Spatiotemporal data cleaning and repairing method and system for ship cooperative situation awareness

Abstract

The application provides a space-time data cleaning and repairing method and system for ship cooperative situation awareness, and relates to the technical field of ship data processing. The method comprises the following steps: S1, aligning non-synchronous data of multiple source heterogeneous sensors to a global discrete time grid to generate a shared time confidence mask; S2, generating a physical statistical fusion mask based on physical extreme value, fluid dynamics slip ratio constraint and statistical characteristics; S3, constructing a double-branch parallel orthogonal decoupling reconstruction model, a time encoder and a channel encoder are used to extract features in parallel, and a decoder is used to output a reconstruction value after adaptive fusion; S4, performing three-grade decision based on an adaptive dynamic threshold to output primary cleaning data; and S5, performing confidence dynamic weighted verification and correction based on spatial neighborhood relationship to obtain final situation data. The application significantly improves the reconstruction accuracy and physical consistency of ship data.

Images