Multi-source heterogeneous ocean data intelligent fusion and ocean disaster prediction method and platform

Abstract

The invention relates to the technical field of data processing, and discloses a multi-source heterogeneous marine data intelligent fusion and marine disaster prediction method and platform, and the method comprises the steps: carrying out the hierarchical deployment and network topology construction of marine environment monitoring equipment, communication relay equipment and data processing equipment, and obtaining a sea-air-shore integrated monitoring network; collecting marine environment data with a space-time identifier; carrying out ocean element modeling and dynamic feature extraction to obtain an ocean environment digital model; performing feature fusion and correlation analysis on the multi-source heterogeneous ocean data to obtain spatio-temporal feature data; and training and multi-model integration are carried out on the typhoon path prediction model, the storm surge prediction model and the red tide prediction model by adopting a transfer learning method to obtain a marine disaster prediction model set, so that the prediction precision and timeliness are improved, the allocation and scheduling of computing resources are optimized, and the real-time requirement of mass data processing is met.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a method and platform for intelligent fusion of multi-source heterogeneous ocean data and prediction of ocean disasters. Background Art

[0002] Traditional marine disaster prediction methods mainly rely on a single data source and simple numerical models, which makes it difficult to accurately grasp the complex dynamic changes in the marine environment, resulting in insufficient prediction accuracy and timeliness.

[0003] The data generated by the current ocean monitoring system has the characteristics of multi-source heterogeneity, uneven temporal and spatial distribution, and severe noise interference. How to effectively process and integrate these data has become a key challenge for marine disaster prediction. At the same time, there are complex interactions between different types of marine disasters, and a single prediction model is difficult to accurately describe the multi-scale dynamic characteristics of the disaster evolution process. In addition, the real-time processing of massive multi-source heterogeneous data places higher demands on computing resources, and the existing data processing methods lack effective resource scheduling and optimization mechanisms, making it difficult to meet the dual needs of marine disaster warning for real-time and accuracy. Summary of the invention

[0004] The present application provides a method and platform for intelligent fusion of multi-source heterogeneous ocean data and marine disaster prediction, thereby improving prediction accuracy and timeliness, optimizing the allocation and scheduling of computing resources, and meeting the real-time requirements of massive data processing.

[0005] The first aspect of the present application provides a method for intelligent fusion of multi-source heterogeneous ocean data and prediction of marine disasters, the method comprising: Hierarchical deployment and network topology construction of marine environment monitoring equipment, communication relay equipment and data processing equipment to obtain an integrated sea, air and shore monitoring network; According to the integrated sea-air-shore monitoring network, a periodic deterministic sampling algorithm is used to collect and standardize water temperature, salinity, sea level, wind speed and air pressure data to obtain marine environmental data with time and space identification; Based on the marine environment data with time and space identification, marine element modeling and dynamic feature extraction are performed to obtain a digital model of the marine environment; Based on the digital model of the ocean environment, a selective caching mechanism and a deep learning algorithm are used to perform feature fusion and correlation analysis on multi-source heterogeneous ocean data to obtain spatiotemporal feature data; According to the spatiotemporal characteristic data, the typhoon path prediction model, the storm surge prediction model and the red tide prediction model are trained and integrated by the transfer learning method to obtain a marine disaster prediction model set.

[0006] The second aspect of the present application provides a multi-source heterogeneous ocean data intelligent fusion and ocean disaster prediction platform, the multi-source heterogeneous ocean data intelligent fusion and ocean disaster prediction platform comprising: The construction module is used to hierarchically deploy marine environment monitoring equipment, communication relay equipment and data processing equipment and construct network topology to obtain an integrated sea, air and shore monitoring network; A collection module is used to collect and standardize water temperature, salinity, sea level, wind speed and air pressure data according to the integrated sea, air and shore monitoring network using a periodic deterministic sampling algorithm to obtain marine environmental data with time and space identification; A feature extraction module, used to perform ocean element modeling and dynamic feature extraction based on the ocean environment data with time and space identification to obtain a digital model of the ocean environment; An association analysis module is used to perform feature fusion and association analysis on multi-source heterogeneous ocean data based on the ocean environment digital model using a selective caching mechanism and a deep learning algorithm to obtain spatiotemporal feature data; The model integration module is used to train and integrate the typhoon path prediction model, the storm surge prediction model and the red tide prediction model according to the spatiotemporal feature data using a transfer learning method to obtain a set of marine disaster prediction models.

[0007] The third aspect of the present application provides an electronic device, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the electronic device executes the above-mentioned multi-source heterogeneous ocean data intelligent fusion and ocean disaster prediction method.

[0008] The fourth aspect of the present application provides a computer-readable storage medium, which stores instructions. When the computer-readable storage medium is run on a computer, it enables the computer to execute the above-mentioned multi-source heterogeneous ocean data intelligent fusion and ocean disaster prediction method.

[0009] Compared with the prior art, the present application has the following beneficial effects: by constructing an integrated sea, air and shore monitoring network, unified collection and transmission of multi-source heterogeneous ocean data is realized, the problems of scattered data acquisition and non-uniform format are solved, and the periodic deterministic sampling algorithm and data standardization processing are adopted to effectively reduce the noise interference in the data collection process and ensure the data quality and timeliness; a digital model of the ocean environment is established based on digital twin technology, which accurately depicts the dynamic correlation between ocean environmental elements and enhances the cognitive ability of the laws of ocean environmental changes; the combination of selective caching mechanism and deep learning algorithm optimizes the utilization efficiency of computing resources and improves the accuracy of feature fusion; through transfer learning methods and multi-model integration technology, the knowledge of pre-trained models is fully utilized to enhance the generalization performance of different types of marine disaster prediction models; a complete model training and optimization process is established to realize the intelligent processing of the entire process from data collection to disaster prediction, which significantly improves the prediction accuracy and timeliness; the distributed computing architecture and edge computing technology are adopted to optimize the allocation and scheduling of computing resources and meet the real-time requirements of massive data processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0011] The structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification so as to facilitate understanding and reading by persons familiar with this technology. They are not used to limit the conditions under which the present invention can be implemented, and therefore have no substantive technical significance. Any structural modification, change in proportion or adjustment of size, without affecting the effects and purposes that can be achieved by the present invention, should still fall within the scope of the technical contents disclosed by the present invention.

[0012] Figure 1 It is a flow chart of a method for intelligent fusion of multi-source heterogeneous ocean data and prediction of ocean disasters provided by an embodiment of the present invention; Figure 2 It is a schematic block diagram of the structure of a multi-source heterogeneous ocean data intelligent fusion and ocean disaster prediction platform provided by an embodiment of the present invention; Figure 3 It is a schematic block diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0013] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0014] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0015] It should also be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0016] It should be further understood that the term "and / or" used in the specification and appended claims of this application refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations. Figure 1 In the embodiments of the present application, an embodiment of the method for intelligent fusion of multi-source heterogeneous ocean data and prediction of marine disasters includes: Step 100: hierarchically deploy marine environment monitoring equipment, communication relay equipment and data processing equipment and construct network topology to obtain a sea-air-shore integrated monitoring network; It is understandable that the execution subject of the present application can be a multi-source heterogeneous ocean data intelligent fusion and ocean disaster prediction platform, or a terminal or a server, which is not limited here. The present application embodiment is described by taking the server as the execution subject as an example.

[0017] Specifically, the monitoring area is divided into three levels: ocean, air and shore, to form monitoring grid units with clear spatial attributes. By analyzing the spatial distribution characteristics of the monitoring grid units, the monitoring density requirements under different environmental conditions are evaluated, and the deployment plan of the monitoring equipment is optimized so that the equipment covers the largest area under resource constraints and can meet the data collection requirements. After the equipment is deployed, the functional requirements of each device are refined and allocated, and its operation is divided into data acquisition module, data preprocessing module and communication module to achieve efficient operation of the equipment functions. Through the analysis of the specific use of the equipment, the data acquisition module is used to collect environmental parameters such as water temperature, salinity, and air pressure in real time, the data preprocessing module cleans and formats the raw data, and the communication module is used for data upload and coordination of communication between devices to ensure the stability and efficiency of the entire monitoring system. According to the functional allocation information of the equipment, it is input into the network topology generator, and the preliminary planning of the network is completed through the multi-level topology construction algorithm to generate a hierarchical network topology structure diagram covering the ocean, air and shore. Based on the network topology diagram, the communication relay equipment is deployed in different regions and the communication protocol is configured to ensure smooth data transmission between different levels. At the ocean level, the communication relay equipment includes the wireless transmitter of the sea surface buoy and the communication station of the unmanned ship; at the air level, drones and satellites are used as efficient data relay tools; and at the shore-based level, fixed communication sites or coastal stations are mainly relied on for data aggregation. In order to improve the transmission efficiency, a layered communication protocol is adopted, and short-distance and high-frequency transmission technology is used between the underlying communication relay equipment, while high-stability and broadband satellite communication protocols are used in long-distance transmission to achieve data synchronization and efficient sharing of the entire network. At the same time, in order to improve the immediacy and distributed processing capabilities of data processing, lightweight prediction models are deployed on edge computing servers in the data transmission network to form local data processing units. These local units mainly perform preliminary processing and analysis on the data collected by the monitoring equipment, reduce the bandwidth requirements for data transmission, and provide fast prediction results to support real-time decision-making. For tasks that require higher computing power and data integration capabilities, the data fusion model of the cloud computing center is used to decompose tasks and allocate computing resources according to the principle of distributed computing, and work with local data processing units through network connections to form a central computing unit. The central unit is responsible for globally integrating the scattered data from various monitoring areas to complete more complex analysis tasks, including intelligent fusion of multi-source data and training and updating of disaster prediction models. The data transmission network, local data processing unit and central computing unit are integrated through unified system integration and communication protocol specifications to form a highly collaborative and scalable integrated sea, air and shore monitoring network.The network can support multi-level monitoring and processing functions, coordinate data collection, transmission, processing and analysis tasks at different levels, and provide full-process services from raw data to prediction results.

[0018] Step 200: Based on the integrated sea-air-shore monitoring network, a periodic deterministic sampling algorithm is used to collect and standardize water temperature, salinity, sea level, wind speed and air pressure data to obtain marine environment data with time and space identification; Specifically, in the integrated sea-air-shore monitoring network, the working time slots of the marine environment monitoring equipment are divided periodically, and the sampling tasks are prioritized by evaluating the functions, data importance and resource constraints of different equipment to generate a sampling scheduling sequence. This sequence can optimize the data collection process of the monitoring equipment, so that high-priority data can be collected first during critical periods, and low-priority tasks can be reasonably arranged to ensure full utilization of system resources and balanced data coverage. According to the generated sampling scheduling sequence, the original data collected by the monitoring equipment are subjected to error calculation and threshold judgment to evaluate the credibility of the data. By introducing error calculation methods, such as statistical analysis based on historical mean deviation, equipment accuracy range and data distribution law, the quality of each data is accurately evaluated and a data credibility index is assigned to it. For records whose data credibility is lower than the preset threshold, they are input into the exception handling module for marking and elimination to avoid interference of abnormal data on subsequent analysis, while ensuring the quality reliability of the valid data set. The exception handling module can adopt a combination of rule detection and machine learning methods, such as filtering outliers by setting reasonable boundary conditions, and further identifying potential anomalies using clustering or classification algorithms. Based on the time distribution characteristics of the effective data set, the data sampling period is dynamically adjusted and synchronously controlled so that the data collection task can better adapt to the dynamic changes of the marine environment. For example, when monitoring emergencies such as typhoons or storm surges, the sampling period is shortened to improve the timeliness of the data, while in a stable state, the sampling interval is appropriately extended to save resources. The dynamic adjustment mechanism is implemented through an adaptive control algorithm, which can flexibly respond to environmental changes and reduce the energy consumption of the equipment. After completing the time series synchronization, the calibrated time series data is cleaned to remove redundant or duplicate records, and the data is converted into a standard format through normalization and numerical quantization. The normalization process maps data of different dimensions to a unified range, thereby eliminating the impact of dimensional differences on model input, while numerical quantization helps to improve data storage efficiency and computational convenience. In data format standardization, a geographic location code is assigned to each data record according to the spatial distribution of the monitoring equipment, and the time information collected by the equipment is used for timestamp marking to form a basic data set with spatiotemporal correlation. Based on the spatiotemporal correlation data, the data change rate and transmission delay between adjacent sampling points are calculated to quantitatively describe the dynamic characteristics and transmission characteristics of the data. The calculation of data change rate is achieved by measuring the incremental changes in a continuous time series, reflecting the dynamic change trend of environmental factors, while the transmission delay is measured by analyzing the transmission time of data from the device to the processing center, reflecting the real-time nature of data acquisition and processing. The spatiotemporal correlation data and its additional timeliness parameters are encapsulated in data packets and encoded using a transmission protocol for efficient transmission over the network. The encapsulation process includes packaging data records according to certain priorities and network protocol requirements, and adding necessary verification information and routing information to improve the reliability and integrity of transmission.Through the above steps, we finally get the marine environment data with time and space identification.

[0019] Step 300: Perform ocean element modeling and dynamic feature extraction based on ocean environment data with time and space identifiers to obtain an ocean environment digital model; It should be noted that the marine environmental data with time and space identification are classified according to different physical elements, and the water temperature field, salinity field, sea level height field, wind speed field and air pressure field are constructed respectively. Based on this, each type of physical element data is digitally encoded to generate a unified data structure. In this process, the physical field theory and data distribution law are combined, and the standardized format and coding rules are used to convert the complex physical environment information into a structured marine element data structure. The physical quantities in the above data structure are mapped in time series and space distribution, and the change law of marine environmental elements in different time and space ranges is revealed by constructing a time and space distribution model. Time series mapping uses dynamic time warping algorithm or Fourier transform technology to extract long-term trends and periodic characteristics, while spatial distribution mapping constructs high-resolution spatial distribution maps through interpolation methods and partition modeling to achieve accurate description of the time and space dynamics of the marine environment. Through this process, the time and space distribution characteristics of marine environmental elements are described. Based on the time and space distribution characteristics, the interaction relationship between marine environmental elements is coupled and the correlation is calculated. Coupling analysis is based on fluid dynamics theory and statistical models. By studying the linkage relationship between water temperature and salinity, wind speed and air pressure, etc., a correlation matrix reflecting the physical mechanism is constructed. The calculation of correlation uses mutual information, Granger causality or correlation coefficient to quantify the coupling strength between different elements and generate a marine element correlation matrix that can reflect the internal interaction of complex systems. The above correlation matrix is ​​input into the dynamic evolution model to analyze the changing law of the marine environmental state. The dynamic evolution model simulates the evolution of the marine environment over time and space by using a set of differential equations, a state space model or a data-driven machine learning algorithm, and generates an environmental evolution sequence. After obtaining the environmental evolution sequence, stability analysis and mutation feature detection are performed to identify abnormal events in the marine environment. For example, by using methods such as principal component analysis and singular value decomposition, the existing system mutations are detected, and the spatial and temporal distribution and impact range of abnormal events are located in combination with physical field models to generate complete abnormal event characteristics of the marine environment. Based on these abnormal event characteristics, a mapping function between the physical marine environment and the digital space is established to achieve the mapping of the physical environment to the digital model. This process is digital twin mapping. Digital twin mapping links physical elements with digital features in virtual space by constructing multi-dimensional mapping relationships, ensuring that the virtual environment can accurately reflect the actual physical state. The digital twin mapping relationship is input into the feature extractor to extract the key features that are most sensitive to environmental status and dynamic changes, and generate a highly condensed marine environment feature vector. Model integration and parameter optimization are performed on the marine environment feature vector and the digital twin mapping relationship. Model integration uses ensemble learning methods (such as Boosting or Bagging) to collaboratively optimize multiple modeling algorithms, while parameter optimization uses techniques such as particle swarm optimization or Bayesian optimization to adjust model parameters to improve prediction accuracy and generalization ability.In this process, by combining the digital twin model with the spatiotemporal distribution characteristics, data from different sources and structures are comprehensively integrated to generate a unified digital model of the ocean environment.

[0020] Step 400: Based on the digital model of the ocean environment, a selective caching mechanism and a deep learning algorithm are used to perform feature fusion and correlation analysis on multi-source heterogeneous ocean data to obtain spatiotemporal feature data; Specifically, in the digital model of the marine environment, according to the characteristics of different marine environmental elements, such as water temperature, salinity, sea level, wind speed and air pressure, these characteristics are systematically classified and layered according to their data types (such as continuous, discrete or categorical) and different patterns of spatiotemporal distribution, and a hierarchical structure of feature data is constructed. On this basis, by analyzing the usage frequency, correlation importance and update cycle of historical data features, a selective cache mechanism is established to cache data in a hierarchical and prioritized manner to generate a hierarchical cache database. Through the cache mechanism, data access latency is effectively reduced, while the processing efficiency of high-priority data is improved. The feature data in the hierarchical cache database is input into a deep neural network for processing to extract potential feature representations and complete nonlinear mapping. Deep neural networks can capture the high-order relationships and complex patterns implicit in feature data, especially nonlinear characteristics, through a multi-layer abstract learning process, thereby generating deep feature representations. These feature representations include spatial features (such as geographic location and regional distribution), temporal features (such as periodic changes and trends), and physical features (such as hydrodynamic characteristics and atmospheric circulation characteristics), which together constitute a multi-dimensional description of multi-source data. After the deep feature representation is generated, the spatial features, temporal features and physical features are subjected to multimodal feature fusion and complementarity analysis. This process uses multimodal learning methods, through joint distribution modeling, collaborative representation and feature alignment, to give full play to the complementary effects between the features of each modality and obtain fused feature vectors. These vectors represent the intrinsic connection and synergy between different features, making the subsequent analysis more comprehensive and accurate. The fused feature vectors are input into the attention calculation unit for feature weight assignment and saliency calculation. The attention mechanism dynamically evaluates the importance of different features, assigns appropriate weights to each feature, and generates a weighted feature set. The weighted feature set is organized according to the time dimension and the space dimension, and an indexing mechanism is established to form a structured spatiotemporal feature organization system. The feature data in the spatiotemporal feature organization system are subjected to similarity measurement and correlation mining. The correlation between different features is identified through similarity measurement, while correlation mining can reveal the potential complex relationship between features. This process generates a feature association map that describes the interactive network of multi-source data in the spatiotemporal dimension. The feature association map is input into the deep spatiotemporal fusion network for cross-modal feature learning and spatiotemporal dependency modeling. The deep spatiotemporal fusion network combines the advantages of convolutional neural networks (CNNs) and recurrent neural networks (RNNs) to capture local patterns of spatial features and model dependencies in time series, thereby generating high-dimensional fused feature tensors. The fused feature tensors are subjected to dimensionality reduction and feature selection. Dimensionality reduction reduces feature redundancy and retains the most representative features through techniques such as principal component analysis, singular value decomposition, or autoencoders.At the same time, feature selection selects the most relevant features based on importance scores or contribution assessments to generate the final spatiotemporal feature data.

[0021] Step 500: Based on the spatiotemporal feature data, a transfer learning method is used to train and integrate the typhoon path prediction model, the storm surge prediction model and the red tide prediction model to obtain a marine disaster prediction model set.

[0022] Specifically, according to the content of spatiotemporal feature data, it is divided into typhoon path prediction data set, storm surge prediction data set and red tide prediction data set, corresponding to different disaster types and prediction needs. In this process, the spatiotemporal characteristics and physical characteristics of the data are combined to ensure that each data set can fully cover the key elements of the target phenomenon. Based on the typhoon path prediction data set, a typhoon path prediction model with an encoder-decoder structure as the core is constructed. The encoder network of the model consists of a 5-layer convolution structure. Each convolution layer uses a ReLU activation function to enhance the nonlinear expression ability, and is equipped with a batch normalization layer to stabilize the training process and accelerate convergence. These convolution layers extract the motion characteristics of the typhoon path, and gradually capture the spatial distribution law and dynamic change trend of the typhoon through layer-by-layer convolution operations. The corresponding decoder network also consists of a 5-layer deconvolution structure, which can generate the prediction results of the typhoon path. In order to avoid feature loss and enhance information transfer, jump connections are added between the encoder and decoder to ensure that high-resolution features can be directly transferred to the decoding stage to generate the initial prediction results of the typhoon path. For the storm surge prediction dataset, a storm surge prediction model is constructed. The model consists of two parts: a temporal feature extraction network and a spatial feature fusion network. The temporal feature extraction network adopts a bidirectional long short-term memory network (Bi-LSTM) structure, which can extract the previous and next dependencies from the time series of tidal level changes and capture the dynamic change pattern and long-term trend of the tidal level. At the same time, the spatial feature fusion network adopts a graph convolutional network (GCN) structure. By performing correlation analysis on multi-point tidal level data, a spatial distribution map of tidal level characteristics in the region is constructed to explore the linkage effect between different regions. Through the synergy of these two sub-networks, the model can generate initial prediction results for storm surges. Based on the red tide prediction dataset, a red tide prediction model is constructed. The model adopts a structure that combines a multi-scale feature extraction module and a spatiotemporal attention module to deal with the complex mechanism and variability of red tide occurrence. In the multi-scale feature extraction module, three parallel residual networks are used, each of which focuses on feature extraction at different scales to ensure that the model can simultaneously capture local and global information of red tide occurrence. The spatiotemporal attention module focuses on analyzing key areas and time periods, and uses the attention mechanism to dynamically adjust the degree of attention to different data to generate initial prediction results for red tides. After all initial prediction models are built, they are input into the transfer learning network together with the pre-trained marine disaster prediction model library for knowledge transfer and model optimization. The transfer learning network contains a feature mapping layer and a task adaptation layer. The feature mapping layer transfers the knowledge of the pre-trained model to the data distribution of the current task, while the task adaptation layer further adjusts the model structure to adapt to the current task requirements. Through transfer learning, the model can make full use of the existing marine disaster prediction experience, optimize the initial prediction results of typhoon paths, storm surges, and red tides, and generate a more accurate set of optimized prediction results.Uncertainty analysis and confidence calculation are performed on the optimized prediction result set. By analyzing the sources of uncertainty in model predictions, such as insufficient data or model bias, confidence intervals are calculated for each prediction result to quantify the reliability of the prediction results. The prediction results containing confidence intervals are input into the integrated learning module for multi-model fusion and weight optimization. The integrated learning module integrates the prediction results of multiple models through weighted voting, stacked generalization and other methods, and uses weight optimization technology to dynamically adjust the contribution of each model to generate a more robust integrated prediction result. On this basis, model integration and parameter calibration are performed on the integrated prediction results and optimized prediction result sets to improve the prediction accuracy and applicability, and finally a marine disaster prediction model set is obtained, which can fully cover the prediction needs of three types of disasters: typhoon path, storm surge and red tide.

[0023] In this embodiment, the weight parameters in the pre-trained marine disaster prediction model library are input into the feature mapping layer of the transfer learning network. The feature mapping layer constructs a shared feature space through general feature analysis. The shared space is designed to capture the common characteristics of three types of disaster prediction tasks: typhoon path, storm surge and red tide, to ensure that the existing knowledge of the pre-trained model can be maximized in the subsequent model optimization process. Based on the shared feature space, the initial prediction results of the typhoon path are mapped to feature vectors. The feature mapping layer uses a 3-layer fully connected network structure for this task. Each layer of the fully connected network processes nonlinear features through the LeakyReLU activation function, which can effectively avoid the gradient disappearance problem while maintaining sensitivity to feature distribution. Through this mapping method, the mapping features of the typhoon path are generated, which are characterized by retaining the core laws of the path dynamics and closely aligning with the shared feature space, providing a unified feature expression for subsequent cross-task migration. The initial prediction results of the storm surge are mapped to feature vectors. The feature mapping layer uses a 3-layer attention mechanism-enhanced convolutional network for the spatiotemporal correlation characteristics of the storm surge. Each layer of convolutional network dynamically adjusts the degree of attention to different feature channels through the attention mechanism, and combines batch normalization and ReLU activation function to achieve stability and efficiency of the training process. This structure can not only extract the spatiotemporal distribution characteristics of storm surge tidal data, but also strengthen the key area characteristics through the attention mechanism, so that the storm surge mapping features have higher expressiveness and distinguishing ability. For the initial prediction results of red tide, the feature mapping layer adopts a three-layer graph neural network (GNN) structure. Each layer of GNN aggregates red tide related features through the GraphSAGE aggregation function to generate red tide mapping features. The GraphSAGE function can use neighborhood information in the process of node feature aggregation, which is suitable for processing red tide, a phenomenon with complex spatial distribution and mutual influence characteristics. After completing the feature mapping, the typhoon path mapping features, storm surge mapping features, and red tide mapping features are input into the task adaptation layer for cross-task knowledge transfer. The task adaptation layer uses an adversarial learning mechanism to design a structure in which the generator and the discriminator confront each other. The generator generates a unified migration feature representation, and the discriminator guides the generator to optimize the migration process by distinguishing the feature distribution of different tasks. This mechanism effectively reduces the distribution difference between the source task and the target task, enables the transfer feature representation to be shared among multiple tasks, and retains the key features of each task. The transfer feature representation is subjected to task relevance analysis and domain adaptive training. The adversarial discriminator structure in the task adaptation layer discriminates the distribution of the transfer features, identifies the features that are specifically related to each task, and further optimizes the distribution of these features to better adapt to the data characteristics of the target task. Through this process, task-specific features are obtained, which are exclusive high-dimensional representations of each task and have the shared knowledge required in the transfer learning process. Based on the task-specific features, the parameters in the pre-trained marine disaster prediction model library are fine-grainedly tuned and gradient updated.In this process, the transfer learning network optimizes the weight parameters of the pre-trained model through back propagation to adapt to the current data distribution and prediction requirements. During the tuning process, an adaptive gradient descent algorithm (such as Adam) is used to accelerate parameter updates, and a learning rate decay strategy is used to ensure the stability of the optimization process. Through parameter updates and model reconstruction, an optimized prediction result set is generated. This result set combines the rich knowledge of the pre-trained model with the specific data characteristics of the target task, significantly improving the prediction accuracy and robustness of typhoon paths, storm surges, and red tides.

[0024] In the embodiment of the present application, by constructing an integrated sea, air and shore monitoring network, unified collection and transmission of multi-source heterogeneous ocean data is achieved, the problems of scattered data acquisition and non-uniform format are solved, and the periodic deterministic sampling algorithm and data standardization processing are adopted to effectively reduce the noise interference in the data collection process and ensure the data quality and timeliness; a digital model of the ocean environment is established based on digital twin technology, which accurately depicts the dynamic correlation between ocean environmental elements and enhances the cognitive ability of the laws of ocean environmental changes; the combination of selective caching mechanism and deep learning algorithm optimizes the efficiency of computing resource utilization and improves the accuracy of feature fusion; through transfer learning methods and multi-model integration technology, the knowledge of pre-trained models is fully utilized to enhance the generalization performance of different types of marine disaster prediction models; a complete model training and optimization process is established to realize the intelligent processing of the entire process from data collection to disaster prediction, which significantly improves the prediction accuracy and timeliness; the distributed computing architecture and edge computing technology are adopted to optimize the allocation and scheduling of computing resources and meet the real-time requirements of massive data processing.

[0025] In a specific embodiment, the process of executing step 100 may specifically include the following steps: The monitoring area is divided into three levels: ocean, air and shore, to obtain monitoring grid units. The density of marine environment monitoring equipment is optimized according to the spatial distribution of monitoring grid units to obtain a monitoring equipment deployment plan. Allocate functions of data acquisition module, data preprocessing module and communication module to each monitoring device in the monitoring device deployment plan to obtain device function configuration information; Input the device function configuration information into the network topology generator to build a multi-level network connection to obtain a network topology diagram, and deploy the communication relay equipment in different areas and configure the communication protocol according to the network topology diagram to obtain a data transmission network; Deploy a lightweight prediction model on the edge computing server of the data processing device in the data transmission network to obtain a local data processing unit, and based on the local data processing unit, distribute and allocate computing resources on the data fusion model of the cloud computing center to obtain a central computing processing unit; The data transmission network, local data processing unit and central computing processing unit are systematically integrated and the protocols are unified to obtain an integrated sea, air and shore monitoring network.

[0026] Specifically, in view of the complexity of the monitoring area, a three-dimensional grid division method is used to divide the area into three levels: ocean, air, and shore. Assume that the ocean range of the monitoring area is a two-dimensional rectangular area , whose boundaries are and Describes the vertical layer height range Subdividing the aerial monitoring layer, the three-dimensional grid division is expressed as: Number of grid cells ; in, is the spatial resolution of the 2D grid, is the resolution in the vertical height direction. After the grid division is completed, the density of the marine environment monitoring equipment is optimized according to the spatial distribution and regional characteristics of the monitoring grid units. The density optimization layout is achieved through the following objective function: ; in, is the total number of grid cells, For the The importance weight of each unit (for example, the grid close to the typhoon path has a greater weight), For the The distance between each unit and the nearest monitoring device. During the optimization process, adjust the layout of the equipment to ensure that more equipment is deployed in high-weight areas and minimize the coverage redundancy between equipment. After the equipment is deployed, assign three functions to each device: data acquisition module, data preprocessing module and communication module. The data acquisition module is responsible for collecting basic environmental parameters such as water temperature, salinity, and air pressure; the data preprocessing module formats, removes outliers, and compresses the collected data; the communication module uploads the data to the relay device through the wireless communication protocol. Assume that the device The amount of data is , the data acquisition rate is , the processing delay is , then the transmission delay of the device It is expressed as: ; After completing the device function assignment, the device function configuration information is input into the network topology generator, and a multi-level network topology diagram is constructed through hierarchical design. Communication relay equipment and monitoring devices, by minimizing the total communication cost Implementing regional deployment: ; in, Indicates a relay device With monitoring equipment The transmission cost between (such as power consumption, delay), Is the connection indicator variable. If the device With relay equipment Connect, then , otherwise 0. By solving this optimization problem, the regional deployment of communication relay equipment is completed, and the communication protocol is configured in combination with the actual network requirements. For example, the ocean buoy uses the LoRa protocol, while the aerial layer equipment uses satellite communication. In the data transmission network, in order to improve the data processing efficiency, a lightweight prediction model is deployed on the edge computing server to form a local data processing unit. Assume that the computing power of the edge server is , the local model calculation requirement is , the following constraints must be met: ; The edge server can predict short-term path changes in real time and transmit the results to the cloud computing center. The cloud computing center deploys the data fusion model and processes the global prediction task through distributed computing resource allocation. Assume that the number of cloud computing nodes is , the computing power of each node is The total amount of tasks is , the distributed resource allocation strategy needs to meet the following requirements: ; By coordinating the task allocation of the cloud computing center and the edge computing server, a central computing processing unit is formed. The data transmission network, local data processing unit and central computing processing unit are systematically integrated through unified communication protocols and interface standards to form an integrated sea, air and shore monitoring network. This network can achieve efficient linkage from data collection to processing, and support real-time monitoring of the marine environment and disaster warning.

[0027] The device function configuration information is input into the network topology generator to construct a multi-level network connection to obtain a network topology diagram, and the communication relay device is deployed in different regions and the communication protocol is configured according to the network topology diagram to obtain a data transmission network, including: state space modeling of the data acquisition module, the data preprocessing module and the communication module in the device function configuration information to obtain an agent environment state set; quantitative calculation of the communication resource consumption, the data transmission delay and the network load is performed according to the environment state set to obtain an action reward function; the environment state set and the action reward function are input into the deep Q network for policy iteration training to obtain a network topology optimization strategy; hierarchical connection construction of the communication links between devices is performed according to the network topology optimization strategy to obtain an initial network topology structure; the initial network topology structure is topologically optimized and updated by a multi-agent collaborative learning algorithm to obtain a network topology diagram; based on the network topology diagram, a density clustering algorithm is used to plan the layout of the communication relay device in different regions to obtain the deployment location of the relay device; according to the deployment location of the relay device, the communication protocol parameters are adaptively configured by a dynamic programming algorithm to obtain a protocol configuration scheme; the protocol configuration scheme is input into the network simulator for performance evaluation and parameter fine-tuning to obtain a data transmission network.

[0028] In a specific embodiment, the process of executing step 200 may specifically include the following steps: The data collection time slots of the marine environment monitoring equipment in the integrated sea-air-shore monitoring network are periodically divided and prioritized to obtain the sampling scheduling sequence; According to the sampling scheduling sequence, the original data collected by the monitoring equipment is subjected to error calculation and threshold judgment to obtain the data credibility index, and the data with a data credibility index less than the preset threshold is input into the abnormal processing module for marking and elimination to obtain a valid data set; Based on the time distribution characteristics of the effective data set, the data sampling period is dynamically adjusted and synchronously controlled to obtain the time series calibration data, and the time series calibration data is cleaned, normalized and numerically quantified to obtain standard format data; The standard format data is geographically encoded and time-stamped according to the spatial distribution of the monitoring equipment to obtain spatiotemporal correlation data, and the data change rate and transmission delay between adjacent sampling points are calculated based on the spatiotemporal correlation data to obtain data timeliness parameters; The spatiotemporal correlation data and data timeliness parameters are encapsulated in data packets and encoded in transmission protocols to obtain marine environmental data with spatiotemporal identification.

[0029] Specifically, the data collection time slots of the marine environment monitoring equipment are periodically divided and prioritized to ensure that high-priority monitoring tasks can obtain resource allocation first. Assume that the monitoring equipment set is , the importance weight of its collection task is , the collection time is The total acquisition period is , then the acquisition time slot of each device The following optimization goals are met: ; in, is a weight set according to the urgency of the task or the importance of the data. For example, in the typhoon path prediction task, the monitoring equipment close to the typhoon center area will be given a higher weight to prioritize the collection and processing of its data. According to the optimized sampling scheduling sequence, the error calculation and threshold judgment are performed on the raw data collected by the monitoring equipment to evaluate the credibility of the data. Assume that the raw data value is The standard value of the data is , the error calculation formula is: ; in, Indicates The relative error of the data points. Exceeding the preset threshold , the data point is considered unreliable and is marked and removed. Through this process, abnormal data is effectively removed and a valid data set with higher credibility is obtained. Based on the time distribution characteristics of the effective data set, the data sampling period is dynamically adjusted and synchronously controlled. Assume that the sampling period is , the temporal distribution of the data is characterized by its autocorrelation function describe: ; in, is the time lag, For time By analyzing to identify the main periodic characteristics of the data and dynamically adjust The adjusted data is synchronized and controlled to ensure that the data timestamps of different devices are aligned, generating time-series calibration data. Based on the time-series calibration data, data cleaning, normalization, and numerical quantization are performed to generate standard format data. Data cleaning ensures the integrity and consistency of the data by removing missing values ​​and outliers; normalization maps the data to a unified range [0, 1], and the formula is: ; in, and are the minimum and maximum values ​​of the data respectively; numerical quantization discretizes continuous data into fixed-precision values ​​to reduce storage and computational burdens. The standard format data is geo-coded and timestamped according to the spatial distribution of the monitoring equipment to generate spatiotemporal correlation data. Assume that the monitoring equipment is located at the longitude and latitude coordinates , and its collection time is , then each piece of data is marked as a triple Based on these spatiotemporal correlation data, the data change rate and transmission delay between adjacent sampling points are calculated. It is expressed as: ; The transmission delay It is expressed as: ; in, and They are the timestamps of data reception and transmission, respectively. These parameters reflect the timeliness and dynamic change characteristics of the data. The spatiotemporal correlation data and data timeliness parameters are encapsulated in data packets and encoded in transmission protocols to generate marine environmental data with spatiotemporal identification. Data packet encapsulation groups data through standard network protocols (such as TCP / IP or MQTT) and adds necessary verification information and header information; transmission protocol encoding selects appropriate communication methods (such as LoRa or satellite communication) according to different devices and network levels to ensure the reliability and efficiency of data transmission.

[0030] In a specific embodiment, the process of executing step 300 may specifically include the following steps: The marine environment data with time and space labels are classified and digitally encoded according to the water temperature field, salinity field, sea level height field, wind speed field and air pressure field to obtain the marine element data structure; Map the time series and spatial distribution of physical quantities in the marine element data structure to obtain the temporal and spatial distribution characteristics of marine environmental elements; Based on the temporal and spatial distribution characteristics, the interaction relationship between marine environmental elements is coupled analyzed and the correlation degree is calculated to obtain the marine element correlation matrix; The marine element correlation matrix is ​​input into the dynamic evolution model to analyze the law of marine environmental state changes, and the environmental evolution sequence is obtained. The stability analysis and mutation feature detection of the environmental evolution sequence are performed to obtain the characteristics of abnormal marine environmental events. Based on the characteristics of abnormal events in the marine environment, a mapping function between the physical marine environment and the digital space is established to obtain a digital twin mapping relationship, and the digital twin mapping relationship is input into a feature extractor to extract key features of the marine environment to obtain a marine environment feature vector; The model integration and parameter optimization of the marine environment characteristic vector and digital twin mapping relationship are carried out to obtain the digital model of the marine environment.

[0031] Specifically, the marine environment data with time and space labels are classified into physical elements according to water temperature field, salinity field, sea level height field, wind speed field and air pressure field. By extracting the physical attributes of the data and grouping them according to their scientific significance, such as water temperature field Describing a specific time and spatial coordinates The seawater temperature distribution and salinity field Describe the salinity distribution in the same space and time, etc. Digitally encode each type of data and convert it into a unified ocean element data structure, which is expressed by the following formula: ; in, represent discrete values ​​of water temperature, salinity, sea level, wind speed and air pressure, respectively. and It is the discretized dimension of time and space. After completing the digital coding, the physical quantities in the marine element data structure are mapped in time series and space distribution to reveal the spatiotemporal distribution characteristics of marine environmental elements. Time series mapping captures long-term trends, periodic fluctuations and mutation points by analyzing continuous values ​​in the time dimension. Its mathematical expression is: ; in, It is a time lag The corresponding autocorrelation function is, Represents the time series Spatial distribution mapping constructs a continuous spatial distribution map through interpolation technology and distribution fitting methods, such as the spatial prediction model based on Gaussian process regression: ; in, is the physical value of the prediction point, is the Gaussian kernel function, As weights. Based on the temporal and spatial distribution characteristics, the interaction relationship between marine environmental elements is coupled and the correlation is calculated. Taking water temperature and salinity as an example, the coupling analysis quantifies the causal relationship between them through Granger causality test or mutual information method. Mutual information The formula is: ; in, represents the joint probability distribution, and are marginal probability distributions respectively. The ocean element correlation matrix generated by correlation calculation It can describe the strength of the relationship between elements, where Indicates and The correlation matrix is ​​input into the dynamic evolution model to analyze the changing law of the marine environment state. The dynamic evolution model describes the spatiotemporal evolution of the system through a set of differential equations. For example, partial differential equations are used to describe the diffusion of the water temperature field: ; in, is the diffusion coefficient, It is an external heat source. Through numerical simulation, the environmental evolution sequence can be generated, and the stability of the evolution sequence can be analyzed through eigenvalue analysis and singular spectrum analysis to extract possible mutation characteristics. Based on the mutation characteristics of environmental evolution, a mapping function between the physical ocean environment and the digital space is established to generate a digital twin mapping relationship. This mapping is fitted through a multivariate regression model or a deep learning network, for example, a multi-layer perceptron (MLP) model is used to establish a nonlinear mapping: ; in, is the weight matrix, is the input feature matrix, is the bias, is a nonlinear activation function. The generated digital twin mapping relationship is input into the feature extractor to further extract key features and generate a high-dimensional marine environment feature vector. Model integration and parameter optimization are performed on the marine environment feature vector and the digital twin mapping relationship. Integration methods such as random forests or gradient boosting decision trees improve prediction performance through multi-model fusion, and parameter optimization uses Bayesian optimization or particle swarm optimization algorithms to ultimately obtain the best-performing marine environment digital model that fully reflects the spatiotemporal dynamic characteristics of the marine environment.

[0032] In a specific embodiment, the process of executing step 400 may specifically include the following steps: The characteristics of marine environmental elements in the marine environmental digital model are classified and constructed according to data types and spatiotemporal distribution to obtain a characteristic data hierarchy, and the historical data characteristics are cached and prioritized based on the characteristic data hierarchy to obtain a hierarchical cache database; Input the feature data in the hierarchical cache database into a deep neural network for potential feature extraction and nonlinear mapping to obtain a deep feature representation; Perform multimodal feature fusion and complementarity analysis on the spatial features, temporal features and physical features in the deep feature representation to obtain a fused feature vector; The fused feature vector is input into the attention calculation unit for feature weight assignment and saliency calculation to obtain a weighted feature set. The weighted feature set is then used to establish a feature organization structure and index mechanism according to the time and space dimensions to obtain a spatiotemporal feature organization system. Perform similarity measurement and correlation mining on the feature data in the spatiotemporal feature organization system to obtain a feature correlation map; The feature association map is input into the deep spatiotemporal fusion network for cross-modal feature learning and spatiotemporal dependency modeling to obtain a fused feature tensor, which is then subjected to dimensionality reduction and feature selection to obtain spatiotemporal feature data.

[0033] Specifically, the characteristics of marine environmental elements in the marine environmental digital model are classified according to data type and spatiotemporal distribution, and a hierarchical structure of feature data is constructed. This process groups different types of data based on the physical properties and statistical characteristics of the data. For example, water temperature and salinity are classified as continuous spatiotemporal characteristics, sea level height is classified as event-driven characteristics, and wind speed and air pressure are classified as dynamic meteorological characteristics. Based on these classifications, feature data are prioritized according to their importance in spatiotemporal distribution, and a hierarchical cache database is constructed by optimizing the following objective function: ; in, is the total number of feature data, For the The importance weight of the feature, is the data update frequency, The amount of data with features. Through this optimization function, the data in the cache is managed hierarchically so that high-priority data can get higher cache resource allocation. The feature data in the hierarchical cache database is input into the deep neural network for potential feature extraction and nonlinear mapping. Assume that the input data is ,in To input feature dimensions, the deep neural network implements feature abstraction through multiple fully connected layers and activation functions. The output of each layer is expressed as: ; in, It is The output of the layer, is the weight matrix, is the bias vector, is a nonlinear activation function (such as ReLU). The final output deep feature representation It contains the potential features of the data and has higher expressive power. Multimodal feature fusion and complementary analysis are performed on the spatial features, temporal features and physical features in the deep feature representation. Assuming that the spatial feature is , the time series characteristics are , the physical characteristics are , and its fusion is achieved through weighted summation: ; in, is the weight coefficient, which reflects the importance of different modal features. Through complementary analysis, the synergy of these features is enhanced to obtain the fusion feature vector The fused feature vector is input into the attention calculation unit for feature weight assignment and saliency calculation. The attention mechanism calculates the importance score of each feature. To dynamically allocate weights, the calculation formula is: ; in, is the score function, and are the learnable parameters, For the feature vectors. Through the attention mechanism, a weighted feature set is generated, and a feature organization structure and index mechanism are established based on the time dimension and space dimension to form a spatiotemporal feature organization system. Based on the spatiotemporal feature organization system, similarity measurement and correlation mining are performed on the feature data. Similarity measurement is achieved through the cosine similarity formula: ; Relevance mining is achieved by constructing a graph structure, where the adjacency matrix of the graph Elements Representation characteristics and The strength of the association between them. Finally, a feature association map is generated to describe the complex relationship between features. The feature association map is input into the deep spatiotemporal fusion network for cross-modal feature learning and spatiotemporal dependency modeling. Assume that the input map is ,in is a node set, is the adjacency matrix. The deep spatiotemporal fusion network captures spatial features through the graph convolution layer and captures temporal dependencies through the recurrent neural network. The graph convolution operation formula is: ; in, For the Layer node features, is the weight matrix, is the activation function. Time series modeling is implemented through LSTM, and its update formula is: ; Through the above operations, a fused feature tensor is generated. The fused feature tensor is subjected to dimensionality reduction and feature selection, such as principal component analysis or recursive feature elimination method, to extract the most important features, and finally obtain high-quality spatiotemporal feature data.

[0034] In a specific embodiment, the process of executing step 500 may specifically include the following steps: The spatiotemporal characteristic data are divided into typhoon track prediction dataset, storm surge prediction dataset and red tide prediction dataset; A typhoon path prediction model is constructed based on the typhoon path prediction dataset. The typhoon path prediction model includes an encoder network and a decoder network. The encoder network uses a 5-layer convolution structure to extract typhoon motion features. Each convolution structure uses a ReLU activation function and a batch normalization layer. The decoder network uses a 5-layer deconvolution structure to generate path prediction results, and a jump connection is set between the encoder network and the decoder network to obtain the initial prediction result of the typhoon path. A storm surge prediction model is established for the storm surge prediction data set. The storm surge prediction model includes a temporal feature extraction network and a spatial feature fusion network. The temporal feature extraction network uses a bidirectional long short-term memory network structure to model tide level changes. The spatial feature fusion network uses a graph convolutional network structure to perform correlation analysis on multi-point tide level data to obtain the initial storm surge prediction results. A red tide prediction model is constructed based on the red tide prediction dataset. The red tide prediction model includes a multi-scale feature extraction module and a spatiotemporal attention module. The multi-scale feature extraction module uses three parallel residual networks to process features of different scales. The spatiotemporal attention module focuses on analyzing key areas and time periods to obtain the initial prediction results of red tides. The pre-trained marine disaster prediction model library is input into the transfer learning network for knowledge transfer. The transfer learning network includes a feature mapping layer and a task adaptation layer. The model is optimized for the initial prediction results of the typhoon path, the initial prediction results of the storm surge, and the initial prediction results of the red tide to obtain an optimized prediction result set. Perform uncertainty analysis and confidence calculation on the optimized prediction result set to obtain the model prediction confidence interval; The model prediction confidence interval is input into the integrated learning module for multi-model fusion and weight optimization to obtain the integrated prediction result. The integrated prediction result and the optimized prediction result set are then model integrated and parameter calibrated to obtain the marine disaster prediction model set.

[0035] Specifically, the spatiotemporal feature data are divided into three sub-datasets, which are used for typhoon path prediction, storm surge prediction, and red tide prediction, respectively. Among them, the typhoon path prediction dataset includes data on the temporal changes of the typhoon center position (longitude and latitude), the storm surge prediction dataset covers multi-point tide observations and meteorological conditions, and the red tide prediction dataset contains spatial distribution data of water temperature, salinity, and nutrient concentration. This division method ensures that each dataset has sufficient expressive power for the corresponding prediction task. Based on the typhoon path prediction dataset, a typhoon path prediction model is constructed. The model adopts an encoder-decoder structure, in which the encoder network contains a 5-layer convolution structure and the decoder network contains a 5-layer deconvolution structure. Assume that the input data is , the convolution process is expressed as: ; in, It is The output feature map of the layer, and are the convolution kernel and bias parameters, respectively. * indicates the convolution operation, BN is batch normalization, and ReLU is the activation function. After the encoder extracts the typhoon motion features, the decoder generates the path prediction result through deconvolution: ; in, is the output feature map of the decoder, and are the weights and biases of deconvolution. In order to avoid information loss, a jump connection is set between the encoder and the decoder so that the high-resolution features are directly passed to the decoder layer to generate the initial prediction results of the typhoon path. For the storm surge prediction dataset, a storm surge prediction model is constructed. The model includes a time series feature extraction network and a spatial feature fusion network. The time series feature extraction network uses a bidirectional long short-term memory network (Bi-LSTM) to model the time dependency of tide level changes. Its state update formula is: ; in, They are forget gate, input gate and output gate respectively. is the cell state, is the hidden state, is the Sigmoid activation function, Represents element-by-element product. The spatial feature fusion network uses a graph convolutional network (GCN) to capture the spatial correlation of tide data through the following formula: ; in, For the The node feature matrix of the layer, is the adjacency matrix, is the weight matrix, is the activation function. Combining temporal and spatial features, the initial prediction results of storm surge are generated. For the red tide prediction dataset, a red tide prediction model is constructed. The model includes a multi-scale feature extraction module and a spatiotemporal attention module. The multi-scale feature extraction module uses 3 parallel residual networks, and the output of each residual network is: ; in, is the input feature, represents the nonlinear transformation through multiple convolutional layers, is a set of weights. The spatiotemporal attention module dynamically assigns weights using the following formula: ; in, is the attention weight, Score for attention, is the input feature, is a learnable parameter. Combined with multi-scale and attention mechanisms, the initial prediction results of red tide are generated. The above initial prediction results are input into the transfer learning network for knowledge transfer. The transfer learning network contains a feature mapping layer and a task adaptation layer. The feature mapping layer maps the general features of the pre-trained model to the current task through the following formula: ; in, and is the mapping layer parameter, is the input feature. The task adaptation layer uses an adversarial learning mechanism to optimize task-specific features and realizes knowledge transfer through interactive training of the generator and the discriminator. Uncertainty analysis and confidence calculation are performed on the optimized prediction result set. The uncertainty of the prediction result is estimated by Monte Carlo sampling, and the confidence interval calculation formula is: ; in, is the predicted mean, is the standard deviation, is the confidence coefficient, is the number of samples. Through uncertainty analysis, the confidence interval of the prediction result is generated. The confidence interval is input into the ensemble learning module, and multi-model fusion is achieved through weight optimization. The prediction result of ensemble learning is: ; in, For the The prediction results of the model are is the weight, is the number of models. Through model integration and parameter calibration, a set of marine disaster prediction models is finally generated.

[0036] In a specific embodiment, the execution step inputs the pre-trained marine disaster prediction model library into the transfer learning network for knowledge transfer, the transfer learning network includes a feature mapping layer and a task adaptation layer, and the model is optimized for the initial prediction results of the typhoon path, the initial prediction results of the storm surge, and the initial prediction results of the red tide. The process of obtaining the optimized prediction result set can specifically include the following steps: The weight parameters in the pre-trained marine disaster prediction model library are input into the feature mapping layer for common feature analysis to obtain a shared feature space; Based on the shared feature space, the initial prediction results of the typhoon path are mapped to feature vectors. The feature mapping layer adopts a 3-layer fully connected network structure. Each layer of the fully connected network structure uses the LeakyReLU activation function to obtain the typhoon path mapping features. The initial storm surge prediction results are mapped to feature vectors. The feature mapping layer uses a 3-layer attention mechanism-enhanced convolutional network. Each layer of the attention mechanism-enhanced convolutional network uses batch normalization and ReLU activation function to obtain storm surge mapping features. Based on the initial prediction results of red tide, feature vector mapping is performed. The feature mapping layer adopts a three-layer graph neural network structure. Each layer of the graph neural network structure uses the GraphSAGE aggregation function to obtain the red tide mapping features. Typhoon path mapping features, storm surge mapping features and red tide mapping features are input into the task adaptation layer for cross-task knowledge transfer. The task adaptation layer adopts an adversarial learning mechanism to obtain the transfer feature representation. Perform task relevance analysis and domain adaptation training on the transferred feature representation. The task adaptation layer uses an adversarial discriminator structure to obtain task-specific features. Based on the task-specific features, the parameters in the pre-trained marine disaster prediction model library are fine-grained tuned and gradient updated to obtain the model optimization parameters, and the parameters are updated and the model is reconstructed based on the model optimization parameters to obtain the optimized prediction result set.

[0037] Specifically, the weight parameters in the pre-trained marine disaster prediction model library are input into the feature mapping layer. The main function of the feature mapping layer is to analyze the common features shared in the model and construct a high-dimensional shared feature space. This process is achieved through the following formula: ; in, is a shared feature representation, is the input feature of the pre-trained model, and are the weight matrix and bias vector respectively, Represents a nonlinear activation function (such as ReLU). After extracting common features through the feature mapping layer, the constructed shared feature space can capture the basic knowledge across tasks in the pre-trained model. Based on the shared feature space, the initial prediction results of the typhoon path are mapped to feature vectors. The feature mapping layer designs a 3-layer fully connected network for the typhoon path task, and the output of each layer is expressed as: ; in, It is The output feature vector of the layer, and are the weight matrix and bias vector respectively. LeakyReLU is a ReLU activation function with leakage parameters, which is used to avoid the gradient vanishing problem. Through three layers of nonlinear mapping, the typhoon path mapping features are finally generated, retaining the high-order representation of the path dynamic features. For the initial storm surge prediction results, the feature mapping layer uses a 3-layer attention mechanism enhanced convolutional network, and the output of each layer is expressed as: ; Among them, * represents the convolution operation, BN is batch normalization, is the attention weight, calculated as: ; in, and are the attention weights and bias parameters, is the score vector, It is Features. By combining convolution operation and attention mechanism, storm surge mapping features are generated, which can accurately capture the spatial correlation in tide distribution. For the initial prediction results of red tide, the feature mapping layer adopts a 3-layer graph neural network structure (GNN), and each layer uses GraphSAGE aggregation function to update node features. The formula is: ; in, Representation Node In the The characteristics of the layer, Is a node The neighbor set of , AGGREGATE is the feature aggregation function (such as averaging or maximum value), and are weights and biases. The red tide mapping features are generated through three-layer iterative updates of the graph structure, which can effectively capture the complex spatial relationship of the red tide occurrence area. The typhoon path mapping features, storm surge mapping features, and red tide mapping features are input into the task adaptation layer for cross-task knowledge transfer. The task adaptation layer achieves feature alignment through the adversarial learning mechanism and generates a migration feature representation. Adversarial learning consists of two parts: the generator and the discriminator. The goal of the generator is to generate a representation consistent with the distribution of target task features. Its optimization goal is: ; in, is the discriminator, is a generator, and They are data distribution and noise distribution respectively. Through adversarial optimization, the transferred feature representation can retain both general knowledge and task-specific characteristics. The transferred feature representation is subjected to task relevance analysis and domain adaptive training. Task relevance is measured by calculating the Kullback-Leibler divergence (KL divergence) of the feature distribution: ; in, and is the feature distribution of the source task and the target task. Domain adaptive training further optimizes the adversarial discriminator structure to make the source task feature distribution more consistent with the target task feature distribution, thereby generating task-specific features. Based on the task-specific features, the parameters in the pre-trained model library are fine-grained tuned and gradient updated. Parameter tuning updates the weights through back propagation, and the formula is: ; in, are model parameters, is the learning rate, is the loss function. The tuned model parameters are used for model reconstruction, and finally an optimized prediction result set is generated.

[0038] The above describes the multi-source heterogeneous ocean data intelligent fusion and marine disaster prediction method in the embodiment of the present application. The following describes the multi-source heterogeneous ocean data intelligent fusion and marine disaster prediction platform 10 in the embodiment of the present application. Figure 2 In the embodiment of the present application, an embodiment of the multi-source heterogeneous ocean data intelligent fusion and ocean disaster prediction platform 10 includes: Construction module 11 is used to hierarchically deploy marine environment monitoring equipment, communication relay equipment and data processing equipment and construct network topology to obtain an integrated sea, air and shore monitoring network; The acquisition module 12 is used to collect and standardize the water temperature, salinity, sea level, wind speed and air pressure data according to the integrated sea, air and shore monitoring network using a periodic deterministic sampling algorithm to obtain marine environmental data with time and space identification; The feature extraction module 13 is used to perform ocean element modeling and dynamic feature extraction based on the ocean environment data with time and space identification to obtain a digital model of the ocean environment; The correlation analysis module 14 is used to perform feature fusion and correlation analysis on multi-source heterogeneous ocean data based on the digital model of the ocean environment by using a selective caching mechanism and a deep learning algorithm to obtain spatiotemporal feature data; The model integration module 15 is used to train and integrate the typhoon path prediction model, storm surge prediction model and red tide prediction model according to the spatiotemporal feature data using the transfer learning method to obtain a set of marine disaster prediction models.

[0039] Through the collaborative cooperation of the above-mentioned components and the construction of an integrated sea, air and shore monitoring network, the unified collection and transmission of multi-source heterogeneous ocean data has been achieved, solving the problems of scattered data acquisition and non-uniform formats. The periodic deterministic sampling algorithm and data standardization processing have been adopted to effectively reduce noise interference in the data collection process and ensure data quality and timeliness. Based on the digital twin technology, a digital model of the marine environment is established to accurately depict the dynamic correlation between marine environmental elements and enhance the ability to understand the laws of marine environmental changes. The combination of selective caching mechanism and deep learning algorithm optimizes the efficiency of computing resources and improves the accuracy of feature fusion. Through transfer learning methods and multi-model integration technology, the knowledge of pre-trained models is fully utilized to enhance the generalization performance of different types of marine disaster prediction models. A complete model training and optimization process has been established to realize intelligent processing of the entire process from data collection to disaster prediction, significantly improving the prediction accuracy and timeliness. The distributed computing architecture and edge computing technology are adopted to optimize the allocation and scheduling of computing resources and meet the real-time requirements of massive data processing.

[0040] See also Figure 3 , Figure 3 This is a schematic block diagram of the structure of an electronic device 300 provided in an embodiment of the present application. The electronic device 300 includes a processor 301 and a memory 302. The processor 301 and the memory 302 are connected via a platform bus 303, wherein the memory 302 may include a non-volatile storage medium and an internal memory.

[0041] The non-volatile storage medium can store a computer program. The computer program includes program instructions, and when the program instructions are executed by the processor 301, the processor 301 can execute any of the above-mentioned multi-source heterogeneous ocean data intelligent fusion and ocean disaster prediction methods.

[0042] The processor 301 is used to provide computing and control capabilities to support the operation of the entire electronic device 300 .

[0043] The internal memory provides an environment for the operation of the computer program in the non-volatile storage medium. When the computer program is executed by the processor 301, the processor 301 can execute any of the above-mentioned multi-source heterogeneous ocean data intelligent fusion and ocean disaster prediction methods.

[0044] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a partial structure related to the present application scheme, and does not constitute a limitation on the electronic device 300 involved in the present application scheme. The specific electronic device 300 may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0045] It should be understood that the processor 301 may be a central processing unit (CPU), and the processor 301 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0046] It should be noted that technicians in the relevant field can clearly understand that for the convenience and simplicity of description, the specific working process of the electronic device 300 described above can refer to the corresponding process of the aforementioned multi-source heterogeneous ocean data intelligent fusion and marine disaster prediction method, and will not be repeated here.

[0047] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by one or more processors, the one or more processors implement the multi-source heterogeneous ocean data intelligent fusion and ocean disaster prediction method provided in the embodiment of the present application.

[0048] The computer-readable storage medium may be an internal storage unit of the electronic device 300 in the aforementioned embodiment, such as a hard disk or memory of the electronic device 300. The computer-readable storage medium may also be an external storage device of the electronic device 300, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped with the electronic device 300.

[0049] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, platforms and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0050] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable an electronic device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.

[0051] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for intelligent fusion of multi-source heterogeneous ocean data and prediction of marine disasters, characterized in that: The method comprises: Hierarchical deployment and network topology construction of marine environment monitoring equipment, communication relay equipment and data processing equipment to obtain an integrated sea, air and shore monitoring network; According to the integrated sea-air-shore monitoring network, a periodic deterministic sampling algorithm is used to collect and standardize water temperature, salinity, sea level, wind speed and air pressure data to obtain marine environmental data with time and space identification; Based on the marine environment data with time and space identification, marine element modeling and dynamic feature extraction are performed to obtain a digital model of the marine environment; Based on the digital model of the ocean environment, a selective caching mechanism and a deep learning algorithm are used to perform feature fusion and correlation analysis on multi-source heterogeneous ocean data to obtain spatiotemporal feature data; According to the spatiotemporal characteristic data, the typhoon path prediction model, the storm surge prediction model and the red tide prediction model are trained and integrated by the transfer learning method to obtain a marine disaster prediction model set.

2. The method for intelligent fusion of multi-source heterogeneous ocean data and prediction of marine disasters according to claim 1 is characterized in that: The marine environment monitoring equipment, communication relay equipment and data processing equipment are hierarchically deployed and the network topology is constructed to obtain an integrated sea, air and shore monitoring network, including: Divide the monitoring area into three grids: ocean, air and shore-based, to obtain monitoring grid units, and optimize the density of marine environment monitoring equipment according to the spatial distribution of the monitoring grid units to obtain a monitoring equipment deployment plan; Allocate functions of a data acquisition module, a data preprocessing module and a communication module to each monitoring device in the monitoring device deployment scheme to obtain device function configuration information; Input the device function configuration information into a network topology generator to construct a multi-level network connection to obtain a network topology structure diagram, and deploy the communication relay device in different areas and configure the communication protocol according to the network topology structure diagram to obtain a data transmission network; Performing lightweight prediction model deployment on the edge computing server of the data processing device in the data transmission network to obtain a local data processing unit, and performing distributed deployment and computing resource allocation on the data fusion model of the cloud computing center based on the local data processing unit to obtain a central computing processing unit; The data transmission network, the local data processing unit and the central computing processing unit are system-integrated and protocol-unified to obtain a sea-air-shore integrated monitoring network.

3. The method for intelligent fusion of multi-source heterogeneous ocean data and prediction of marine disasters according to claim 2 is characterized in that: According to the integrated sea, air and shore monitoring network, a periodic deterministic sampling algorithm is used to collect and standardize water temperature, salinity, sea level, wind speed and air pressure data to obtain marine environmental data with time and space identification, including: Periodically dividing and prioritizing the data collection time slots of the marine environment monitoring equipment in the integrated sea-air-shore monitoring network to obtain a sampling scheduling sequence; According to the sampling scheduling sequence, error calculation and threshold judgment are performed on the raw data collected by the monitoring device to obtain a data credibility index, and the data whose data credibility index is less than a preset threshold is input into the abnormal processing module for marking and elimination to obtain a valid data set; Dynamically adjusting and synchronously controlling the data sampling period based on the time distribution characteristics of the effective data set to obtain time series calibration data, and performing data cleaning, normalization and numerical quantization processing on the time series calibration data to obtain standard format data; The standard format data is geographically encoded and time-stamped according to the spatial distribution of the monitoring equipment to obtain spatiotemporal correlation data, and the data change rate and transmission delay between adjacent sampling points are calculated based on the spatiotemporal correlation data to obtain data timeliness parameters; The spatiotemporal correlation data and the data timeliness parameters are packaged into data packets and encoded with a transmission protocol to obtain ocean environment data with spatiotemporal identification.

4. The method for intelligent fusion of multi-source heterogeneous ocean data and prediction of marine disasters according to claim 3 is characterized in that: The ocean element modeling and dynamic feature extraction based on the ocean environment data with time and space identification to obtain the ocean environment digital model includes: The marine environment data with time and space identification are classified and digitally encoded according to the water temperature field, salinity field, sea level height field, wind speed field and air pressure field to obtain a marine element data structure; Performing time series and spatial distribution mapping on the physical quantities in the ocean element data structure to obtain the temporal and spatial distribution characteristics of the ocean environment elements; Based on the temporal and spatial distribution characteristics, coupling analysis and correlation calculation are performed on the interaction relationship between the marine environmental elements to obtain a marine element correlation matrix; Inputting the marine element association matrix into the dynamic evolution model to analyze the law of marine environmental state changes, obtain an environmental evolution sequence, and perform stability analysis and mutation feature detection on the environmental evolution sequence to obtain the characteristics of marine environmental abnormal events; A mapping function between the physical ocean environment and the digital space is established based on the characteristics of the abnormal event in the ocean environment to obtain a digital twin mapping relationship, and the digital twin mapping relationship is input into a feature extractor to extract key features of the ocean environment to obtain an ocean environment feature vector; Model integration and parameter optimization are performed on the marine environment feature vector and the digital twin mapping relationship to obtain a digital model of the marine environment.

5. The method for intelligent fusion of multi-source heterogeneous ocean data and prediction of ocean disasters according to claim 4 is characterized in that: Based on the digital model of the ocean environment, a selective caching mechanism and a deep learning algorithm are used to perform feature fusion and correlation analysis on multi-source heterogeneous ocean data to obtain spatiotemporal feature data, including: The marine environment element features in the marine environment digital model are classified and constructed according to data types and spatiotemporal distribution to obtain a feature data hierarchy structure, and the historical data features are cached and prioritized based on the feature data hierarchy structure to obtain a hierarchical cache database; Inputting the feature data in the hierarchical cache database into a deep neural network for potential feature extraction and nonlinear mapping to obtain a deep feature representation; Performing multimodal feature fusion and complementarity analysis on the spatial features, temporal features and physical features in the deep feature representation to obtain a fused feature vector; Inputting the fused feature vector into the attention calculation unit for feature weight assignment and saliency calculation to obtain a weighted feature set, and establishing a feature organization structure and indexing mechanism for the weighted feature set according to the time dimension and the space dimension to obtain a spatiotemporal feature organization system; Performing similarity measurement and correlation mining on the feature data in the spatiotemporal feature organization system to obtain a feature correlation map; The feature association map is input into a deep spatiotemporal fusion network for cross-modal feature learning and spatiotemporal dependency modeling to obtain a fused feature tensor, and the fused feature tensor is subjected to dimensionality reduction and feature selection to obtain spatiotemporal feature data.

6. The method for intelligent fusion of multi-source heterogeneous ocean data and prediction of ocean disasters according to claim 5 is characterized in that: According to the spatiotemporal feature data, a transfer learning method is used to train and integrate the typhoon path prediction model, the storm surge prediction model and the red tide prediction model to obtain a marine disaster prediction model set, including: The spatiotemporal characteristic data are divided into a typhoon path prediction data set, a storm surge prediction data set and a red tide prediction data set; A typhoon path prediction model is constructed based on the typhoon path prediction dataset, wherein the typhoon path prediction model includes an encoder network and a decoder network, wherein the encoder network uses a 5-layer convolution structure to extract typhoon motion features, and each convolution structure uses a ReLU activation function and a batch normalization layer, and the decoder network uses a 5-layer deconvolution structure to generate a path prediction result, and a jump connection is set between the encoder network and the decoder network to obtain an initial typhoon path prediction result; A storm surge prediction model is established for the storm surge prediction data set. The storm surge prediction model includes a temporal feature extraction network and a spatial feature fusion network. The temporal feature extraction network uses a bidirectional long short-term memory network structure to model tide level changes. The spatial feature fusion network uses a graph convolutional network structure to perform correlation analysis on multi-point tide level data to obtain an initial storm surge prediction result. A red tide prediction model is constructed according to the red tide prediction data set, wherein the red tide prediction model includes a multi-scale feature extraction module and a spatiotemporal attention module, wherein the multi-scale feature extraction module uses three parallel residual networks to process features of different scales, and the spatiotemporal attention module performs a focused analysis on key areas and time periods to obtain an initial red tide prediction result; Inputting the pre-trained marine disaster prediction model library into the transfer learning network for knowledge transfer, the transfer learning network includes a feature mapping layer and a task adaptation layer, and optimizing the model of the initial prediction result of the typhoon path, the initial prediction result of the storm surge, and the initial prediction result of the red tide to obtain an optimized prediction result set; Performing uncertainty analysis and confidence calculation on the optimized prediction result set to obtain a model prediction confidence interval; The model prediction confidence interval is input into the integrated learning module for multi-model fusion and weight optimization to obtain an integrated prediction result, and the integrated prediction result and the optimized prediction result set are subjected to model integration and parameter calibration to obtain a marine disaster prediction model set.

7. The method for intelligent fusion of multi-source heterogeneous ocean data and prediction of ocean disasters according to claim 6 is characterized in that: The pre-trained marine disaster prediction model library is input into the transfer learning network for knowledge transfer. The transfer learning network includes a feature mapping layer and a task adaptation layer. The model is optimized for the initial prediction result of the typhoon path, the initial prediction result of the storm surge, and the initial prediction result of the red tide to obtain an optimized prediction result set, including: The weight parameters in the pre-trained marine disaster prediction model library are input into the feature mapping layer for common feature analysis to obtain a shared feature space; Based on the shared feature space, feature vector mapping is performed on the initial prediction result of the typhoon path, the feature mapping layer adopts a 3-layer fully connected network structure, and each layer of the fully connected network structure uses a LeakyReLU activation function to obtain typhoon path mapping features; Performing feature vector mapping on the initial storm surge prediction result, the feature mapping layer adopts a 3-layer attention mechanism enhanced convolutional network, each layer of the attention mechanism enhanced convolutional network uses batch normalization and ReLU activation function to obtain storm surge mapping features; Performing feature vector mapping according to the initial red tide prediction result, the feature mapping layer adopts a 3-layer graph neural network structure, and each layer of the graph neural network structure uses a GraphSAGE aggregation function to obtain red tide mapping features; Inputting the typhoon path mapping feature, the storm surge mapping feature and the red tide mapping feature into a task adaptation layer for cross-task knowledge transfer, wherein the task adaptation layer adopts an adversarial learning mechanism to obtain a transfer feature representation; Performing task relevance analysis and domain adaptation training on the migration feature representation, wherein the task adaptation layer uses an adversarial discriminator structure to obtain task-specific features; Based on the task-specific features, fine-grained tuning and gradient updating are performed on the parameters in the pre-trained marine disaster prediction model library to obtain model optimization parameters, and parameter updating and model reconstruction are performed based on the model optimization parameters to obtain an optimized prediction result set.

8. A multi-source heterogeneous ocean data intelligent fusion and ocean disaster prediction platform, characterized by: Used to execute the multi-source heterogeneous ocean data intelligent fusion and ocean disaster prediction method as described in any one of claims 1 to 7, the multi-source heterogeneous ocean data intelligent fusion and ocean disaster prediction platform comprises: The construction module is used to hierarchically deploy marine environment monitoring equipment, communication relay equipment and data processing equipment and construct network topology to obtain an integrated sea, air and shore monitoring network; A collection module is used to collect and standardize water temperature, salinity, sea level, wind speed and air pressure data according to the integrated sea, air and shore monitoring network using a periodic deterministic sampling algorithm to obtain marine environmental data with time and space identification; A feature extraction module, used to perform ocean element modeling and dynamic feature extraction based on the ocean environment data with time and space identification to obtain a digital model of the ocean environment; An association analysis module is used to perform feature fusion and association analysis on multi-source heterogeneous ocean data based on the ocean environment digital model using a selective caching mechanism and a deep learning algorithm to obtain spatiotemporal feature data; The model integration module is used to train and integrate the typhoon path prediction model, the storm surge prediction model and the red tide prediction model according to the spatiotemporal feature data using a transfer learning method to obtain a set of marine disaster prediction models.

9. An electronic device, characterized in that: The electronic device comprises: a memory and at least one processor, wherein instructions are stored in the memory; The at least one processor calls the instructions in the memory so that the electronic device executes the multi-source heterogeneous ocean data intelligent fusion and ocean disaster prediction method as described in any one of claims 1-7.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the multi-source heterogeneous ocean data intelligent fusion and ocean disaster prediction method as described in any one of claims 1 to 7 is implemented.

Images