Visual analysis semantic label method based on navigation logic division

  Objectives  This paper aims to explore new methods for enhancing the abnormal data processing of real ships in inland rivers, improving data comprehension and assisting in ship behavior recognition research.

  Methods  By constructing a navigation logic level, the time series data is divided to obtain the semantic label of the ship behavior. A navigation logic visualization analysis system is designed on the basis of semantic labels, and the navigation status of the ship is combined with data visualization to assist in analyzing data problems and studying ship characteristics. Relying on a digital waterway, the data of working ships with complex behavior in an inland waterway is selected for example-based testing, and the system is used to analyze abnormal data and conduct research on ship behavior.

  Results  Through the interactive visualization of navigation logic, the causes and characteristics of abnormal data with position jumping can be effectively determined, thereby enhancing abnormal data processing. In addition, the qualitative analysis of features and quantitative analysis of thresholds effectively divides the berthing and direct sailing status data, further enriching the semantic labels of ship behavior.

  Conclusions   The visual analysis system designed with semantic labels of ship behavior proposed herein improves data comprehensibility through free human-computer interaction. It can enhance abnormal data analysis and processing, assist in ship behavior recognition research and provide new research tools for data analysts.