CN-121977642-A - Real-time monitoring system and method for ocean currents

CN121977642ACN 121977642 ACN121977642 ACN 121977642ACN-121977642-A

Abstract

The invention relates to the technical field of ocean monitoring, in particular to a real-time ocean current monitoring system and method for ocean currents, comprising a data acquisition module for acquiring historical ocean current data, real-time ocean current data and real-time environmental parameters of a target sea area; the model accuracy analysis module is used for determining whether the prediction accuracy of the standard ocean current monitoring model on ocean current is up to standard or not based on the ocean current offset, determining whether the prediction accuracy of the standard ocean current monitoring model is influenced by the environmental parameters of a target ocean area based on the environmental interference coefficient to obtain an environmental influence analysis module for correcting the ocean current monitoring model, determining whether the interference resistance of the corrected ocean current monitoring model is up to standard based on the ocean current parameter fluctuation value to determine a model fluctuation assessment module for adjusting regularization intensity, and determining whether the prediction accuracy of the corrected ocean current monitoring model is up to standard based on the deviation characteristic value to determine the model accuracy analysis module for optimizing the corrected ocean current monitoring model. The invention improves the accuracy of real-time monitoring of ocean currents.

Inventors

WANG SHANWU
ZHANG YUE
WANG XIAO
SU ZHAOJUN
YIN LINLIN
CUI YUE
WEI YANLIANG
ZHANG KAI

Assignees

滨州市海洋发展研究院

Dates

Publication Date: 20260505
Application Date: 20260121

Claims (10)

1. A real-time monitoring system for ocean currents, comprising: the data acquisition module is used for acquiring historical ocean current data, real-time ocean current data and real-time environmental parameters of a target sea area; The ocean current model construction module is used for carrying out feature extraction and labeling on the historical ocean current data to construct a sample data set so as to train an initial monitoring model by using the sample data set to obtain a reference ocean current monitoring model; The ocean current prediction module is used for inputting real-time ocean current data into the reference ocean current monitoring model to obtain predicted ocean current parameters, determining ocean current offset based on the predicted ocean current parameters, and determining whether the accuracy of the reference ocean current monitoring model for ocean current prediction meets the standard or not; the environment influence analysis module is used for determining an environment interference coefficient based on the real-time environment parameter so as to determine whether the target sea area environment parameter affects the prediction accuracy of the reference ocean current monitoring model, and determining to embed the real-time environment parameter as a weight duty ratio into the reference ocean current monitoring model according to the difference value between the environment interference coefficient and the preset environment interference coefficient so as to obtain a corrected ocean current monitoring model; The model fluctuation evaluation module is used for determining a current parameter fluctuation value based on the corrected current monitoring model so as to determine whether the anti-interference performance of the corrected current monitoring model meets the standard, and determining and adjusting the regularization strength of the corrected current monitoring model according to the relative difference between the current parameter fluctuation value and the preset current parameter fluctuation value; The model precision analysis module is used for determining a deviation characteristic value based on the corrected ocean current monitoring model so as to determine whether the prediction precision of the corrected ocean current monitoring model meets the standard, and determining the training period or the learning rate of the optimized corrected ocean current monitoring model according to the ratio of the deviation characterization parameter to the preset deviation characterization parameter.
2. The real-time monitoring system for ocean currents according to claim 1, wherein the ocean current prediction module determines that the accuracy of the reference ocean current monitoring model for ocean current prediction does not reach the standard based on a comparison result that the ocean current deviation is greater than a preset ocean current deviation.
3. The real-time ocean current monitoring system according to claim 2, wherein the environmental impact analysis module determines that the target sea area environmental parameter affects the prediction accuracy of the reference ocean current monitoring model based on a comparison result that the environmental interference coefficient is greater than a preset environmental interference coefficient.
4. The real-time ocean current monitoring system according to claim 3 wherein the environmental impact analysis module determines to embed the real-time environmental parameter as a weight ratio into the reference ocean current monitoring model based on a comparison result of a difference value of the environmental interference coefficient and a preset difference value under a condition that it is determined that the target ocean current environmental parameter affects the prediction accuracy of the reference ocean current monitoring model, wherein, Determining to embed real-time environment parameters into an input layer of a reference ocean current monitoring model in a first weight ratio based on a comparison result that the difference is smaller than or equal to the preset difference; And determining to embed the real-time environment parameters into the input layer of the reference ocean current monitoring model at a second weight ratio based on the comparison result that the difference is larger than the preset difference.
5. The real-time ocean current monitoring system according to claim 4, wherein the model fluctuation evaluation module determines that the interference immunity of the corrected ocean current monitoring model does not reach the standard based on a comparison result that the ocean current parameter fluctuation value is greater than a preset ocean current parameter fluctuation value.
6. The real-time ocean current monitoring system according to claim 5, wherein the model fluctuation assessment module determines to adjust regularized strength of the corrected ocean current monitoring model based on a comparison result of a relative difference of the ocean current parameter fluctuation value and a preset relative difference under a condition that the anti-interference performance of the corrected ocean current monitoring model does not reach the standard, wherein, Determining to increase the regularization strength of the corrected ocean current monitoring model by a first preset strength adjustment coefficient based on a comparison result that the relative difference is smaller than or equal to a preset relative difference; And determining to increase the regularization strength of the corrected ocean current monitoring model by a second preset strength adjustment coefficient based on the comparison result that the relative difference is larger than the preset relative difference.
7. The real-time ocean current monitoring system according to claim 6, wherein the model accuracy analysis module determines that the prediction accuracy of the corrected ocean current monitoring model does not reach the standard based on a comparison result that the deviation characterization parameter is greater than a preset deviation characterization parameter.
8. The real-time ocean current monitoring system according to claim 7, wherein the model accuracy analysis module determines to increase the training period of the corrected ocean current monitoring model by a preset period adjustment coefficient based on a comparison result that a ratio of the deviation characterization parameter to the preset deviation characterization parameter is smaller than or equal to a preset ratio under a condition that it is determined that the prediction accuracy of the corrected ocean current monitoring model does not reach the standard.
9. The real-time ocean current monitoring system according to claim 8, wherein the model accuracy analysis module determines to reduce the learning rate of the corrected ocean current monitoring model by a preset learning rate adjustment coefficient based on a comparison result that a ratio of the deviation characterization parameter to the preset deviation characterization parameter is greater than a preset ratio under a condition that it is determined that the prediction accuracy of the corrected ocean current monitoring model does not reach the standard.
10. A real-time monitoring method for ocean currents, employing the real-time monitoring system for ocean currents according to any one of claims 1 to 9, comprising: acquiring historical ocean current data, real-time ocean current data and real-time environmental parameters of a target sea area; performing feature extraction and labeling on the historical ocean current data to construct a sample data set, and training an initial monitoring model by using the sample data set to obtain a reference ocean current monitoring model; Inputting real-time ocean current data into the reference ocean current monitoring model to obtain predicted ocean current parameters, determining ocean current offset based on the predicted ocean current parameters, and determining whether the accuracy of ocean current prediction by the reference ocean current monitoring model meets the standards or not; determining an environment interference coefficient based on the real-time environment parameter to determine whether the target sea area environment parameter affects the prediction accuracy of the reference ocean current monitoring model, and determining to embed the real-time environment parameter as a weight ratio into the reference ocean current monitoring model according to the difference value between the environment interference coefficient and a preset environment interference coefficient to obtain a corrected ocean current monitoring model; Determining a current parameter fluctuation value based on the corrected current monitoring model to determine whether the anti-interference performance of the corrected current monitoring model meets the standard, and determining and adjusting regularization strength of the corrected current monitoring model according to the relative difference between the current parameter fluctuation value and a preset current parameter fluctuation value; and determining a deviation characteristic value based on the corrected ocean current monitoring model to determine whether the prediction precision of the corrected ocean current monitoring model meets the standard, and determining the training period or learning rate of the optimized corrected ocean current monitoring model according to the ratio of the deviation characterization parameter to the preset deviation characterization parameter.

Description

Real-time monitoring system and method for ocean currents Technical Field The invention relates to the technical field of ocean monitoring, in particular to a real-time monitoring system and method for ocean currents. Background Real-time and accurate ocean current monitoring has important significance in the fields of ocean scientific research, shipping safety, fishery resource management, pollutant diffusion tracking, climate prediction, military oceanography and the like, and traditional ocean current monitoring mainly depends on field observation means such as investigation ships, anchor buoys, seabed bases, acoustic Doppler flow profilers and the like, and the methods can obtain more accurate data, but have inherent limitations such as high cost, limited coverage range, difficulty in realizing large-area synchronous real-time observation and the like. With the development of satellite remote sensing technology and numerical simulation technology, the ocean current monitoring capability is remarkably improved, and satellite altimeters, scatterometers and the like can provide large-scale sea surface height and wind field data so as to calculate surface ocean current information. With the advent of artificial intelligence technology, and in particular the successful application of deep learning in time series prediction and spatial data analysis, a number of historical data-driven artificial intelligence ocean current prediction models have emerged. The models can rapidly realize prediction of ocean currents by learning complex nonlinear relations in historical ocean current data, and a new technical path is provided for real-time monitoring. CN118175454a discloses a real-time marine monitoring system and method based on AIOT technology, which comprises an information acquisition unit for acquiring monitoring information of the current time of a target sea area, wherein the monitoring information at least comprises ocean current speed, submarine topography data and sea water density data in the target sea area, a serialization unit for serializing the monitoring information to generate monitoring sequence data, compressing the monitoring sequence data and uploading the compressed monitoring sequence data to a satellite system, and a ground processing unit for receiving the compressed monitoring sequence data sent by the satellite system and inputting the monitoring sequence data into a trained monitoring model so as to output early warning information through the monitoring model. However, the prior art has the following problems that the influence of natural environment factors on the real-time ocean current monitoring accuracy is low, so that the real-time ocean current monitoring accuracy is low. Disclosure of Invention Therefore, the invention provides a real-time monitoring system and a real-time monitoring method for ocean currents, which are used for solving the problems that the accuracy of the real-time monitoring of the ocean currents is low due to the fact that the influence of natural environment factors on the real-time monitoring accuracy of the ocean currents is lacked in the prior art. To achieve the above object, the present invention provides a real-time monitoring system for ocean currents, comprising: the data acquisition module is used for acquiring historical ocean current data, real-time ocean current data and real-time environmental parameters of a target sea area; The ocean current model construction module is used for carrying out feature extraction and labeling on the historical ocean current data to construct a sample data set so as to train an initial monitoring model by using the sample data set to obtain a reference ocean current monitoring model; The ocean current prediction module is used for inputting real-time ocean current data into the reference ocean current monitoring model to obtain predicted ocean current parameters, determining ocean current offset based on the predicted ocean current parameters, and determining whether the accuracy of the reference ocean current monitoring model for ocean current prediction meets the standard or not; the environment influence analysis module is used for determining an environment interference coefficient based on the real-time environment parameter so as to determine whether the target sea area environment parameter affects the prediction accuracy of the reference ocean current monitoring model, and determining to embed the real-time environment parameter as a weight duty ratio into the reference ocean current monitoring model according to the difference value between the environment interference coefficient and the preset environment interference coefficient so as to obtain a corrected ocean current monitoring model; The model fluctuation evaluation module is used for determining a current parameter fluctuation value based on the corrected current monitoring model so as to determine whether the anti-interference performance of t