maritime Vessel Identification in Complex Scenarios

Maritime vessel recognition is a crucial task in maritime monitoring and traffic management, supporting various applications such as vessel tracking, operational safety, and anomaly detection. Traditional vessel detection and recognition processes rely heavily on manual inspection, which is constrained by environmental noise, low visibility, and the need for real-time performance, making high accuracy and reliability difficult to achieve. This study develops a multimodal classification method that effectively integrates image data with vessel identification numbers to improve the accuracy of automated vessel recognition. A comprehensive recognition system, integrating vessel detection and classification, has been successfully developed and deployed in a maritime monitoring center. With the inclusion of vessel identification numbers as auxiliary data, the system achieved an accuracy of 89% in practical applications. The innovation of this research lies in the application of a multimodal model to address the challenges of vessel recognition, significantly enhancing recognition accuracy and establishing an intelligent recognition system suitable for real-world maritime monitoring and analysis.