A marine hydro-meteorological multi-source data fusion analysis and early warning method and system

Abstract

The application discloses a kind of ocean hydrographic weather multi-source data fusion analysis early warning method and system, it is related to ocean hydrographic weather data analysis technical field, including: through the multi-source heterogeneous data intelligent processing analysis platform collection ocean current velocity, baric field distribution and so on multidimensional original data, after classified screening extraction calibration data subset, input hydrographic weather coupling collaborative analysis model completes coupling correlation operation, again through time series data dynamic correction fusion algorithm carries out data correction and fusion, will the data after processing input ocean environment multidimensional coupling early warning model and carry out risk grade quantization analysis, finally generate and push multidimensional early warning information.Coupling collaborative analysis, dynamic correction fusion and multidimensional early warning model form core technical support, the application realizes the efficient integration and depth mining of multi-source heterogeneous data, improves the systematicness and accuracy of ocean hydrographic weather warning, applicable to ocean environment safety guarantee related scene.

Description

Technical Field

[0001] This invention relates to the field of marine hydrological and meteorological data analysis technology, and in particular to a method and system for early warning through multi-source fusion analysis of marine hydrological and meteorological data. Background Technology

[0002] The stability of the marine hydro-meteorological environment directly impacts the safe operation of numerous sectors, including marine shipping, fisheries production, and marine engineering construction. Accurate multi-source data fusion analysis and early warning are crucial for mitigating marine disaster risks. With the rapid development of marine observation technology, various hydro-meteorological monitoring devices have emerged, resulting in massive amounts of heterogeneous data across multiple dimensions, including ocean currents, pressure field distribution, temperature-salinity vertical profiles, and wind shear. These data are scattered, diverse in type, and complex in spatiotemporal characteristics, making efficient integration and in-depth analysis difficult with traditional data processing methods. There is an urgent need to construct a comprehensive technical system integrating specialized models and intelligent analysis platforms to achieve precise coupling, dynamic correction, and scientific early warning of multi-source data, meeting the high requirements of marine environmental safety for data processing timeliness, correlation, and early warning reliability.

[0003] Existing marine hydrological and meteorological data processing and early warning technologies have two significant shortcomings: First, data fusion lacks a systematic collaborative mechanism. Existing technologies mostly analyze single types of data or individual environmental elements independently, failing to fully consider the coupling and correlation characteristics between hydrological and meteorological elements. The classification, screening, and cross-coupling processing of multi-source heterogeneous data are not comprehensive enough, making it difficult for the data fusion results to accurately reflect the complex dynamic changes in the marine environment. Second, early warning models lack dynamic adaptability and multi-dimensional coupling capabilities. Traditional early warning methods mostly use fixed parameter models, lacking dynamic correction mechanisms for time-series data, and have not formed a comprehensive quantitative analysis system for multi-dimensional environmental elements. They cannot effectively integrate the associated risk information of multiple elements such as current velocity, air pressure, temperature and salinity, and wind direction, resulting in insufficient pertinence and accuracy of early warning results, making it difficult to meet the precise early warning needs in complex and ever-changing marine environmental scenarios. Summary of the Invention

[0004] In order to overcome the shortcomings and deficiencies of existing technologies, this invention provides a method and system for early warning of marine hydrological and meteorological multi-source data fusion analysis.

[0005] The technical solution adopted in this invention is a method for multi-source data fusion analysis and early warning of marine hydrology and meteorology, comprising the following steps: S1 for acquiring multi-dimensional raw data; collecting multi-dimensional raw data in the field of marine hydrology and meteorology through a multi-source heterogeneous data intelligent processing and analysis platform, wherein the multi-dimensional raw data includes ocean current velocity, pressure field distribution, wave period, temperature-salinity vertical profile, wind shear and precipitation intensity correlation data; S2 for filtering to obtain a calibration data subset; classifying and filtering the collected multi-dimensional raw data using the multi-source heterogeneous data intelligent processing and analysis platform, and extracting the calibration data subset coupled with marine hydrology and meteorology; S3 for completing the coupling correlation operation of hydrological and meteorological elements; A subset of calibration data is input into the hydro-meteorological coupled collaborative analysis model, and the data coupling and correlation operation is completed through the multi-factor correlation mapping mechanism built into the model; S4 is used to perform time-series dynamic correction of the coupling results and realize multi-source fusion; the time-series data dynamic correction fusion algorithm is used to dynamically correct the coupling operation results and perform multi-source data fusion processing; S5 is used to perform multi-dimensional risk quantification; the fused data is input into the marine environment multi-dimensional coupled early warning model to perform quantitative analysis of marine hydro-meteorological risk levels; S6 is used to generate and push early warning information; multi-dimensional early warning information is generated based on the quantitative analysis results, and the early warning information is pushed through the output module of the multi-source heterogeneous data intelligent processing and analysis platform.

[0006] Furthermore, the expression for the hydro-meteorological coupled collaborative analysis model is as follows: ,in, The results are from a coupled hydrological and meteorological analysis. This refers to the number of data acquisition nodes. The first The coupling weighting coefficient between node flow velocity and temperature-salinity data; For the first Nodal ocean current velocity calibration data; For the first Nodal pressure field distribution data; For the first Vertical profile data of temperature and salinity at nodes; For the first Node wind shear data; For the first Nodal precipitation intensity correlation data; For velocity-pressure field coupling operator; For thermo-salinity-wind shear cooperation operator; This is an operator for integrating the coupling results with precipitation data.

[0007] Furthermore, the expression for the time-series data dynamic correction and fusion algorithm is as follows: ,in, For time-series dynamic correction and fusion of data; These are the start and end times of the time series data, respectively. For time-coupled data, a dynamic weighting function is used. The data is the hydrological and meteorological coupled data at time t; This is the error correction coefficient function; Let t be the error gradient vector of the coupled data; This refers to the time decay factor for time-series data. This is a spatiotemporal matching matrix for multi-source data.

[0008] Furthermore, the expression for the multi-dimensional coupled early warning model of the marine environment is: ,in, Quantified values ​​for marine environmental early warning levels; This is the amplification factor for the warning level; The number of warning dimensions; For the first Dimensional marine environmental basic state values; For the first Dimensional environmental state influence coefficient; For the first Weights based on the rate of change of dimensional states; For the first Dimensional environmental state change; This is a dynamic early warning correction coefficient; For the first Dimensional early warning response sensitivity; For the first Cumulative amount of dimensional risk; For the first Dimensional risk accumulation rate.

[0009] Furthermore, the data processing model expression of the multi-source heterogeneous data intelligent processing and analysis platform is as follows: ,in, Output results for platform data processing; The number of heterogeneous data source types; For the first Calibration data for heterogeneous data sources; A multi-source data mapping matrix; Operators for classifying and filtering heterogeneous data; Optimize the processing coefficients for the data; Enhance operators for calibration data.

[0010] Furthermore, the comprehensive evaluation model expression for the multi-source marine hydrological and meteorological data fusion analysis and early warning is as follows: ,in, This is a comprehensive evaluation value for early warning based on multi-source data fusion analysis; To assess the number of indicators; For the first Weight of each evaluation indicator; For the first The indicators integrate data values; For the first Coupling coefficient of the item index; For the first The time decay factor of the item indicator; For the first Spatial influence coefficient of the indicator; The benchmark coefficient is used for comprehensive evaluation; To dynamically evaluate the correction factor; For the first Real-time monitoring values ​​of the indicators; For the first Rate of change of each indicator.

[0011] Further, step S3 includes the following sub-steps: S31, based on the calibration data subset output by the multi-source heterogeneous data intelligent processing and analysis platform, extract the marine hydro-meteorological calibration influencing factors corresponding to each data, and establish a factor-data association index table; S32, divide the data in the association index table into hydro-meteorological element types such as flow velocity, air pressure, temperature and salinity, wind direction, and precipitation, and determine the coupling priority of each type of subset; S33, input each type of subset into the hydro-meteorological coupling collaborative analysis model according to the coupling priority order, and complete the cross-coupling operation between different types of data through the multi-factor association mapping mechanism built into the model to generate preliminary coupling results; S34, verify the data association of the preliminary coupling results, remove data items with coupling association degree lower than the set threshold, and retain effective coupling data that conforms to the coupling law of marine hydro-meteorology.

[0012] Further, step S4 includes the following sub-steps: S41, acquiring the effective coupled data output from S3, extracting the temporal feature parameters of the data, establishing a time-series data sequence library, and clarifying the timestamp distribution pattern of the data; S42, based on the time-series data dynamic correction and fusion algorithm, setting a dynamic correction window according to the timestamp distribution pattern, and performing time-period error detection and identification on the data in the time-series data sequence library; S43, using the algorithm's built-in dynamic correction mechanism to specifically correct the identified error data, and simultaneously employing a multi-source data fusion strategy to perform spatiotemporal fusion on the corrected data from different time periods; S44, performing temporal consistency verification on the fused data to ensure the continuity and correlation of the data in the time dimension, and generating time-series corrected fused data.

[0013] Further, S5 includes the following sub-steps: S51, splitting the time-series corrected fusion data generated in S4 according to the marine environmental early warning dimensions to obtain the corresponding dimensional data for current velocity early warning, air pressure early warning, temperature and salinity early warning, wind direction early warning, and precipitation early warning; S52, inputting the data of each dimension into the marine environmental multi-dimensional coupled early warning model, and performing preliminary risk level quantification on the data of each dimension through the model's dimensional quantification module; S53, activating the model's multi-dimensional coupling mechanism, performing cross-correlation calculations on the preliminary quantification results of each dimension, and fusing the risk information of each dimension to generate comprehensive risk quantification data; S54, based on the comprehensive risk quantification data, referring to the marine hydrological and meteorological early warning level classification standard, determining the final risk level quantification result.

[0014] A multi-source marine hydrological and meteorological data fusion analysis and early warning system is disclosed. This system is applied to a multi-source marine hydrological and meteorological data fusion analysis and early warning method, comprising: a multi-source heterogeneous marine hydrological and meteorological data acquisition unit, used to collect data on ocean current velocity, pressure field distribution, wave cycle, vertical profile of temperature and salinity, wind shear, and precipitation intensity correlation, and establishing a bidirectional data transmission connection with a multi-source heterogeneous data intelligent processing and analysis unit; and a multi-source heterogeneous data intelligent processing and analysis unit, which receives the data transmitted by the acquisition unit, classifies and filters it, extracts a calibrated data subset, and establishes data output connections with a hydrological and meteorological coupled collaborative analysis unit and a time-series data dynamic correction and fusion unit, respectively. The hydro-meteorological coupled collaborative analysis unit receives a subset of calibration data and completes coupling operations through a built-in model. Its output is connected to the input of the time-series data dynamic correction and fusion unit. The time-series data dynamic correction and fusion unit performs dynamic correction and multi-source fusion on the coupling operation results. Its output is connected to the marine environment multi-dimensional coupled early warning unit. The marine environment multi-dimensional coupled early warning unit receives the fused data and performs risk level quantitative analysis. Its output is connected to the multi-dimensional early warning information push unit. The multi-dimensional early warning information push unit receives the quantitative analysis results and generates multi-dimensional early warning information, which is then pushed to the target terminal through a preset transmission channel.

[0015] Beneficial Effects: This invention proposes a method and system for early warning based on multi-source marine hydrological and meteorological data fusion analysis. By constructing an intelligent processing and analysis platform for multi-source heterogeneous data, it connects a coupled and collaborative analysis model for hydrological and meteorological data, a dynamic correction and fusion algorithm for time-series data, and a multi-dimensional coupled early warning model for the marine environment, forming a complete technical system. Addressing the lack of a systematic collaborative mechanism in traditional data fusion, this invention classifies and filters multi-dimensional raw data and extracts calibrated data subsets. It utilizes a multi-factor correlation mapping mechanism to achieve deep coupling calculations of hydrological and meteorological elements, fully exploring the correlation characteristics between various data points, breaking the limitations of independent analysis of single data, and accurately reflecting the complex dynamic changes in the marine environment. Facing the shortcomings of traditional early warning models in terms of dynamic adaptability and multi-dimensional coupling capabilities, the invention uses a dynamic correction and fusion algorithm for time-series data to correct errors and perform spatiotemporal fusion on the coupling results, ensuring the continuity and accuracy of the data's temporal sequence. Finally, the multi-dimensional coupled early warning model for the marine environment integrates multi-factor risk information to complete a comprehensive quantitative analysis, replacing fixed-parameter models and improving the targeting and adaptability of early warnings. The system's various units form a closed-loop collaboration through bidirectional or unidirectional data transmission, seamlessly connecting data acquisition, processing, coupling, correction to early warning push. This enables efficient integration and in-depth mining of multi-source heterogeneous data, significantly improving the scientific nature and reliability of marine hydrological and meteorological early warnings, and fully meeting the needs of marine environmental safety assurance. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart illustrating the overall method steps provided in this embodiment of the invention; Figure 2 A flowchart of method step S3 provided in an embodiment of the present invention; Figure 3 A flowchart of method step S4 provided in an embodiment of the present invention; Figure 4 A flowchart of method step S5 provided in an embodiment of the present invention; Figure 5 This is a system unit composition diagram provided for an embodiment of the present invention. Detailed Implementation

[0018] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0019] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0020] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0021] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0022] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0023] like Figure 1 As shown, a method for early warning based on multi-source marine hydrological and meteorological data fusion analysis includes the following steps: S1. Multi-dimensional raw data in the field of marine hydrology and meteorology are collected through a multi-source heterogeneous data intelligent processing and analysis platform. The multi-dimensional raw data includes ocean current velocity, pressure field distribution, wave cycle, temperature and salinity vertical profile, wind shear and precipitation intensity correlation data. Specifically, step S1 involves collecting multi-dimensional raw marine hydrological and meteorological data through a multi-source heterogeneous data intelligent processing and analysis platform. The data collection covers six core categories: ocean current velocity, pressure field distribution, wave cycle, vertical temperature and salinity profile, wind shear, and precipitation intensity correlation data. During the data collection process, the platform utilizes 30-50 fixed monitoring nodes and 10-15 mobile monitoring devices distributed across different sea areas to acquire continuous data at a frequency of once every 10 minutes. Ocean current velocity data collection includes stratified monitoring at depths of 0-50 meters; pressure field distribution data covers a 100×100 km grid monitoring area; wave cycle data focuses on an effective period range of 1-20 seconds; vertical temperature and salinity profile data is collected at 2-meter intervals at depths of 0-100 meters; wind shear data records horizontal and vertical wind direction changes within a height range of 0-30 meters; and precipitation intensity correlation data simultaneously collects information such as precipitation time periods and durations within the monitoring area. During the data collection process, the platform receives and stores various types of data in real time to ensure the integrity and timeliness of the data, providing comprehensive and continuous raw data support for subsequent data processing. The collected data is uniformly formatted as a standardized binary format, which facilitates subsequent classification and filtering operations.

[0024] S2 utilizes a multi-source heterogeneous data intelligent processing and analysis platform to classify and filter the collected multi-dimensional raw data, and extract a subset of calibration data coupled with marine hydrology and meteorology. Specifically, step S2 utilizes a multi-source heterogeneous data intelligent processing and analysis platform to classify and filter the collected multi-dimensional raw data. The core objective is to extract a subset of calibration data coupled with marine hydrological and meteorological data. In practice, the platform first categorizes the raw data into six types based on data type: current velocity dataset, pressure field dataset, wave cycle dataset, temperature-salinity vertical profile dataset, wind direction shear dataset, and precipitation intensity correlation dataset. Then, the built-in correlation analysis module calculates the correlation coefficients between each type of data and marine hydrological and meteorological data. These correlation coefficients can be calculated using Pearson correlation coefficient, mutual information, covariance-based coupling correlation coefficient, or equivalent alternatives, specifically implemented by the platform's correlation analysis module. A correlation coefficient threshold of 0.75 is set, and data with correlation coefficients higher than this threshold are selected as candidate calibration data. Subsequently, the candidate calibration data underwent data integrity verification, removing data records with a missing rate exceeding 5%, as well as data exceeding reasonable value ranges. Specifically, the reasonable ranges for ocean current velocity were set at 0-5 m / s, pressure field distribution at 950-1050 hPa, wave period at 1-20 seconds, temperature at -2-30°C and salinity at 30-38‰ in the vertical temperature-salinity profile, wind shear at 0-360 degrees, and duration of precipitation intensity in the correlation data at 0-72 hours. Except for the explicit weighting coefficients for current velocity and temperature-salinity data, the weights of elements such as pressure field, wave period, wind shear, and precipitation can be reflected through the correlation mapping matrix or internal parameters of the coupling / coordination operator. The coupling operation results can be a data structure containing the coupling strength coefficients of each element and a calibrated multidimensional feature vector / time series. Through the above multi-layered screening process, a subset of calibration data that accurately reflects the coupling characteristics of marine hydrology and meteorology is finally extracted, providing high-quality data input for subsequent coupling operations.

[0025] S3 inputs a subset of calibration data into the hydro-meteorological coupled collaborative analysis model, and completes the data coupling and correlation operation through the multi-factor correlation mapping mechanism built into the model. Specifically, step S3 inputs the selected calibration data subset into the hydrological and meteorological coupled collaborative analysis model, and completes the data coupling and correlation calculation through the model's built-in multi-factor correlation mapping mechanism. During implementation, firstly, various types of data in the calibration data subset are aligned according to the acquisition node and timestamp to ensure consistency of different data types in the spatiotemporal dimension. The timestamp alignment accuracy is controlled within 1 minute, and the acquisition node matching error does not exceed 1 kilometer. Subsequently, the model calls the built-in multi-factor correlation mapping mechanism, which pre-sets coupling and correlation rules between ocean current velocity, pressure field distribution, wave cycle, temperature-salinity vertical profile, wind shear, and precipitation intensity data. These rules include negative correlation rules between current velocity and pressure field, positive correlation rules between temperature-salinity data and wave cycle, and collaborative correlation rules between wind shear and precipitation intensity. Based on these association rules, the model performs cross-coupling operations on various types of calibration data. During the operation, differentiated coupling weights are assigned to data from different acquisition nodes, with weight values ​​ranging from 0.1 to 0.9. Nearshore monitoring node data has a higher weight than offshore monitoring node data, and real-time monitoring data has a higher weight than historical backtracking data. Through multiple rounds of iterative operations, the model integrates the coupling association information of various data types, generating coupling association operation results that comprehensively reflect the interactions of hydrological and meteorological elements. The operation results include key information such as coupling strength, association trends, and cooperative change characteristics, providing basic data for subsequent dynamic correction and fusion.

[0026] S4, a time-series data dynamic correction and fusion algorithm is used to dynamically correct the coupled operation results and perform multi-source data fusion processing; Specifically, step S4 employs a time-series data dynamic correction and fusion algorithm to dynamically correct the coupled operation results and perform multi-source data fusion processing. In practice, the time-series characteristics of the coupled operation results are first extracted, including data change cycles, fluctuation amplitudes, and trend slopes. Based on these time-series characteristics, a time-series data sequence is constructed, with the sequence length determined by the collection duration, ranging from a minimum of 24 hours to a maximum of 7 days. Subsequently, the algorithm sets a dynamic correction window, with the window size adaptively adjusted according to the data change frequency. When the data fluctuation frequency is high, the window size is set to 1 hour; when the fluctuation frequency is low, the window size is set to 6 hours. Error detection is performed on the time-series data by sliding the window, with detection indicators including data deviation, abrupt change coefficient, and trend consistency. The data deviation threshold is set to 5%, and the abrupt change coefficient threshold is set to 0.3. Data exceeding these thresholds is considered abnormal data. For the detected abnormal data, the algorithm corrects it through a built-in dynamic correction mechanism. Correction methods include interpolation correction based on historical data from the same period, smoothing correction based on adjacent time period data, and logical correction based on coupling association rules. After correcting the abnormal data, the algorithm uses a spatiotemporal fusion strategy to fuse the corrected data from different collection nodes and time periods. During the fusion process, the spatiotemporal weights of the data are comprehensively considered. The spatial weights are set according to the coverage and importance of the monitoring nodes, and the temporal weights are set according to the timeliness of the data. After fusion, fused data with continuous time sequence and spatiotemporal consistency is generated, with a data resolution of 10 minutes / time and 1 km × 1 km, ensuring the integrity and correlation of the data in the time and spatial dimensions.

[0027] S5 inputs the fused data into the multi-dimensional coupled early warning model of the marine environment to conduct quantitative analysis of marine hydrological and meteorological risk levels; Specifically, step S5 inputs the fused data into the multi-dimensional coupled early warning model of the marine environment for quantitative analysis of marine hydro-meteorological risk levels. During implementation, the fused data is first split according to early warning dimensions, into six major early warning dimensions: current velocity risk, pressure field risk, wave cycle risk, temperature-salinity vertical profile risk, wind direction shear risk, and precipitation intensity risk. Subsequently, the model quantifies the data for each early warning dimension individually. During quantification, referring to the marine hydro-meteorological risk level classification standards, the data for each dimension is converted into a quantitative score of 0-10, where 0-3 indicates low risk, 3-6 indicates medium risk, and 6-10 indicates high risk. The quantification process considers the rate of change and cumulative effect of the data for each dimension, with the rate of change weight set at 0.4 and the cumulative effect weight set at 0.6. After completing single-dimensional quantification, the model initiates a multi-dimensional coupling mechanism, integrating the quantification results from each dimension through weighted summation and collaborative computation. The weight allocation is determined based on the marine environmental characteristics of different sea areas. In nearshore waters, the weights for current velocity risk and precipitation intensity risk are each set to 0.2, while in offshore waters, the weights for wave cycle risk and pressure field risk are each set to 0.25. The remaining weights for other dimensions are distributed proportionally. Through multi-dimensional coupling computation, the model generates a comprehensive risk level quantification value, also ranging from 0 to 10. Simultaneously, it outputs auxiliary information such as the risk contribution ratio of each dimension and risk change trends, providing a quantitative basis for generating early warning information. The accuracy of the quantitative analysis is controlled above 90%, and the risk level misjudgment rate does not exceed 8%.

[0028] S6 generates multi-dimensional early warning information based on quantitative analysis results, and pushes the early warning information through the output module of the multi-source heterogeneous data intelligent processing and analysis platform.

[0029] Specifically, step S6 generates multi-dimensional early warning information based on the quantitative analysis results, and pushes the early warning information through the output module of the multi-source heterogeneous data intelligent processing and analysis platform. During implementation, the early warning level is first determined based on the comprehensive risk level quantification value: 0-3 points correspond to a blue early warning, 3-6 points to a yellow early warning, and 6-10 points to an orange early warning. Simultaneously, the main risk factors are identified based on the risk contribution ratio of each dimension; for example, if the flow velocity risk ratio exceeds 40%, it is marked as a flow velocity-dominated early warning. Subsequently, multi-dimensional early warning information is generated, including core elements such as the early warning level, main risk factors, risk impact range, risk duration, and trend prediction. The risk impact range is precisely marked using latitude and longitude coordinates with an error not exceeding 5 kilometers, and the risk duration prediction accuracy is controlled within 1 hour. The early warning information format combines standardized text format with visual chart format. The text format facilitates terminal reception and parsing, while the visual chart format includes risk distribution heatmaps and trend change curves for intuitive display of the risk situation. Ocean wave period data can be incorporated as sea state characteristics into the vertical temperature-salinity profile data or as internal parameters of coupled / cooperative operators. When needed, it can also be expanded into a separate ocean wave period term for computation. Finally, the platform output module pushes early warning information through preset transmission channels, including satellite communication, wireless communication, and wired network communication. Satellite communication is used for early warning pushes in far-sea areas, wireless communication is used for pushes to mobile terminals in near-shore areas, and wired network communication is used for pushes between coastal monitoring stations and command centers. The push delay does not exceed 3 minutes, ensuring that early warning information can be delivered to target terminals in a timely manner, providing timely guidance for marine environmental safety protection.

[0030] Preferably, the expression of the hydro-meteorological coupled collaborative analysis model is: ,in, The results are from a coupled hydrological and meteorological analysis. This refers to the number of data acquisition nodes. The first The coupling weighting coefficient between node flow velocity and temperature-salinity data; For the first Nodal ocean current velocity calibration data; For the first Nodal pressure field distribution data; For the first Vertical profile data of temperature and salinity at nodes; For the first Node wind shear data; For the first Nodal precipitation intensity correlation data; For velocity-pressure field coupling operator; For thermo-salinity-wind shear cooperation operator; This is an operator for integrating the coupling results with precipitation data.

[0031] Specifically, the hydro-meteorological coupled collaborative analysis model is used to achieve deep correlation calculations between multiple marine hydrological and meteorological elements. Its implementation utilizes a subset of calibration data output from a multi-source heterogeneous data intelligent processing and analysis platform. During model operation, the number of data acquisition nodes is first determined, ranging from 30 to 50 based on the monitoring area. Each node corresponds to a complete set of multi-dimensional data acquisition units. For each node's data related to ocean current velocity, pressure field distribution, vertical profile of temperature and salinity, wind shear, and precipitation intensity, corresponding coupling weight coefficients are assigned. The weight coefficient for current velocity data ranges from 0.3 to 0.9, and the weight coefficient for temperature and salinity data ranges from 0.2 to 0.7. These weight coefficients are dynamically adjusted based on the importance of the sea area where the node is located and the reliability of the data; the weight coefficient for key nearshore monitoring nodes is higher than that for ordinary offshore nodes. Subsequently, a specific coupling operator was used to perform cross-correlation calculations between flow velocity and pressure field data. This operator, designed based on the coupling principles of fluid mechanics and meteorology, can accurately capture the negative correlation between the two. A collaborative operator was used to achieve the collaborative fusion of temperature-salinity vertical profile and wind shear data, highlighting their combined impact on the marine environment. Finally, a correlation integration operator was used to integrate the above two types of calculation results with precipitation intensity correlation data, forming a complete hydro-meteorological coupled collaborative analysis result. During the model calculation process, each iteration cycle was set to 5 minutes, with a total of 8-12 iterations to ensure that the calculation results fully reflect the dynamic coupling relationship between various elements, providing accurate coupling basis data for subsequent data processing.

[0032] Preferably, the expression for the time-series data dynamic correction and fusion algorithm is: ,in, For time-series dynamic correction and fusion of data; These are the start and end times of the time series data, respectively. For time-coupled data, a dynamic weighting function is used. The data is the hydrological and meteorological coupled data at time t; This is the error correction coefficient function; Let t be the error gradient vector of the coupled data; This refers to the time decay factor for time-series data. This is a spatiotemporal matching matrix for multi-source data.

[0033] Specifically, the time-series data dynamic correction and fusion algorithm is used to optimize and fuse hydrological and meteorological coupled calculation results over time. During implementation, the time span of the time-series data is first defined, with the start and end times set according to actual monitoring needs, covering a minimum of 24 hours and a maximum of 7 days. In the initial stage of algorithm operation, a dynamic weight function and an error correction coefficient function are constructed. The former is set according to the timeliness of the data, with a function value of 0.7-0.9 for real-time data and 0.1-0.3 for historical data; the latter is dynamically adjusted based on historical statistical results of data errors, with a function value of 0.6-0.8 for periods of high error incidence and 0.2-0.4 for periods of stable data. Subsequently, the error gradient vector of the coupled data is calculated. This vector is determined by the ratio of the difference between adjacent data points to the time interval, used to accurately locate the trend and magnitude of data error changes. Simultaneously, a time decay factor is set, calculated based on the interval between the data acquisition time and the current time. The factor value is 1.0 within a 1-hour interval, decreasing by 0.1 for each additional hour, down to a minimum of 0.3, thus highlighting the influence weight of recent data. The algorithm accumulates the product of coupled data, error gradient vector, and time decay factor through integral calculation, and then correlates it with a spatiotemporal matching matrix. This matrix is ​​constructed with a spatial resolution of 1 km × 1 km and a temporal resolution of 10 minutes, including correlation strength information of data at different spatiotemporal locations. During algorithm implementation, the dynamic correction window adaptively adjusts according to the data fluctuation frequency. When the fluctuation frequency is higher than 3 times per hour, the window is set to 1 hour; when it is lower than 1 time per hour, the window is set to 6 hours, ensuring the targeted nature of error detection and correction. The final output is fused data with continuous time sequence and controllable error. The dynamic correction window is an algorithm running parameter that can be configured by the platform or adaptively determined based on the data fluctuation frequency; it is not an explicit variable in the formula.

[0034] Preferably, the expression for the multi-dimensional coupled early warning model of the marine environment is: ,in, Quantified values ​​for marine environmental early warning levels; This is the amplification factor for the warning level; The number of warning dimensions; For the first Dimensional marine environmental basic state values; For the first Dimensional environmental state influence coefficient; For the first Weights based on the rate of change of dimensional states; For the first Dimensional environmental state change; This is a dynamic early warning correction coefficient; For the first Dimensional early warning response sensitivity; For the first Cumulative amount of dimensional risk; For the first Dimensional risk accumulation rate.

[0035] Specifically, the multi-dimensional coupled early warning model for the marine environment is used for the comprehensive quantification of marine hydro-meteorological risk levels. During implementation, the number of early warning dimensions is first determined, divided into six core dimensions based on key influencing factors of the marine environment, covering key indicators such as current velocity, pressure field, wave cycle, vertical profile of temperature and salinity, wind shear, and precipitation intensity. During model operation, an early warning level amplification factor is first set, ranging from 1.2 to 1.8, adjusted according to the risk tolerance of different sea areas. For densely populated nearshore areas and port areas, the value is 1.6 to 1.8, while for offshore operation areas, it is 1.2 to 1.4. For each early warning dimension, a basic marine environmental state value is determined, calculated by weighting historical average data and real-time monitoring data, with real-time data having a weighting of 0.6 to 0.8. An environmental state influence coefficient is then set, ranging from 0.5 to 1.5, assigned according to the degree of influence of each factor on the marine environment. The coefficients corresponding to current velocity and wave cycle are 1.3 to 1.5, and the coefficients corresponding to temperature and salinity data are 0.5 to 0.8. Simultaneously, the weights of the rate of change of state and the change in environmental state are set, with the former ranging from 0.3 to 0.7 and the latter calculated by the difference between the current data and the historical average. The dynamic early warning correction coefficient ranges from 0.8 to 1.2, dynamically adjusted according to the recent early warning accuracy. The early warning response sensitivity ranges from 1.0 to 1.5, with higher values ​​for high-risk areas than for low-risk areas. The basic quantitative results of each dimension are integrated through product and square root operations, and then combined with the square root result of the sum of squares of the risk accumulation rate to finally generate a comprehensive early warning level quantitative value. During model calculation, the quantitative results are updated every 10 minutes to ensure the real-time nature of the early warning. The quantitative values ​​accurately correspond to different risk levels, providing a core basis for generating early warning information. In this embodiment, the number of early warning dimensions m is preferably 6, corresponding to six dimensions: flow velocity, pressure field, wave cycle, vertical profile of temperature and salinity, wind shear, and precipitation intensity. The quantitative values ​​of each dimension are represented by Sj (or equivalent variables) and fused in the model through summation / square root methods.

[0036] Preferably, the data processing model expression of the multi-source heterogeneous data intelligent processing and analysis platform is: ,in, Output results for platform data processing; The number of heterogeneous data source types; For the first Calibration data for heterogeneous data sources; A multi-source data mapping matrix; Operators for classifying and filtering heterogeneous data; Optimize the processing coefficients for the data; Enhance operators for calibration data.

[0037] Specifically, the data processing model of the multi-source heterogeneous data intelligent processing and analysis platform is used to achieve efficient integration and optimized processing of multiple types of heterogeneous data. During implementation, the number of heterogeneous data source types is first determined, categorized into 6-8 types based on data acquisition equipment and methods, including fixed monitoring stations, mobile monitoring equipment, satellite remote sensing, radar detection, and other sources. During model execution, the calibration data from various heterogeneous data sources are first aggregated using a union operation to integrate the raw data from all sources, constructing a comprehensive dataset. Subsequently, a multi-source data mapping matrix is ​​introduced. This matrix is ​​constructed according to the correspondence between data types and marine hydrological and meteorological elements, with dimensions equal to the number of heterogeneous data types multiplied by the number of core elements. Matrix element values ​​range from 0.1 to 0.9, representing the correlation strength between data and elements. Preliminary data classification and filtering are achieved through matrix multiplication, selecting effective data with a correlation higher than 0.7 with the core elements. Simultaneously, intersection operations are performed on various types of data to extract common core information from data from different sources, assigning data optimization processing coefficients with values ​​ranging from 1.1 to 1.5 to strengthen the weight of core information. A calibration data enhancement operator is introduced, based on algorithms such as data smoothing, noise filtering, and detail enhancement, which can improve the accuracy and usability of the data. The results of union and intersection processing are integrated through addition operations to finally generate the platform's data processing output. During model implementation, data processing latency is controlled within 3 minutes, the integrity of processed data reaches over 98%, and data accuracy is improved by 30%-50%, providing high-quality data support for subsequent coupling analysis, correction fusion, and early warning quantification. The platform's data processing output refers to the calibration / fusion data results after the platform completes classification, screening, mapping, and enhancement of multi-source heterogeneous data, which is used as input for subsequent coupling analysis and early warning quantification; the generation and push of early warning information are completed in step S6.

[0038] Preferably, the comprehensive evaluation model expression for the multi-source marine hydrological and meteorological data fusion analysis and early warning is as follows: ,in, This is a comprehensive evaluation value for early warning based on multi-source data fusion analysis; To assess the number of indicators; For the first Weight of each evaluation indicator; For the first The indicators integrate data values; For the first Coupling coefficient of the item index; For the first The time decay factor of the item indicator; For the first Spatial influence coefficient of the indicator; The benchmark coefficient is used for comprehensive evaluation; To dynamically evaluate the correction factor; For the first Real-time monitoring values ​​of the indicators; For the first Rate of change of each indicator.

[0039] Specifically, a comprehensive evaluation model for multi-source marine hydrological and meteorological data fusion analysis and early warning is used to achieve a comprehensive evaluation and optimization of early warning effectiveness. During implementation, the number of evaluation indicators is first determined, divided into 8-10 indicators based on core early warning needs, including key dimensions such as data fusion accuracy, coupling correlation accuracy, time series correction effectiveness, risk quantification accuracy, and early warning response timeliness. During model operation, the weights of each evaluation indicator are set, ranging from 0.08 to 0.15. Specifically, the weights corresponding to early warning accuracy and response timeliness are 0.13-0.15, and the weight corresponding to data fusion accuracy is 0.10-0.12. The fused data values ​​are directly obtained from the results of multi-source data fusion. The coupling coefficient ranges from 0.6 to 0.9, set according to the correlation strength between each indicator and the early warning effectiveness. Simultaneously, a time decay factor and a spatial influence coefficient are set. The former ranges from 0.7 to 0.95, with the value decreasing as the data collection time increases. The latter ranges from 0.8 to 1.2, with higher values ​​for nearshore areas than for offshore areas. The comprehensive evaluation benchmark coefficient ranges from 1.0 to 1.3, determined based on industry standards and historical data. The dynamic evaluation correction factor ranges from 0.9 to 1.1, dynamically adjusted according to the deviation between recent evaluation results and actual conditions. The basic evaluation results of each indicator are integrated through summation and division operations, and then combined with the product of the indicator change rates to finally generate a comprehensive evaluation value for multi-source data fusion analysis and early warning. During model implementation, the evaluation results are updated hourly, and the consistency between the evaluation value and the early warning effect reaches over 90%, accurately reflecting the operating status of the early warning system and providing a scientific basis for system optimization and adjustment. The comprehensive evaluation model is an optional operational evaluation / feedback module, which can evaluate the system's operating effect after completing the early warning output (e.g., after step S6) and adjust parameters such as weights and thresholds based on the evaluation results; it does not affect the implementation of the main process described in claim 1.

[0040] Preferred, such as Figure 2As shown, step S3 includes the following sub-steps: S31, based on the calibration data subset output by the multi-source heterogeneous data intelligent processing and analysis platform, extract the marine hydrological and meteorological calibration influencing factors corresponding to each data, and establish a factor-data association index table; S32, divide the data in the association index table into hydrological and meteorological element types into flow velocity, air pressure, temperature and salinity, wind direction, and precipitation subsets, and determine the coupling priority of each subset; S33, input each subset into the hydrological and meteorological coupling collaborative analysis model according to the coupling priority order, and complete the cross-coupling operation between different types of data through the multi-factor association mapping mechanism built into the model to generate preliminary coupling results; S34, verify the data association of the preliminary coupling results, remove data items with coupling association degree lower than the set threshold, and retain effective coupling data that conforms to the coupling law of marine hydrological and meteorological data.

[0041] Specifically, step S3 achieves precise implementation of hydro-meteorological coupled collaborative analysis through four sub-steps. S31 first utilizes the calibration data subset output by the multi-source heterogeneous data intelligent processing and analysis platform to comprehensively extract key marine hydro-meteorological influencing factors corresponding to each data point. These factors include core parameters such as ocean current intensity, pressure gradient, vertical temperature and salinity variation rate, wind shear angle, and precipitation intensity amplitude. Based on these factors, a factor-data association index table is established. The index table includes key information such as factor number, data source, corresponding monitoring node, and collection timestamp. The query response time of the index table is controlled within 0.5 seconds to ensure efficient subsequent computation. S32 precisely divides the data in the association index table into hydro-meteorological element types: current velocity, pressure, temperature and salinity, wind direction, and precipitation. The coupling priority is determined based on the influence weight of each subset on the marine environment coupling analysis. The current velocity and pressure subsets are set to priority level one, and the temperature and salinity and wind direction subsets are set to priority level two. The precipitation category is set as Level 3, and the priority division is determined by a combination of expert scoring and historical data verification. S33 inputs various subsets of data into the hydro-meteorological coupled collaborative analysis model according to the coupling priority order of Level 1, Level 2, and Level 3. Through the model's built-in multi-factor correlation mapping mechanism, based on the coupling laws of fluid mechanics and meteorology, cross-coupling operations between different types of data are completed. During the operation, the participation of each subset of data is allocated according to priority: Level 1 datasets account for 40%, Level 2 for 35%, and Level 3 for 25%, generating preliminary coupling results. S34 uses the correlation threshold judgment method to verify the data correlation of the preliminary coupling results, setting the correlation threshold to 0.8. By calculating the coupling correlation coefficient between different data items, data items with a coupling correlation coefficient lower than this threshold are eliminated, retaining effective coupled data that conforms to the coupling laws of marine hydro-meteorology. The effective data ratio is not less than 92%, providing high-quality coupled data support for subsequent time-series correction and fusion.

[0042] Preferred, such as Figure 3 As shown, S4 specifically includes the following sub-steps: S41, acquiring the effective coupled data output from S3, extracting the temporal feature parameters of the data, establishing a time-series data sequence library, and clarifying the timestamp distribution pattern of the data; S42, based on the time-series data dynamic correction and fusion algorithm, setting a dynamic correction window according to the timestamp distribution pattern, and performing time-period error detection and identification on the data in the time-series data sequence library; S43, using the algorithm's built-in dynamic correction mechanism to specifically correct the identified error data, and simultaneously employing a multi-source data fusion strategy to perform spatiotemporal fusion on the corrected data from different time periods; S44, performing temporal consistency verification on the fused data to ensure the continuity and correlation of the data in the time dimension, and generating time-series corrected fused data.

[0043] Specifically, step S4 implements a systematic approach to dynamic correction and fusion of time-series data through four sub-steps. S41 first acquires the effective coupled data output from step S3, comprehensively extracting the time-series characteristic parameters of the data. These parameters include data time intervals, change cycles, fluctuation amplitudes, and trend slopes. Based on these parameters, a time-series data sequence library is established. The sequence library is divided into sub-libraries according to the monitoring area, with each sub-library containing continuous coupled data for the corresponding area within 24 hours. The data storage format adopts standardized binary to ensure compatibility between data reading and processing. S42, based on the dynamic correction and fusion algorithm for time-series data, adaptively sets the dynamic correction window according to the data time interval and fluctuation frequency. When the data time interval is 10 minutes and the fluctuation frequency is higher than 3 times per hour, the correction window size is set to 1 hour; when the data time interval is 30 minutes and the fluctuation frequency is lower than 1 time per hour, the correction window size is set to 6 hours. The sliding window is used to perform time-by-time error detection and identification on the data in the time-series data sequence library. The criteria include the degree of data deviation from the mean, the magnitude of abrupt changes, and trend consistency. The false detection rate during the detection process is controlled within 3%. S43 uses the algorithm's built-in dynamic correction mechanism to specifically correct the identified error data. For random errors, a moving average correction method is used; for systematic errors, a historical data comparison correction method is used; and for abrupt changes, a coupled correlation logic correction method is used. The error of the corrected data is controlled within 5%. At the same time, a spatiotemporal fusion strategy is used to fuse the corrected data from different time periods in a spatiotemporal dimension. During the fusion process, spatial weights are allocated according to the coverage of monitoring nodes, and temporal weights are allocated according to the timeliness of the data. S44 performs a time-series consistency verification on the fused data. By calculating the difference between data from adjacent time periods and the trend fit, the continuity and correlation of the data in the time dimension are verified. The difference threshold is set at 0.1, and the trend fit threshold is set at 0.9 to ensure that the data that passes the verification meets the time-series consistency requirements. Finally, time-series corrected fusion data with continuous time series and controllable error is generated.

[0044] Preferred, such as Figure 4As shown, S5 specifically includes the following sub-steps: S51, the time-series corrected fusion data generated in S4 is split according to the marine environmental early warning dimensions to obtain the corresponding dimensional data for current velocity early warning, air pressure early warning, temperature and salinity early warning, wind direction early warning, and precipitation early warning; S52, the data of each dimension are input into the marine environmental multi-dimensional coupled early warning model, and the risk level of each dimension data is initially quantified by the model's dimensional quantification module; S53, the multi-dimensional coupling mechanism of the model is activated to perform cross-correlation calculations on the preliminary quantification results of each dimension, and the risk information of each dimension is fused to generate comprehensive risk quantification data; S54, based on the comprehensive risk quantification data and referring to the marine hydrological and meteorological early warning level classification standard, the final risk level quantification result is determined.

[0045] Specifically, step S5 achieves precise quantification of multi-dimensional coupled early warning of the marine environment through four sub-steps. S51 precisely splits the time-series corrected and fused data generated in step S4 according to the dimensions of marine environmental early warning. Based on the impact type of marine hydrological and meteorological disasters, corresponding dimensional data for current velocity warning, air pressure warning, temperature and salinity warning, wind direction warning, and precipitation warning are obtained. Each dimension includes key information such as basic values, rate of change, and cumulative amount. The accuracy of data splitting reaches 99%, ensuring no overlap or omission of data across dimensions. S52 inputs the data of each dimension into the corresponding module of the multi-dimensional coupled early warning model of the marine environment. The model's dimension quantification module performs preliminary risk level quantification of each dimension. The quantification process refers to the marine hydrological and meteorological risk level classification standard, converting each dimension into a quantification score of 0-10. During quantification, parameters such as the ratio of real-time values ​​to historical extreme values ​​and the ratio of rate of change to safety thresholds are considered. The quantification time for each dimension does not exceed [a certain threshold]. 1 minute; S53 initiates the multi-dimensional coupling mechanism of the model. Based on the correlation characteristics of data from each dimension, it uses a combination of weighted summation and collaborative computation to perform cross-correlation calculations on the preliminary quantitative results of each dimension. The weight allocation is determined according to the contribution of each dimension to marine disasters. The weight of the current velocity warning dimension is set to 0.25, the air pressure warning dimension to 0.2, the temperature and salinity warning dimension to 0.15, the wind direction warning dimension to 0.2, and the precipitation warning dimension to 0.2. The risk information from each dimension is integrated to generate comprehensive risk quantitative data. S54, based on the comprehensive risk quantitative data, strictly refers to the national marine hydrological and meteorological warning level classification standard, and divides the quantitative values ​​into three levels: low risk (0-3), medium risk (3-6), and high risk (6-10). The final risk level quantitative result is determined, and auxiliary information such as the risk contribution ratio of each dimension and the risk development trend is output. The update cycle of the quantitative results is consistent with the data collection cycle to ensure the real-time performance and accuracy of the warning.

[0046] like Figure 5The system describes a multi-source marine hydrological and meteorological data fusion analysis and early warning system. This system is applied to a multi-source marine hydrological and meteorological data fusion analysis and early warning method, comprising: a multi-source heterogeneous marine hydrological and meteorological data acquisition unit, used to collect data related to ocean current velocity, pressure field distribution, wave period, temperature-salinity vertical profile, wind shear, and precipitation intensity, and establishing a bidirectional data transmission connection with a multi-source heterogeneous data intelligent processing and analysis unit; and a multi-source heterogeneous data intelligent processing and analysis unit, which receives the data transmitted by the acquisition unit, classifies and filters it, extracts a calibration data subset, and establishes data output connections with a hydrological and meteorological coupled collaborative analysis unit and a time-series data dynamic correction and fusion unit, respectively. The system consists of: a hydro-meteorological coupled collaborative analysis unit, which receives a subset of calibration data and performs coupled calculations using a built-in model; its output is connected to the input of a time-series data dynamic correction and fusion unit; a time-series data dynamic correction and fusion unit, which performs dynamic correction and multi-source fusion on the coupled calculation results; and a marine environment multi-dimensional coupled early warning unit, which receives the fused data and performs risk level quantitative analysis; and a multi-dimensional early warning information push unit, which receives the quantitative analysis results and generates multi-dimensional early warning information, pushing the early warning information to the target terminal through a preset transmission channel.

[0047] The formulas in this invention enable unified calculation of different scalar and vector parameters, constructing a coupling mechanism based on physical correlation and dimensional adaptation. By pre-setting targeted operators and weight allocations, the essential differences between scalars and vectors are resolved. Taking the hydro-meteorological coupled collaborative analysis model as an example, the direction and magnitude information of vector parameters such as ocean current velocity and pressure field distribution are transformed into a quantitative form compatible with scalar parameters such as temperature-salinity vertical profiles and precipitation intensity through coupling operators. The design of the operators strictly follows the inherent laws among marine hydro-meteorological elements, such as the negative correlation between current velocity and pressure field, and the synergistic effect of temperature-salinity data and wind shear, so that the directional characteristics of vectors are transformed into correlation strength coefficients, and the magnitude characteristics are transformed into quantitative contribution values. At the same time, by assigning differentiated weights to different types of parameters, such as giving higher weights to vector current velocity data from nearshore monitoring nodes than to scalar pressure data from offshore areas, the spatial characteristics of vector parameters are preserved while achieving unified integration with scalar parameters at the numerical level. This ensures that various parameters participate in the calculation according to their actual degree of influence, rather than being simply superimposed.

[0048] Furthermore, the formula constructs a dynamic coordination system for scalar and vector parameters by introducing dynamic adaptation factors and correlation mapping matrices, further ensuring the rationality of unified calculation. For example, in the dynamic correction and fusion algorithm for time-series data, the error gradient vector of vector data is correlated with the time decay factor of scalar data through the error correction coefficient function. The direction information of the gradient vector is transformed into the direction guide for error correction, and the magnitude information is transformed into the correction amplitude. Then, the spatiotemporal alignment with other scalar time-series data is achieved through the spatiotemporal matching matrix. In the multi-dimensional coupled early warning model of the marine environment, the rate of change of vector wind shear and the cumulative amount of scalar precipitation intensity are adjusted through dynamic early warning correction coefficients and early warning response sensitivity to form a unified risk quantification dimension. The spatial variation characteristics of vector parameters and the numerical variation characteristics of scalar parameters are jointly transformed into risk level quantification values. This design respects both the numerical attributes of scalar parameters and the spatiotemporal attributes of vector parameters, and through a multi-layered adaptation mechanism, enables various parameters to form an organic whole in the formula, accurately reflecting the coupled correlation of multiple marine hydrological and meteorological elements, and achieving efficient unified calculation of cross-type parameters.

[0049] A method and system for multi-source marine hydrological and meteorological data fusion analysis and early warning is proposed. This system constructs a multi-model collaborative and closed-loop technical framework, achieving efficient processing and accurate early warning of multi-source heterogeneous data through deep integration of a specialized platform and algorithms. The system utilizes a multi-source heterogeneous data intelligent processing and analysis platform to collect and classify multi-dimensional raw data, accurately extracting and calibrating data subsets to provide high-quality data support for subsequent analysis. A hydrological and meteorological coupled collaborative analysis model is used to uncover deep correlations between various elements. Combined with a time-series data dynamic correction and fusion algorithm, dynamic optimization and spatiotemporal fusion of data are achieved. Finally, a multi-dimensional coupled early warning model of the marine environment completes multi-dimensional risk quantification, forming a complete chain technical architecture from data input to early warning output. This significantly improves the systematic nature of data processing and the scientific rigor of early warning results.

[0050] This method and system address the lack of collaborative mechanisms in traditional data fusion. By classifying and filtering multi-dimensional data and establishing a multi-factor correlation mapping mechanism, it breaks through the limitations of independent analysis of single elements, fully integrates the coupling and correlation characteristics of multiple types of data such as flow velocity, air pressure, temperature, and salinity, and accurately reproduces the complex dynamic changes of the marine environment. Addressing the insufficient dynamic adaptability and multi-dimensional coupling capabilities of traditional early warning models, it corrects time-series data errors through a dynamic correction mechanism, ensuring data continuity and accuracy. Furthermore, it utilizes a multi-dimensional coupled early warning model to integrate risk information from various dimensions, replacing fixed-parameter models, significantly improving the targeting and adaptability of early warnings, and fully meeting the needs of marine environmental safety assurance for precise early warning.

[0051] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0052] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0053] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0054] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0055] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0056] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0057] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0058] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0059] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for early warning based on multi-source marine hydrological and meteorological data fusion analysis, characterized in that, Includes the following steps: S1. Multi-dimensional raw data in the field of marine hydrology and meteorology are collected through a multi-source heterogeneous data intelligent processing and analysis platform. The multi-dimensional raw data includes ocean current velocity, pressure field distribution, wave cycle, temperature and salinity vertical profile, wind shear and precipitation intensity correlation data. S2 utilizes a multi-source heterogeneous data intelligent processing and analysis platform to classify and filter the collected multi-dimensional raw data, and extract a subset of calibration data coupled with marine hydrology and meteorology. S3 inputs a subset of calibration data into the hydro-meteorological coupled collaborative analysis model, and completes the data coupling and correlation operation through the multi-factor correlation mapping mechanism built into the model. S4, a time-series data dynamic correction and fusion algorithm is used to dynamically correct the coupled operation results and perform multi-source data fusion processing; S5 inputs the fused data into the multi-dimensional coupled early warning model of the marine environment to conduct quantitative analysis of marine hydrological and meteorological risk levels; S6 generates multi-dimensional early warning information based on quantitative analysis results, and pushes the early warning information through the output module of the multi-source heterogeneous data intelligent processing and analysis platform.

2. The method for early warning based on multi-source marine hydrological and meteorological data fusion analysis according to claim 1, characterized in that, The expression for the hydro-meteorological coupled collaborative analysis model is: ,in, The results are from a coupled hydrological and meteorological analysis. This refers to the number of data acquisition nodes. The first The coupling weighting coefficient between node flow velocity and temperature-salinity data; For the first Nodal ocean current velocity calibration data; For the first Nodal pressure field distribution data; For the first Vertical profile data of temperature and salinity at nodes; For the first Node wind shear data; For the first Nodal precipitation intensity correlation data; For velocity-pressure field coupling operator; For thermo-salinity-wind shear cooperation operator; This is an operator for integrating the coupling results with precipitation data.

3. The method for early warning based on multi-source marine hydrological and meteorological data fusion analysis according to claim 1, characterized in that, The expression for the time-series data dynamic correction and fusion algorithm is: ,in, For time-series dynamic correction and fusion of data; These are the start and end times of the time series data, respectively. For time-coupled data, a dynamic weighting function is used. The data is the hydrological and meteorological coupled data at time t; This is the error correction coefficient function; Let t be the error gradient vector of the coupled data; This refers to the time decay factor for time-series data. This is a spatiotemporal matching matrix for multi-source data.

4. The method for early warning based on multi-source marine hydrological and meteorological data fusion analysis according to claim 1, characterized in that, The expression for the multi-dimensional coupled early warning model of the marine environment is: ,in, Quantified values ​​for marine environmental early warning levels; This is the amplification factor for the warning level; The number of warning dimensions; For the first Dimensional marine environmental basic state values; For the first Dimensional environmental state influence coefficient; For the first Weights based on the rate of change of dimensional states; For the first Dimensional environmental state change; This is a dynamic early warning correction coefficient; For the first Dimensional early warning response sensitivity; For the first Cumulative amount of dimensional risk; For the first Dimensional risk accumulation rate.

5. The method for early warning based on multi-source marine hydrological and meteorological data fusion analysis according to claim 1, characterized in that, The data processing model expression of the multi-source heterogeneous data intelligent processing and analysis platform is: ,in, Output results for platform data processing; The number of heterogeneous data source types; For the first Calibration data for heterogeneous data sources; A multi-source data mapping matrix; Operators for classifying and filtering heterogeneous data; Optimize the processing coefficients for the data; To enhance operators for calibration data.

6. The method for early warning based on multi-source marine hydrological and meteorological data fusion analysis according to claim 1, characterized in that, The comprehensive evaluation model expression for the marine hydrological and meteorological multi-source data fusion analysis and early warning is as follows: ,in, This is a comprehensive evaluation value for early warning based on multi-source data fusion analysis; To assess the number of indicators; For the first Weight of each evaluation indicator; For the first The indicators integrate data values; For the first Coupling coefficient of the item index; For the first The time decay factor of the item indicator; For the first Spatial influence coefficient of the indicator; The benchmark coefficient is used for comprehensive evaluation; To dynamically evaluate the correction factor; For the first Real-time monitoring values ​​of the indicators; For the first Rate of change of each indicator.

7. The method for early warning based on multi-source marine hydrological and meteorological data fusion analysis according to claim 1, characterized in that, S3 includes the following steps: S31, based on the calibration data subset output by the multi-source heterogeneous data intelligent processing and analysis platform, extract the marine hydro-meteorological calibration influencing factors corresponding to each data, and establish a factor-data association index table; S32, divide the data in the association index table into hydro-meteorological element types such as flow velocity, air pressure, temperature and salinity, wind direction, and precipitation, and determine the coupling priority of each type of subset; S33, input each type of subset into the hydro-meteorological coupling collaborative analysis model according to the coupling priority order, and complete the cross-coupling operation between different types of data through the multi-factor association mapping mechanism built into the model to generate preliminary coupling results; S34, verify the data association of the preliminary coupling results, remove data items with coupling association degree lower than the set threshold, and retain effective coupling data that conforms to the coupling law of marine hydro-meteorology.

8. The method for early warning based on multi-source marine hydrological and meteorological data fusion analysis according to claim 1, characterized in that, S4 specifically includes the following sub-steps: S41, acquiring the effective coupled data output from S3, extracting the temporal feature parameters of the data, establishing a time-series data sequence library, and clarifying the timestamp distribution pattern of the data; S42, based on the time-series data dynamic correction and fusion algorithm, setting a dynamic correction window according to the timestamp distribution pattern, and performing time-period error detection and identification on the data in the time-series data sequence library; S43, using the algorithm's built-in dynamic correction mechanism to specifically correct the identified error data, and simultaneously employing a multi-source data fusion strategy to perform spatiotemporal fusion on the corrected data from different time periods; S44, performing temporal consistency verification on the fused data to ensure the continuity and correlation of the data in the time dimension, and generating time-series corrected fused data.

9. The method for multi-source fusion analysis and early warning of marine hydrological and meteorological data according to claim 1, characterized in that, S5 specifically includes the following steps: S51, splitting the time-series corrected fusion data generated in S4 according to the marine environmental early warning dimensions to obtain the corresponding dimensional data for current velocity early warning, air pressure early warning, temperature and salinity early warning, wind direction early warning, and precipitation early warning; S52, inputting the data of each dimension into the marine environmental multi-dimensional coupled early warning model, and performing preliminary risk level quantification on the data of each dimension through the model's dimensional quantification module; S53, activating the model's multi-dimensional coupling mechanism, performing cross-correlation calculations on the preliminary quantification results of each dimension, and fusing the risk information of each dimension to generate comprehensive risk quantification data; S54, based on the comprehensive risk quantification data, referring to the marine hydrological and meteorological early warning level classification standards, determining the final risk level quantification result.

10. A marine hydrological and meteorological multi-source data fusion analysis and early warning system, characterized in that, This system is applied to the marine hydrological and meteorological multi-source data fusion analysis and early warning method described in claim 1, comprising: a multi-source heterogeneous marine hydrological and meteorological data acquisition unit, used to acquire ocean current velocity, pressure field distribution, wave period, temperature-salinity vertical profile, wind shear, and precipitation intensity correlation data, and establish a bidirectional data transmission connection with a multi-source heterogeneous data intelligent processing and analysis unit; a multi-source heterogeneous data intelligent processing and analysis unit, which receives the data transmitted by the acquisition unit, classifies and filters it, extracts a calibration data subset, and establishes data output connections with a hydrological and meteorological coupled collaborative analysis unit and a time-series data dynamic correction and fusion unit, respectively; the hydrological and meteorological coupled... The collaborative analysis unit receives a subset of calibration data and performs coupling operations through a built-in model. Its output is connected to the input of the time-series data dynamic correction and fusion unit. The time-series data dynamic correction and fusion unit performs dynamic correction and multi-source fusion on the coupling operation results. Its output is connected to the marine environment multi-dimensional coupling early warning unit. The marine environment multi-dimensional coupling early warning unit receives the fused data and performs risk level quantitative analysis. Its output is connected to the multi-dimensional early warning information push unit. The multi-dimensional early warning information push unit receives the quantitative analysis results and generates multi-dimensional early warning information, which is then pushed to the target terminal through a preset transmission channel.

Images