基于视觉大模型的船舶身份识别研究

Research on Ship Identity Recognition Using Large Vision Model

本文针对内河流域船舶监管中船舶识别准确性低的问题,提出了识别船舶身份的方法,并通过大量真实数据开展针对性训练,通过目标识别小模型与视觉大模型分析的有效结合,大幅提升了视频处理效率和大语言模型分析效率,节约了网络带宽与AI算力资源。研究结果表明,本文提出的方法在提高船舶身份识别准确性和效率方面具有显著优势,可以为内河船舶监管提供有效技术支持。

This article proposes a method for identifying ship identities in the supervision of ships in inland river basins,addressing the issue of low accuracy in ship recognition.Targeted training is conducted using a large amount of real data,and the effective combination of target recognition small models and visual large models analysis significantly improves video processing efficiency and large language model analysis efficiency,saving network bandwidth and AI computing resources.The research results indicate that the method proposed in this article has significant advantages in improving the accuracy and efficiency of ship identity recognition,and can provide effective technical support for inland river ship supervision.