Water depth mapping method and system based on multi-source data fusion analysis

By employing a multi-source data fusion analysis method based on Transformer-LSTM and DeepBathy-UNet, the high cost and low efficiency of traditional water depth measurement methods are addressed, enabling high-frequency updates and high-precision water depth mapping, thus supporting marine management and shipping safety.

CN122156351APending Publication Date: 2026-06-05WUHAN UNIV OF TECH
View PDF 7 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
WUHAN UNIV OF TECH
Filing Date
2026-05-08
Publication Date
2026-06-05

AI Technical Summary

Technical Problem

Traditional water depth measurement methods suffer from high measurement costs, low operational efficiency, limited coverage, and difficulty in achieving high-frequency updates. They are particularly difficult to implement in vast sea areas, complex waterways, and remote islands and reefs, resulting in outdated nautical chart data and missing water depth information, which seriously restricts shipping safety and refined marine management.

Method used

A Transformer-LSTM hybrid neural network model was used to repair missing segments of ship AIS data and to perform spatiotemporal fusion matching with ship depth sounding data. Discrete water depth point sets were retrieved by combining tidal data, and a continuous water depth surface model was generated by DeepBathy-UNet deep learning interpolation model. Multi-source water depth data were introduced for fusion correction, and an accuracy evaluation system was established.

Benefits of technology

It significantly improves the quality and coverage of water depth data, reduces surveying costs, and enables rapid updates, providing high spatiotemporal resolution water depth data support for dynamic port dredging, intelligent ship stowage, low-carbon route optimization, and marine environmental monitoring.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122156351A_ABST
    Figure CN122156351A_ABST
Patent Text Reader

Abstract

The application provides a water depth mapping method and system based on multi-source data fusion analysis, and relates to the technical field of marine geographic information surveying and mapping. The steps of the method include obtaining ship AIS data and ship sounding data, and respectively performing pretreatment; a hybrid neural network model is used to repair the missing section of the ship AIS data track; the repaired ship AIS data and the ship sounding data are spatiotemporally fused and matched to obtain AIS composite water depth data; based on the AIS composite water depth data, the ship draft is corrected at the voyage section level, and combined with the tide data, a discrete water depth point set with the chart depth datum as the reference is inversely obtained; the discrete water depth point set is spatially interpolated to generate a continuous water depth surface model; multi-source water depth data is introduced to fuse and correct the continuous water depth surface model, an accuracy evaluation system is established to quantitatively evaluate the corrected continuous water depth surface model, and water depth mapping results and accuracy evaluation reports are output.
Need to check novelty before this filing date? Find Prior Art