Intelligent Fusion and Optimization Processing Method for Marine Physical Data

Abstract

The present invention relates to the technical field of marine environmental monitoring, and specifically relates to an intelligent fusion and optimization processing method for marine physical data, including the following steps: Real-time collection of marine physical data through a variety of sensors, including temperature, salinity, flow velocity, tide, and acoustic wave characteristic data, and the use of a deep residual-graph convolutional hybrid model for spatio-temporal feature extraction and spatial correlation analysis. Using the quantum annealing optimization algorithm to optimize the parameters of the multi-source fusion data set to generate an optimal parameter set to improve the fusion accuracy of the data. Identifying abnormal data nodes through an adaptive threshold detection algorithm to further ensure the accuracy and reliability of the data. Finally, constructing a three-dimensional visualization marine environment model according to the optimized parameter set and realizing the dynamic mapping relationship between the model parameters and physical characteristics to achieve real-time update and intuitive display. The present invention effectively improves the processing accuracy of marine environmental data, provides accurate marine environmental prediction, and has broad application prospects.

Description

Technical Field

[0001] The present invention relates to the technical field of marine environment monitoring, and particularly to an intelligent fusion and optimization processing method for marine physical data. Background Art

[0002] The changes in the marine environment are affected by various factors, including physical characteristics such as temperature, salinity, flow velocity, tides, and acoustic wave propagation. To obtain accurate marine environment data, modern marine research relies on the collaborative work of multi-source sensors, such as temperature sensors, salinity sensors, current meters, tide gauges, and sonar systems. These sensors can collect various physical data in the ocean in real time.

[0003] Currently, some methods have been proposed for the fusion and calibration of multi-source marine data, but there are still many technical problems. First, most existing methods are limited to the processing of a single data source and lack an efficient algorithm for multi-source data fusion. Especially in the case of a large amount of data with complex sources, the existing algorithms are difficult to ensure the accuracy and consistency of data fusion. Second, the calibration problem of spatio-temporal data has not been effectively solved, especially how to eliminate the spatio-temporal deviation between sensors and ensure the accuracy of each data point in the time and space dimensions. Although some methods use deep learning and machine learning for feature extraction, in practical applications, the training and optimization of the model still face high computational costs and low accuracy problems. Third, the existing anomaly detection methods often use fixed thresholds, lacking adaptability and flexibility, resulting in relatively serious false alarm and missed alarm phenomena. Summary of the Invention

[0004] The present invention provides an intelligent fusion and optimization processing method for marine physical data.

[0005] The intelligent fusion and optimization processing method for marine physical data includes the following steps:

[0006] S1: Perform standardized preprocessing on the marine physical data collected by multi-source sensors to generate a standardized matrix containing temperature, salinity, flow velocity, tide, and acoustic wave characteristics;

[0007] S2: Construct a spatio-temporal calibration matrix based on a dynamic sliding window, perform dynamic spatio-temporal baseline calibration on the standardized matrix, and generate a spatio-temporal calibration matrix;

[0008] S3: Input the spatio-temporal calibration matrix into a deep residual-graph convolution hybrid model for feature fusion, and output a multi-source fusion data set;

[0009] S4: Construct a multi-dimensional anomaly feature correlation matrix based on the multi-source fusion data set, and use an adaptive threshold detection algorithm to identify abnormal data nodes;

[0010] S5: Use the quantum annealing optimization algorithm to perform parameter optimization iterations on the multi-source fusion data set to generate an optimized parameter set;

[0011] S6: Construct a three-dimensional visualization marine environment model according to the optimized parameter set, and establish a dynamic mapping relationship between the model parameters and physical characteristics.

[0012] Optionally, the S1 includes:

[0013] S11, Data acquisition: By setting multiple different types of marine sensors, collect marine physical data in real time, including temperature, salinity, flow velocity, tide, and acoustic wave characteristic data;

[0014] S12, Data formatting: Format the collected marine physical data to form a data set with time-space matching;

[0015] S13, Missing value filling: For sensor data with missing or abnormal values, fill the missing values by interpolation method or smoothing method based on adjacent data;

[0016] S14, Normalization processing: Perform normalization processing on the collected marine physical data;

[0017] S15, Feature extraction: Extract temporal features and spatial features from the standardized data;

[0018] S16, Construct a standardized matrix: Combine the normalized marine physical data and the extracted eigenvalue to form a standardized matrix.

[0019] Optionally, the S2 includes:

[0020] S21, Sliding window setting and data partitioning: According to the acquisition frequency and spatial resolution of the marine physical data, set the size and step length of the dynamic sliding window;

[0021] The window size is adjusted according to the time span and spatial range;

[0022] According to the set sliding window, divide the standardized matrix into multiple spatio-temporal sub-matrices in time and space dimensions. Each sub-matrix contains a data set within a preset time period and a preset spatial range;

[0023] S22, Spatio-temporal reference data selection: Among multiple spatio-temporal sub-matrices, select representative reference data as the reference benchmark for spatio-temporal calibration.

[0024] S23, Spatio-temporal deviation analysis: Analyze the spatio-temporal data deviation by comparing the data of each spatio-temporal sub-matrix with the reference data.

[0025] S24, Dynamic calibration model construction: According to the results of spatio-temporal deviation analysis, a spatio-temporal calibration model is established using statistical modeling methods;

[0026] S25, Spatio-temporal calibration calculation: Apply the spatio-temporal calibration model to the data of each spatio-temporal sub-matrix. Through model calculation, obtain the calibration parameters of each spatio-temporal window, and perform spatio-temporal calibration on the data of each sub-matrix;

[0027] S26, Spatio-temporal calibration matrix generation: Merge multiple calibrated spatio-temporal sub-matrices to generate a complete spatio-temporal calibration matrix;

[0028] S27, Calibration result evaluation and optimization: Evaluate the quality of the generated spatio-temporal calibration matrix.

[0029] Optionally, S3 includes:

[0030] S31, Deep residual network module construction and feature extraction: Construct a deep residual network module, and extract the temporal features in the spatio-temporal calibration matrix through the residual learning mechanism;

[0031] S32, Graph convolutional network module construction and spatial feature extraction: Construct a graph convolutional network module, and use graph convolutional methods to capture the spatial features in the spatio-temporal calibration matrix.

[0032] S33, Weighted fusion and multi-source fusion dataset generation: Perform weighted fusion on the temporal features and spatial features extracted by the deep residual network module and the graph convolutional network module respectively to form a comprehensive feature representation.

[0033] The comprehensive feature representation formed after fusion will be output as a multi-source fusion dataset, which contains the temporal and spatial features fused from multiple sensors.

[0034] S34, Fusion result evaluation and optimization: Evaluate the quality of the output multi-source fusion dataset, and the evaluation metrics include data accuracy, stability, and fusion effect.

[0035] Optionally, S4 includes:

[0036] S41, Construct a multi-dimensional abnormal feature correlation matrix: According to multiple features in the multi-source fusion dataset, construct a multi-dimensional abnormal feature correlation matrix;

[0037] S42, Adopt an adaptive threshold detection algorithm: Adopt an adaptive threshold detection algorithm to analyze the data in the multi-dimensional abnormal feature correlation matrix and identify abnormal data nodes.

[0038] S43, Abnormal data node identification and output: After being processed by the adaptive threshold detection algorithm, identify all abnormal data nodes and output these abnormal data nodes;

[0039] S44, Abnormal Detection Result Evaluation and Optimization: Evaluate the recognition results of abnormal data.

[0040] Optionally, the S5 includes:

[0041] S51, Define the Optimization Objective Function: Determine the optimization objective function, which is used to measure the performance of each parameter during the optimization process;

[0042] S52, Quantum Annealing Algorithm Initialization: Use the quantum annealing algorithm for parameter optimization, initialize the parameter set, and represent the initial state of each parameter on qubits. The goal of quantum annealing is to gradually search for the optimal parameter combination through the evolution of qubit states;

[0043] S53, Quantum Annealing Evolution Process: The evolution process of quantum annealing performs annealing operations on qubit states, continuously reducing the "temperature" of the system and gradually reaching the optimization goal;

[0044] S54, Measurement and Parameter Update of Qubits: After the quantum system undergoes the annealing process, at each moment when the temperature drops to the lowest, measure the qubits to obtain the optimized parameter set.

[0045] Optionally, the S6 includes:

[0046] S61, Input and Initialization of the Optimized Parameter Set: Use the generated optimized parameter set as the basic parameters for constructing the three-dimensional visualized ocean environment model, and initialize the basic information required for the ocean environment model. The basic information includes the boundary range of the ocean area and the meshing parameters;

[0047] S62, Three-Dimensional Grid Construction: Construct a three-dimensional grid based on the boundary information of the ocean area;

[0048] S63, Dynamic Mapping of Physical Characteristics and Model Parameters: Assign corresponding physical characteristic values to each grid node according to the optimized parameter set;

[0049] S64, Construction of the Three-Dimensional Visualized Ocean Environment Model: Based on the established grid and physical characteristic values, construct a three-dimensional visualized ocean environment model.

[0050] S65, Visualization Result Display and Analysis: Display the generated three-dimensional visualized ocean environment model to the user, supporting interactive operations.

[0051] Advantages of the present invention:

[0052] In the present invention, marine physical data is collected by multi-source sensors, and the data is processed and optimized using deep learning models, quantum annealing optimization algorithms, and adaptive threshold detection algorithms. A hybrid model of a deep residual network and a graph convolutional network is used to achieve deep feature extraction of spatio-temporal data and capture spatial dependence relationships. The parameter set is dynamically optimized through the quantum annealing algorithm, thereby realizing efficient multi-source data fusion. This method effectively improves the accuracy and consistency of the data, can make full use of the data characteristics of different sensors, and realizes a more accurate analysis of the marine physical environment.

[0053] In the present invention, through a spatio-temporal correction method based on a dynamic sliding window, the spatio-temporal deviation in the sensor data is accurately corrected to ensure the reliability and consistency of the data. Combined with the adaptive threshold detection algorithm, abnormal nodes in the data are effectively identified to ensure the high-quality construction of the marine environment model. This method can adaptively adjust the detection threshold, and improve the sensitivity of detection through local anomaly detection, reduce the false alarm rate and the missed alarm rate, further improve the recognition accuracy of abnormal data, and ensure the stability and reliability of data processing.

[0054] In the present invention, by constructing a three-dimensional visual marine environment model and dynamically establishing the mapping relationship between model parameters and physical characteristics, an intuitive and efficient marine environment display platform is provided. The optimized parameter set is obtained through the quantum annealing optimization algorithm and mapped onto a three-dimensional grid, supporting real-time updates and interactive operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only those of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0056] Figure 1 It is a schematic flowchart of the method according to an embodiment of the present invention;

[0057] Figure 2 It is a schematic flowchart of the S3 process according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0058] The following will describe the present invention in detail with reference to the drawings and specific embodiments. At the same time, it should be noted here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments. For some well-known technologies, those skilled in the art can also adopt other alternative methods for implementation; moreover, the drawings are only for more specifically describing the embodiments and are not intended to specifically limit the present invention.

[0059] Such as Figure 1 - Figure 2As shown in the figure, the intelligent fusion and optimization processing method for ocean physical data includes the following steps:

[0060] S1: Perform standardized preprocessing on the ocean physical data collected by multi-source sensors to generate a standardized matrix containing temperature, salinity, flow velocity, tide, and acoustic wave characteristics;

[0061] S2: Construct a spatio-temporal correction matrix based on a dynamic sliding window to perform dynamic spatio-temporal reference correction on the standardized matrix and generate a spatio-temporal correction matrix;

[0062] S3: Input the spatio-temporal correction matrix into a deep residual-graph convolution hybrid model for feature fusion and output a multi-source fusion data set;

[0063] S4: Construct a multi-dimensional abnormal feature association matrix based on the multi-source fusion data set and use an adaptive threshold detection algorithm to identify abnormal data nodes;

[0064] S5: Use the quantum annealing optimization algorithm to perform parameter optimization iteration on the multi-source fusion data set and generate an optimization parameter set;

[0065] S6: Construct a three-dimensional visualization ocean environment model according to the optimization parameter set and establish a dynamic mapping relationship between the model parameters and physical characteristics.

[0066] S1 includes:

[0067] S11, Data collection: By setting multiple different types of ocean sensors, collect ocean physical data in real time, including temperature, salinity, flow velocity, tide, and acoustic wave characteristic data;

[0068] Temperature data is collected by a CTD sensor (Conductivity, Temperature, Depth sensor), which can simultaneously measure the temperature, conductivity, and depth of seawater and calculate salinity based on conductivity;

[0069] Salinity data is collected by a CTD sensor or a conductivity sensor, and the conductivity sensor uses the conductivity value to estimate the seawater salinity;

[0070] Flow velocity data is collected by an acoustic Doppler current profiler, which uses the acoustic Doppler effect to calculate the water body flow velocity and can provide flow velocity information at different depths;

[0071] Tide data is collected by a pressure sensor or a tide gauge. The pressure sensor measures the water level change, and the tide gauge is specifically used to monitor the sea level change to obtain tide data;

[0072] Acoustic wave characteristic data is collected by a sonar system or an ocean acoustic sensor. The sonar system can measure the propagation speed, reflection intensity, and other acoustic characteristics of acoustic waves in water;

[0073] S12, Data Formatting: Format the collected ocean physical data to ensure that different data sources have unified timestamp and spatial location markings, forming a dataset with time-space matching;

[0074] S13, Missing Value Filling: For sensor data with missing or abnormal values, fill in the missing values through interpolation or smoothing methods based on adjacent data to ensure data continuity and integrity;

[0075] S14, Normalization: Normalize the collected ocean physical data so that the value ranges of various data indicators are unified between [0, 1]. The specific method is as follows:

[0076] Temperature: Perform min-max normalization on the collected temperature data. The formula is:

[0077] ;

[0078] where T is the original temperature data value, the minimum value of temperature in the dataset, the maximum value of temperature in the dataset, the normalized temperature data, with a range between [0, 1];

[0079] Salinity: Perform Z-score standardization on the salinity data. The formula is:

[0080] ;

[0081] where S is the original salinity data value, is the mean value of the salinity data, is the standard deviation of the salinity data, is the standardized salinity data;

[0082] Flow Velocity: Adopt the standard deviation normalization method. The formula is:

[0083] ;

[0084] where V is the original flow velocity data value, is the standard deviation of the flow velocity data, is the normalized flow velocity data;

[0085] Tide: Perform periodic normalization on the tide data, and make normalization adjustments with daily, monthly, and annual cycles to ensure that the influence of periodic changes is eliminated;

[0086] Acoustic Wave Characteristics: After performing logarithmic transformation on the acoustic wave characteristics, then perform normalization processing to reduce the influence of extreme values in the data;

[0087] S15, Feature extraction: Extract temporal features and spatial features from the standardized data, such as mean value, standard deviation, maximum value, minimum value, and the change rate of the data, etc.;

[0088] S16, Construct a standardized matrix: Combine the normalized ocean physical data and the extracted eigenvalue to form a standardized matrix, which serves as the input for subsequent spatio-temporal correction processing.

[0089] S2 includes:

[0090] S21, Sliding window setting and data partitioning: According to the acquisition frequency and spatial resolution of the ocean physical data, set the size and step of the dynamic sliding window;

[0091] The window size is adjusted according to the time span (such as hours, days, months) and spatial range (such as longitude and latitude range or water depth range);

[0092] According to the set sliding window, divide the standardized matrix into multiple spatio-temporal sub-matrices in the time and space dimensions. Each sub-matrix contains a data set within a preset time period and a preset spatial range. The preset spatial range is jointly defined by the latitude range, longitude range, and water depth range. For example, based on the coastline of the actual sea area, delimit the outer edge of the research area, and then divide this area into multiple adjacent local units according to established rules; These local units can be several grid blocks of similar size, or can be divided according to the sensor distribution in the sea area, the homogeneity of ocean elements, or administrative divisions. No matter which division method is adopted, it is necessary to ensure that the ocean physical data monitored within each local unit has internal consistency, so that the local unit can represent the environmental state of a certain local sea area; When further divided in combination with the time dimension, "spatio-temporal sub-matrices" can be formed within each local unit for subsequent dynamic sliding window analysis and data correction;

[0093] S22, Selection of spatio-temporal reference data: Among multiple spatio-temporal sub-matrices, select representative reference data as the reference benchmark for spatio-temporal correction. The reference data is stable and representative ocean physical data, usually selecting data from time periods and spatial positions with good sensor conditions and small environmental changes. The purpose of selecting reference data is to ensure the consistency and accuracy of the subsequent correction process.

[0094] S23, Spatio-temporal deviation analysis: By comparing the data of each spatio-temporal sub-matrix with the reference data, analyze the spatio-temporal deviation of the data. The specific method uses correlation analysis, regression analysis, or error matrix calculation to find the spatio-temporal deviation caused by factors such as sensor error, environmental change, or data noise. This analysis helps to identify the specific sources of data deviation and provides a basis for the construction of a dynamic correction model.

[0095] S24, Dynamic Calibration Model Construction: According to the results of spatio-temporal deviation analysis, a spatio-temporal calibration model is established using statistical modeling methods (such as the least squares method, Kalman filtering method, etc.). This model automatically adjusts spatio-temporal data based on spatio-temporal data and deviation information, eliminates the deviation in the data, and can dynamically adjust the calibration parameters according to the actually collected environmental data to ensure the real-time and accuracy of calibration;

[0096] S25, Spatio-Temporal Calibration Calculation: Apply the spatio-temporal calibration model to the data of each spatio-temporal sub-matrix. Through model calculation, obtain the calibration parameters of each spatio-temporal window, and perform spatio-temporal calibration on the data of each sub-matrix. The calibration process includes the adjustment of data such as temperature, salinity, flow velocity, tide, and acoustic wave characteristics to ensure the consistency of data in time and space and eliminate the deviation caused by factors such as sensor errors and environmental fluctuations;

[0097] S26, Spatio-Temporal Calibration Matrix Generation: Merge multiple calibrated spatio-temporal sub-matrices to generate a complete spatio-temporal calibration matrix. This matrix contains calibrated ocean physical data (such as temperature, salinity, flow velocity, tide, etc.) to ensure that each data point is dynamically calibrated in the spatio-temporal dimension and meets the unified benchmark standard;

[0098] S27, Calibration Result Evaluation and Optimization: Evaluate the quality of the generated spatio-temporal calibration matrix. The evaluation criteria include the smoothness, consistency, precision, and accuracy of the data, etc. Evaluate the calibration effect by calculating indicators such as the standard deviation, deviation, and error range of the data. If the calibration result does not meet the expectations, adjust the sliding window parameters, the algorithm of the calibration model, or the selection of reference data to further optimize the calibration effect.

[0099] S3 includes:

[0100] S31, Deep Residual Network Module Construction and Feature Extraction: Construct a deep residual network module to extract the temporal features in the spatio-temporal calibration matrix through the residual learning mechanism. The residual network adopts a residual block structure, which can alleviate the problem of gradient disappearance and extract more meaningful features. The deep residual network module includes the following steps:

[0101] Residual Block Design: Each residual block contains two convolutional layers and a skip connection. The input is directly added to the output through the skip connection to enhance the learning ability of the network.

[0102] Feature Extraction: Extract high-level temporal and spatial features from the spatio-temporal calibration matrix by stacking multiple residual blocks.

[0103] Activation Function Application: Apply a non-linear activation function (such as ReLU) after each convolutional layer to enhance the expression ability of the features.

[0104] The features processed by the deep residual network module will be used for subsequent graph convolution processing;

[0105] S32, Graph Convolution Network Module Construction and Spatial Feature Extraction: Construct a graph convolution network module and use graph convolution methods to capture the spatial features in the spatio-temporal correction matrix. The specific steps are as follows:

[0106] Graph Structure Construction: Convert the spatial information in the spatio-temporal correction matrix (such as the positions of ocean sensors, the spatial dependence relationships between data, etc.) into a graph structure. Each spatio-temporal data point serves as a node in the graph, and the edges between the nodes represent the spatial associations between the data.

[0107] Graph Convolution Operation: In the graph convolution network, the information of each node's neighborhood is aggregated through convolution operations. Specifically, for each node, the graph convolution layer updates the feature representation of the node by aggregating the information of neighboring nodes (such as the correlations of physical quantities such as temperature and salinity).

[0108] Adjacency Matrix and Weight Matrix: The core operation of the graph convolution network is convolution based on the adjacency matrix (representing the connection relationships between nodes) and the weight matrix (representing the strength of each connection). This process can learn the complex spatial relationships between nodes.

[0109] The spatial features processed by the graph convolution network module will be fused with the temporal features extracted by the deep residual network module.

[0110] S33, Weighted Fusion and Multi-source Fusion Dataset Generation: Perform weighted fusion on the temporal features and spatial features extracted by the deep residual network module and the graph convolution network module respectively to form a comprehensive feature representation. The specific steps include:

[0111] Weighted Fusion: Perform weighted summation on the temporal features and spatial features output by the deep residual network and the graph convolution network. The weights are obtained through training. Weighted fusion can dynamically adjust the contribution ratios of the temporal features and spatial features in the final fusion result according to their importance.

[0112] Fused Feature Representation: The feature representation obtained through weighted fusion can retain both the temporal information and spatial information of the spatio-temporal correction matrix, has stronger representation ability, and can better reflect the comprehensive features of the ocean physical environment.

[0113] The comprehensive feature representation formed after fusion will be output as a multi-source fusion dataset. The multi-source fusion dataset contains the temporal and spatial features fused from multiple sensors (such as temperature, salinity, flow velocity, tide, and acoustic wave features), providing high-quality input data for subsequent anomaly detection, prediction modeling, or other tasks.

[0114] S34, Fusion Result Evaluation and Optimization: Evaluate the quality of the output multi-source fusion dataset. The evaluation metrics include data accuracy, stability, and fusion effect. Evaluate by calculating the error between the fused data and the real data, and using clustering or classification algorithms for verification. If the evaluation result does not meet the requirements, further improve the fusion effect by optimizing the network structure (such as adjusting the number of layers of the residual block, modifying the parameters of the graph convolutional layer, etc.).

[0115] S4 includes:

[0116] S41, Construct a Multidimensional Anomaly Feature Correlation Matrix: According to multiple features in the multi-source fusion dataset (such as temperature, salinity, flow velocity, tide, and acoustic wave features), construct a multidimensional anomaly feature correlation matrix. The specific steps include:

[0117] Feature Correlation Analysis: Calculate the linear correlation between each feature using the Pearson correlation coefficient. The formula is as follows:

[0118] ;

[0119] where, is the correlation coefficient between feature i and feature j, is the measurement data of the i-th sensor at the k-th moment, is the mean value of feature i, and n is the total number of data points.

[0120] By calculating the correlation, determine the relationship between each feature, and construct a symmetric correlation matrix R. Each element in the matrix represents the correlation between feature i and feature j;

[0121] Construct the Correlation Matrix: Organize the correlation calculation results of multiple features into an m×m matrix, where m is the number of features. Each matrix element represents the correlation between feature i and feature j;

[0122] Feature Weighting: Assign a weight to each feature, and weight it according to the importance of the feature and its relevance to other features. The specific method is to use the average correlation of the feature to determine the weight. The formula is as follows:

[0123] ;

[0124] where, is the weight of feature i, is the correlation between feature i and feature j;

[0125] S42. Adaptive threshold detection algorithm is adopted: By using the adaptive threshold detection algorithm, the data in the multi-dimensional abnormal feature correlation matrix is analyzed to identify abnormal data nodes. The specific steps are as follows:

[0126] Determination of adaptive threshold: For each feature data point , its mean value is calculated and standard deviation , and the dynamic threshold is determined based on the 3σ rule:

[0127] ;

[0128] Among them, is the threshold of feature i, is the adaptive factor, which adjusts the sensitivity of the threshold. Usually, takes values between 2 and 4;

[0129] Abnormality determination: If the data point exceeds the threshold , it is determined as an abnormal data node. The judgment condition is as follows:

[0130] ;

[0131] Local anomaly detection: For local anomalies, the local weighted method is used for adjustment. Specifically, the abnormal points in the local area are weighted and calculated, and the local mean value and standard deviation are used to update the threshold calculation:

[0132] ;

[0133] Among them, is the local threshold, is the local anomaly sensitivity factor, usually taking 2.

[0134] S43. Identification and output of abnormal data nodes: After being processed by the adaptive threshold detection algorithm, all abnormal data nodes are identified and these abnormal data nodes are output. The output results include:

[0135] Position and marking of abnormal data nodes: Mark the specific positions of each abnormal data node and associate them with the corresponding physical features (such as temperature, salinity, etc.);

[0136] Analysis of abnormal data types: Classify the identified abnormal data and analyze the reasons for their anomalies (such as equipment failures, environmental interferences, etc.) to assist subsequent processing or optimization;

[0137] S44. Evaluation and optimization of abnormal detection results: Evaluate the results of abnormal data identification. The evaluation criteria include:

[0138] Anomaly detection accuracy: Evaluate the accuracy of anomaly detection by comparing the error between the recognition result and the true data;

[0139] False alarm rate and miss rate: Calculate the false alarm rate and miss rate to evaluate the sensitivity and specificity of the algorithm;

[0140] Optimization and adjustment: If the evaluation result does not meet the expectation, optimize the detection result by adjusting the parameters in the adaptive threshold algorithm or improving the construction method of the anomaly feature correlation matrix.

[0141] S5 includes:

[0142] S51, Define the optimization objective function: Determine the optimization objective function used to measure the performance of each parameter during the optimization process. This objective function reflects the overall effect of data processing and fusion and takes into account the adjustability of the parameters;

[0143] Let the optimization objective function be , where is the parameter set to be optimized, and the objective function is defined as a comprehensive index of fusion accuracy, data consistency, and anomaly detection accuracy in the multi-source fusion dataset. Specifically:

[0144] Fusion accuracy Data consistency Anomaly detection accuracy ;

[0145] Among them, the fusion accuracy represents the accuracy of the data after multi-source data fusion, the data consistency (p) represents the consistency between the fused data and the original data, and the anomaly detection accuracy (p) represents the recognition accuracy of the abnormal data nodes. , , are the weighting coefficients that control the relative importance of each index. The optimization objective function can be obtained through methods such as data fitting and error minimization;

[0146] S52, Quantum annealing algorithm initialization: Use the quantum annealing algorithm for parameter optimization, initialize the parameter set , and represent the initial state of each parameter on the qubits. The goal of quantum annealing is to gradually search for the optimal parameter combination through the evolution of the qubit state. The initialization steps are as follows:

[0147] Initialize the parameter set , and each parameter value is randomly selected within a predetermined range. Set the search range of the parameter as , where and They are the minimum and maximum values of the \(i\)-th parameter respectively.

[0148] In the quantum annealing algorithm, each parameter is mapped to the state of a qubit , and the initial state can be set to a uniform superposition state;

[0149] ;

[0150] Among them, and are the ground states of the qubit, is the superposition state of the qubit, representing a random selection of the parameter between the minimum and maximum values;

[0151] S53, Quantum annealing evolution process: The evolution process of quantum annealing is to perform an annealing operation on the qubit state, continuously reducing the "temperature" of the system and gradually reaching the optimization goal. The specific evolution process is as follows:

[0152] Hamiltonian design: Design the quantum Hamiltonian , which is associated with the objective function . The Hamiltonian guides the qubit to find the lowest energy state in the parameter space, thereby minimizing the objective function. The Hamiltonian can be expressed as:

[0153] ;

[0154] Among them, is the adjustment term, controlling the deviation degree between the objective function and the initial parameter, is the objective function, representing the optimization error or cost, serves as an additional constraint term to ensure that the optimization process does not deviate far from the initial value.

[0155] Temperature scheduling: Gradually reduce the energy of the system through the "temperature" scheduling of quantum annealing. The temperature \(T\) of the quantum system gradually decreases over time. During the annealing process, the system transitions from a high-temperature state (with strong randomness) to a low-temperature state (with enhanced stability), and finally converges to the optimal solution;

[0156] ;

[0157] Among them, is the system temperature at the \(t\)-th moment, is the initial temperature, is the maximum temperature, and \(t\) is the current time step;

[0158] S54, Measurement and Parameter Update of Quantum Bits: After the quantum system undergoes the annealing process, at each moment when the temperature drops to the lowest, the quantum bits are measured to obtain the optimized parameter set , and the specific steps are as follows:

[0159] Quantum Measurement: Measure the quantum bits to obtain the optimized parameter set , that is, by measuring the state of the quantum bits to determine the optimal value of each parameter . The measurement of the quantum bits can be obtained by calculating the probability amplitude:

[0160] ;

[0161] where is the probability of the measurement result, is the probability that the quantum bit is in the ground state ;

[0162] Parameter Update: Update the optimized parameter set according to the measurement result , and use it as the new parameter value to input into the next round of optimization process;

[0163] The optimization process continuously iterates and updates the parameters through the quantum annealing algorithm until one of the following termination conditions is met:

[0164] Optimization Convergence: When the change in the objective function is less than the preset threshold , that is:

[0165] ;

[0166] where and are the parameter sets of the current and the previous round of optimization respectively, is the convergence threshold;

[0167] Maximum Number of Iterations: If the maximum number of iterations is reached, the optimization process is terminated;

[0168] The optimization result is the generated optimized parameter set , and this parameter set can be used for subsequent feature fusion and data analysis tasks.

[0169] S6 includes:

[0170] S61. Optimize parameter set input and initialization: Use the generated optimized parameter set as the basic parameters for constructing a three-dimensional visual marine environment model. The optimized parameter set contains the optimal parameters after quantum annealing optimization on the multi-source fusion data set. These parameters are the core inputs for constructing the model and initialize the basic information required for the marine environment model. The basic information includes the boundary range of the marine area and the meshing parameters, such as the boundary range of the marine area (such as longitude, latitude, depth, etc.) and the meshing parameters.

[0171] Let the longitude and latitude range of the marine area be: ;

[0172] The depth range is: ;

[0173] S62. Three-dimensional grid construction: According to the boundary information of the marine area, construct a three-dimensional grid. The resolution of the grid depends on the meshing parameters in the optimized parameter set (such as the horizontal and vertical resolutions of the grid). This step divides the marine space into multiple small units to construct a discrete three-dimensional coordinate system for subsequent data mapping and visualization;

[0174] Grid division: Assume that the marine area is divided into grids, where is the horizontal resolution, is the vertical resolution, is the resolution in the depth direction;

[0175] Grid node calculation: Each grid node represents a position in the marine area, where , , is the depth interval;

[0176] S63. Dynamic mapping of physical characteristics and model parameters: According to the optimized parameter set, assign corresponding physical characteristic values (such as temperature, salinity, flow velocity, tide, sound wave, etc.) to each grid node . The dynamic mapping relationship between the model parameters and the physical characteristics is established based on the sensor data characteristics and their variation laws during data acquisition. The specific steps are as follows:

[0177] Physical characteristic model: According to the optimized parameters, use interpolation or regression algorithms to map the physical characteristics to the three-dimensional grid nodes. Suppose there is temperature data . Then, according to the changes in space and time, use an interpolation method (such as Kriging interpolation) to calculate the temperature value for each grid node, expressed as:

[0178] ;

[0179] Among them, is the temperature data of the known node m, is the weight based on the spatial distance, is the grid node at the temperature value;

[0180] Similarly, other physical characteristics (such as salinity, flow velocity, etc.) can be mapped to each grid node through a similar interpolation method;

[0181] S64, construction of a three-dimensional visualization ocean environment model: Based on the established grid and physical characteristic values, construct a three-dimensional visualization ocean environment model;

[0182] This model generates a visual image or animation by mapping the physical characteristics of each grid node into three-dimensional space for intuitive analysis. The specific steps include:

[0183] Surface rendering: Use three-dimensional drawing tools (such as OpenGL, Matplotlib 3D, etc.) to draw the surface of the ocean environment according to the physical characteristics of each grid node, and display the distribution of data such as temperature and salinity.

[0184] Volume rendering: Adopt volume rendering technology to stack each layer of the three-dimensional grid (such as different depths, different temperature layers, etc.) to form a complete ocean environment model.

[0185] S65, visualization result display and analysis: Display the generated three-dimensional visualization ocean environment model to the user, support interactive operations, and the user can adjust the viewing angle, zoom, rotation, etc. through the interface to view the ocean environment characteristics of different regions and depths.

[0186] Interactive interface design: Design a user-friendly interactive interface that allows the user to select different ocean physical characteristics for display (such as temperature, salinity, flow velocity, etc.) and supports viewing from different perspectives.

[0187] The present invention covers any substitutions, modifications, equivalent methods, and solutions made within the spirit and scope of the present invention. To enable the public to have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, and those skilled in the art can fully understand the present invention without these detailed descriptions. Additionally, to avoid unnecessary confusion to the essence of the present invention, well-known methods, processes, procedures, components, and circuits, etc. are not described in detail.

[0188] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. The method for intelligent fusion and optimization processing of ocean physical data is characterized by: The following steps are involved: S1: Standardize and preprocess the ocean physical data collected by multi-source sensors to generate a standardized matrix containing temperature, salinity, current velocity, tide and acoustic wave characteristics; S2: Construct a spatiotemporal correction matrix based on a dynamic sliding window, perform spatiotemporal benchmark dynamic correction on the standardized matrix, and generate a spatiotemporal correction matrix, which specifically includes: S21, sliding window setting and data division: according to the acquisition frequency and spatial resolution of ocean physical data, the size and step length of the dynamic sliding window are set; The window size is adjusted according to the time span and spatial range; According to the set sliding window, the standardized matrix is ​​divided into multiple spatiotemporal sub-matrices according to the time and space dimensions, and each sub-matrix contains a data set within a time period and a preset spatial range; S22, spatiotemporal benchmark data selection: selecting representative benchmark data from multiple spatiotemporal sub-matrices as reference benchmarks for spatiotemporal correction; S23, spatiotemporal deviation analysis: analyzing the deviation of spatiotemporal data by comparing the data of each spatiotemporal submatrix with the benchmark data; S24, dynamic correction model construction: according to the results of spatiotemporal deviation analysis, a spatiotemporal correction model is established using statistical modeling methods; S25, spatiotemporal correction calculation: applying the spatiotemporal correction model to the data of each spatiotemporal submatrix, obtaining the correction parameters of each spatiotemporal window through model calculation, and performing spatiotemporal correction on the data of each submatrix; S26, generating a space-time correction matrix: merging the corrected multiple space-time sub-matrices to generate a complete space-time correction matrix; S27, correction result evaluation and optimization: perform quality evaluation on the generated spatiotemporal correction matrix; S3: Input the spatiotemporal correction matrix into the deep residual-graph convolution hybrid model for feature fusion and output a multi-source fusion dataset; S4: Construct a multi-dimensional abnormal feature association matrix based on multi-source fusion data sets, and use an adaptive threshold detection algorithm to identify abnormal data nodes; S5: Use the quantum annealing optimization algorithm to iterate the parameters of the multi-source fusion data set and generate an optimized parameter set; S6: Construct a 3D visual ocean environment model based on the optimized parameter set, and establish a dynamic mapping relationship between model parameters and physical characteristics, including: S61, input and initialization of optimization parameter set: using the generated optimization parameter set as the basic parameters for constructing a three-dimensional visualized ocean environment model, and initializing the basic information required for the ocean environment model, the basic information including the boundary range of the ocean area and gridding parameters; S62, 3D grid construction: constructing a 3D grid according to the boundary information of the ocean area; S63, dynamic mapping of physical characteristics and model parameters: assigning corresponding physical characteristic values ​​to each grid node according to the optimized parameter set; S64, construction of a three-dimensional visual ocean environment model: constructing a three-dimensional visual ocean environment model based on the established grid and physical eigenvalues; S65, visualization result display and analysis: Display the generated 3D visualization ocean environment model to the user and support interactive operation.

2. The method for intelligent fusion and optimization processing of ocean physical data according to claim 1 is characterized in that: The S1 includes: S11, data collection: by setting up multiple different types of ocean sensors, real-time collection of ocean physical data, including temperature, salinity, current velocity, tide and acoustic wave characteristic data; S12, data formatting: formatting the collected ocean physical data to form a data set that matches time and space; S13, missing value filling: For sensor data with missing or abnormal data, the missing values ​​are filled by interpolation or smoothing method based on adjacent data; S14, normalization processing: normalizing the collected ocean physical data; S15, feature extraction: extracting temporal features and spatial features from the standardized data; S16, constructing a standardized matrix: combining the normalized ocean physical data and the extracted eigenvalues ​​to form a standardized matrix.

3. The method for intelligent fusion and optimization processing of ocean physical data according to claim 1 is characterized in that: The S3 includes: S31, deep residual network module construction and feature extraction: construct a deep residual network module and extract the temporal features in the spatiotemporal correction matrix through the residual learning mechanism; S32, Graph Convolutional Network Module Construction and Spatial Feature Extraction: Construct a graph convolutional network module and use the graph convolution method to capture the spatial features in the spatiotemporal correction matrix; S33, weighted fusion and multi-source fusion dataset generation: weighted fusion of temporal features and spatial features extracted by the deep residual network module and graph convolutional network module to form a comprehensive feature representation; The comprehensive feature representation formed after fusion will be output as a multi-source fusion dataset, which contains the temporal and spatial features fused from multiple sensors; S34, fusion result evaluation and optimization: perform quality evaluation on the output multi-source fusion data set. The evaluation indicators include data accuracy, stability and fusion effect.

4. The method for intelligent fusion and optimization processing of ocean physical data according to claim 3 is characterized in that: The S4 includes: S41, constructing a multi-dimensional abnormal feature association matrix: constructing a multi-dimensional abnormal feature association matrix according to multiple features in the multi-source fusion data set; S42, using an adaptive threshold detection algorithm: using an adaptive threshold detection algorithm to analyze the data in the multi-dimensional abnormal feature association matrix and identify abnormal data nodes; S43, abnormal data node identification and output: after being processed by the adaptive threshold detection algorithm, all abnormal data nodes are identified and output; S44, anomaly detection result evaluation and optimization: evaluate the abnormal data identification results.

5. The method for intelligent fusion and optimization processing of ocean physical data according to claim 4 is characterized in that: The S5 includes: S51, define optimization objective function: determine the optimization objective function, which is used to measure the performance of each parameter in the optimization process; S52, quantum annealing algorithm initialization: Use the quantum annealing algorithm to optimize parameters, initialize the parameter set, and represent the initial state of each parameter on the quantum bit. The goal of quantum annealing is to gradually search for the optimal parameter combination through the evolution of the quantum bit state; S53, quantum annealing evolution process: The quantum annealing evolution process continuously lowers the "temperature" of the system and gradually reaches the optimization goal by annealing the quantum bit state; S54, measurement and parameter update of quantum bits: After the quantum system has undergone the annealing process, each time the temperature drops to the lowest, the quantum bits are measured to obtain the optimized parameter set.

Images