A coastal zone ecological security assessment system and method

Abstract

The present invention relates to the field of data processing and analysis, and in particular to a coastal ecological security assessment system and method. It includes: obtaining an environmental state vector and modeling and estimating the environmental state, performing data variability analysis on the environmental state vector; formulating a data acquisition strategy in combination with the environmental state vector and data variability analysis; sampling according to the data acquisition strategy, and mapping the sampled data to a high-dimensional space and optimizing it; restoring the data in the high-dimensional space back to the original dimension; modeling the complexity and variability of the coastal environment, and predicting ecological trends; and comprehensively analyzing the environmental state vector, data variability analysis, and ecological trend prediction results to perform risk assessment and formulate response strategies. It solves the problems that the existing technology cannot accurately capture the rapid and complex changes in the coastal environment; lacks an effective mechanism to dynamically adjust the data sampling strategy; and cannot provide in-depth and accurate analysis in terms of environmental pattern recognition and trend prediction.

Description

Technical Field

[0001] The present invention relates to the field of data processing and analysis, and in particular to a coastal zone ecological security assessment system and method. Background Art

[0002] Coastal ecosystems are highly complex and dynamic, affected by a variety of natural and human factors, such as climate change, coastal construction, industrial pollution, natural disasters, etc. These factors have caused coastal ecosystems to face many challenges, such as loss of biodiversity, destruction of ecological balance and environmental pollution.

[0003] Existing coastal ecological monitoring technologies usually rely on traditional environmental data collection methods, such as regular sampling and manual observation. These methods often fail to accurately capture rapid changes in the environment in real time, resulting in inconsistent and inaccurate data. In addition, these methods also have limitations in data processing and analysis, especially in terms of efficiency and accuracy when processing large amounts of multi-source data and performing complex environmental pattern recognition.

[0004] The patent application number of my country is CN201510679233.3, and the publication date is 2016.01.13. It discloses a coastal ecological security assessment method based on spatial variables and logistic regression, which includes the following steps: 1) obtaining spatial variable factor data affecting ecological security; 2) obtaining sample points in the study area through regular sampling, and obtaining the ecological security index value; 3) performing logistic regression calculation on all spatial variable factors and expert scoring data, and establishing a logistic regression model; 4) dividing ecological security into 5 levels according to the evaluation value; 5) using the established logistic regression model, modeling and predicting the ecological security level of the entire study area, and outputting an ecological security assessment map; 6) using the Pontius matrix to analyze the ecological security evolution process; 7) outputting and saving the simulation results. Compared with the prior art, the influencing factors used in this invention are mainly spatial distance variables, and the ecological security assessment modeling can be completed with fewer influencing variables.

[0005] However, the above technologies have at least the following technical problems: the existing technology cannot accurately capture the rapid and complex changes in the coastal environment, resulting in the collected data being insufficient to fully reflect the environmental status; there is a lack of effective mechanisms to dynamically adjust data sampling strategies, which limits the ability to respond promptly to environmental mutations or key events; in terms of environmental pattern recognition and trend prediction, it is unable to provide in-depth and accurate analysis, limiting the ability to predict future trends in the ecosystem. Summary of the invention

[0006] The present invention provides a coastal zone ecological security assessment system and method, which solves the technical problems that the existing technology cannot accurately capture the rapid and complex changes in the coastal environment, resulting in insufficient collected data to fully reflect the environmental status; lacks an effective mechanism to dynamically adjust the data sampling strategy, limiting the ability to respond to environmental mutations or key events in a timely manner; and cannot provide in-depth and accurate analysis in terms of environmental pattern recognition and trend prediction, limiting the ability to predict future trends of the ecosystem. A comprehensive coastal zone ecological security assessment system is realized, which effectively improves the accuracy of monitoring and assessment of complex coastal environmental changes by integrating intelligent data collection, high-dimensional data processing and deep learning technology.

[0007] A coastal zone ecological security assessment system and method of the present invention specifically includes the following technical solutions:

[0008] A coastal zone ecological security assessment system includes the following parts:

[0009] Data acquisition module, variation analysis module, sampling optimization module, high-dimensional mapping module, tensor decomposition module, restoration module, Gaussian modeling module, ecological prediction module, ecological security assessment module;

[0010] The data acquisition module is used to obtain the environmental state vector and use nonlinear system theory to model and estimate the environmental state through polynomial expansion, so that the data acquisition strategy can adapt to environmental changes; the data acquisition module is connected to the variation analysis module, high-dimensional mapping module and ecological security assessment module through data transmission;

[0011] The variation analysis module is used to determine whether the sampling frequency has increased by analyzing the variability of time series data; the variation analysis module is connected with the sampling optimization module and the ecological safety assessment module through data transmission;

[0012] The sampling optimization module is used to formulate data collection strategies by combining the environmental state vector and data variability analysis; the sampling optimization module is connected to the data collection module by means of data transmission;

[0013] The high-dimensional mapping module is used to map the collected data into a high-dimensional space, and perform conversion and optimization to obtain a high-dimensional mapping result; the high-dimensional mapping module is connected to the tensor decomposition module by means of data transmission;

[0014] A tensor decomposition module is used to refine and optimize high-dimensional mapping results using a multi-dimensional data optimization method based on tensor decomposition to obtain spatial conversion optimization results. The tensor decomposition module is connected to the restoration module by means of data transmission;

[0015] The restoration module is used to restore the spatial transformation optimization results back to the original dimension using deep learning and nonlinear inverse transformation to obtain the restored data features. The restoration module is connected to the Gaussian modeling module by data transmission;

[0016] Gaussian modeling module, which is used to model the complexity and variability of the coastal environment based on the restored data characteristics, capture the combination of various environmental factors and their impact on the ecosystem. The Gaussian modeling module is connected to the ecological prediction module through data transmission;

[0017] Ecological prediction module, used to predict the future trend of coastal ecosystems by applying deep learning technology and nonlinear functions, and generate ecological trend prediction results. The ecological prediction module is connected to the ecological security assessment module through data transmission;

[0018] The ecological security assessment module is used to integrate environmental state vectors, data variability analysis, and ecological trend prediction results to carry out risk assessment and response strategy formulation.

[0019] A coastal zone ecological security assessment method comprises the following steps:

[0020] S1. Obtain the environment state vector and model and estimate the environment state, and then perform data variability analysis on the environment state vector; combine the environment state vector and data variability analysis to formulate a data collection strategy;

[0021] S2. Sampling is performed according to the data collection strategy, and the collected data is mapped to a high-dimensional space, and then optimized using a multi-dimensional data optimization method based on tensor decomposition, and the data in the high-dimensional space is further restored to the original dimension to obtain the restored data features;

[0022] S3. Model the complexity and variability of the coastal environment based on the restored data characteristics, and make ecological trend predictions. Combine the environmental state vector, data variability analysis, and ecological trend prediction results to conduct risk assessment and formulate response strategies.

[0023] Preferably, the S1 specifically includes:

[0024] Obtain the environmental state vector, and use nonlinear system theory to model and estimate the environmental state through polynomial expansion; use a polynomial model to describe the evolution of the environmental state over time, where the coefficient of each term corresponds to the environmental state and the influence of the control input respectively.

[0025] Preferably, the S1 further includes:

[0026] By analyzing the variability of time series data based on the environmental state vector, it is determined whether the sampling frequency should be increased; and the data variability calculation method is used to obtain the data variability index.

[0027] Preferably, the S1 further includes:

[0028] In the process of implementing the data collection strategy, a sampling decision function based on regression analysis and gating mechanism is introduced.

[0029] Preferably, the S2 specifically includes:

[0030] In the process of mapping the collected data into high-dimensional space, high-order kernel methods and multi-dimensional feature expansion technology are used for high-dimensional mapping. The kernel function is selected as the radial basis function, and polynomial expansion is added; the polynomial expansion is constructed by combining different powers and interaction terms of the collected data.

[0031] Preferably, the S2 further includes:

[0032] Use deep learning and nonlinear inverse transformations to restore data in high-dimensional space back to its original dimensions.

[0033] Preferably, the S2 further includes:

[0034] In the process of restoring data in high-dimensional space, deep neural networks are used for multi-level nonlinear mapping.

[0035] Preferably, the S3 specifically includes:

[0036] The Gaussian mixture model is used to model the complexity and variability of the coastal environment. The mixture of Gaussian distribution is used to describe the probability density under different environmental states to obtain the results of complex environmental impact modeling.

[0037] Preferably, the S3 further includes:

[0038] Based on the results of complex environmental impact modeling, ecological trend prediction is performed by applying deep learning technology and nonlinear functions.

[0039] The beneficial effects of the technical solution of the present invention are:

[0040] 1. By using nonlinear system theory and polynomial expansion to model environmental states, the complex dynamics of the coastal environment can be captured more accurately, which not only improves the accuracy of data collection, but also enables data collection strategies to better adapt to environmental changes; by analyzing the variability of time series data, key environmental change events can be identified in a timely manner and the sampling frequency can be adjusted accordingly, which helps to capture important environmental events in a timely manner and avoid information omissions caused by insufficient or outdated data collection;

[0041] 2. Combining the environmental state vector and data variability analysis, the sampling optimization module can formulate the optimal data sampling strategy, realize the dynamic adjustment of data collection, ensure the representativeness and effectiveness of the data, and thus improve the accuracy of the overall evaluation; through high-dimensional mapping and tensor decomposition, it can process and integrate complex data from multiple sensors and observation platforms, which not only improves the efficiency of data processing, but also extracts more comprehensive information from multi-source data;

[0042] 3. The use of Gaussian mixture models and deep learning technology can deeply identify and predict environmental change patterns, help to better understand and predict the future trends of coastal ecosystems, and provide a more solid scientific basis for ecological security assessments; by comprehensively considering environmental status, data variability and ecological trends, comprehensive ecological security assessments can be carried out, which can provide more comprehensive and accurate safety risk information and provide important support for related decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 A structural diagram of a coastal zone ecological security assessment system provided by one embodiment of the present invention;

[0044] Figure 2 A flow chart of a coastal zone ecological security assessment method provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0045] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all 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.

[0046] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0047] The specific scheme of a coastal zone ecological security assessment system and method provided by the present invention is described in detail below with reference to the accompanying drawings.

[0048] Refer to the attached Figure 1 , which shows a structure diagram of a coastal zone ecological security assessment system provided by an embodiment of the present invention, the system includes the following parts:

[0049] Data acquisition module, variation analysis module, sampling optimization module, high-dimensional mapping module, tensor decomposition module, restoration module, Gaussian modeling module, ecological prediction module, ecological security assessment module;

[0050] The data acquisition module is used to obtain the environmental state vector and use nonlinear system theory to model and estimate the environmental state through polynomial expansion to capture the complex dynamics of the coastal environment, so that the data acquisition strategy can better adapt to environmental changes; the data acquisition module is connected to the variation analysis module, high-dimensional mapping module, and ecological security assessment module through data transmission;

[0051] The variation analysis module is used to determine when the sampling frequency needs to be increased to capture key environmental changes by analyzing the variability of time series data; the variation analysis module is connected to the sampling optimization module and the ecological security assessment module through data transmission;

[0052] The sampling optimization module is used to combine the environmental state vector and data variability analysis to formulate the optimal data collection strategy to achieve dynamic adjustment of data collection; the sampling optimization module is connected to the data collection module through data transmission;

[0053] The high-dimensional mapping module is used to map the collected multi-source data into a high-dimensional space, perform conversion and optimization, and obtain a high-dimensional mapping result; the high-dimensional mapping module is connected to the tensor decomposition module by means of data transmission;

[0054] A tensor decomposition module is used to refine and optimize high-dimensional mapping results using a multi-dimensional data optimization method based on tensor decomposition to obtain spatial conversion optimization results; the tensor decomposition module is connected to the restoration module by means of data transmission;

[0055] The restoration module is used to restore the spatial transformation optimization results back to the original dimension using deep learning and nonlinear inverse transformation to obtain the restored data features; the restoration module is connected to the Gaussian modeling module by data transmission;

[0056] Gaussian modeling module, used to model the complexity and variability of coastal environment based on the restored data characteristics, capturing the combination of various environmental factors and their impact on the ecosystem; the Gaussian modeling module is connected to the ecological prediction module through data transmission;

[0057] The ecological prediction module is used to predict the future trend of the coastal ecosystem by applying deep learning technology and nonlinear functions, and generate ecological trend prediction results; the ecological prediction module is connected to the ecological security assessment module through data transmission;

[0058] The ecological security assessment module is used to integrate environmental state vectors, data variability analysis, and ecological trend prediction results to conduct comprehensive risk assessment and response strategy formulation.

[0059] Refer to the attached Figure 2, which shows a flow chart of a coastal zone ecological security assessment method provided by an embodiment of the present invention, the method comprising the following steps:

[0060] S1. Obtain the environment state vector and model and estimate the environment state, and then perform data variability analysis on the environment state vector; combine the environment state vector and data variability analysis to formulate a data collection strategy;

[0061] The coastal environment is complex and changeable. In order to solve the problems of inconsistency and inaccuracy in data collection in coastal ecosystem monitoring, the dynamic changes of the coastal environment are captured by intelligently adjusting the data collection strategy, analyzing environmental data in real time, and automatically adjusting the sampling frequency and accuracy to optimize data quality.

[0062] The environmental state vector is obtained through the data acquisition module, and the environmental state is modeled and estimated through polynomial expansion using nonlinear system theory to capture the complex dynamics of the coastal environment, so that the data acquisition strategy can better adapt to environmental changes. The polynomial model is used to describe the evolution of the environmental state over time, where the coefficient of each term corresponds to the influence of the environmental state and the control input. The specific formula is:

[0063]

[0064] Among them, X t+1 Represents the predicted environment state vector at the next moment, X t represents the environmental state vector at time t, u t represents the control input at time t, including external influencing factors or system parameters, a 1i is the polynomial coefficient of the state vector, indicating the influence of the state variable, a 2j is the polynomial coefficient of the control input, reflecting the impact of the control input on the environmental state, W t is the environmental noise or uncertainty at time t, a3 and a4 are linear adjustment parameters used to simulate environmental noise and deviation, and i and j are the degree indicators of the polynomial.

[0065] The variation analysis module determines when it is necessary to increase the sampling frequency to capture key environmental changes by analyzing the variability of time series data. When environmental data show significant changes, it indicates that an important environmental event has occurred and more intensive data collection is needed to capture the changes that have occurred; the significant changes refer to major fluctuations in coastal environmental parameters or states in a short period of time, which may be caused by natural events (such as tidal changes, storms, sediment transfer) or human activities (such as coastal construction, pollution emissions). The data variability calculation method is used to calculate the relative rate of change of data between consecutive time points, compare historical data patterns, and apply statistical thresholds to determine whether the changes are significant enough. Specifically, the magnitude and rate of change in the time series data, as well as the degree of deviation of the change from the established threshold or expected pattern are analyzed. If it is higher than the threshold or in an unexpected pattern, it is a significant change. The data variability index calculation formula is:

[0066]

[0067] Among them, V t represents the data variability index at time t, ω k is the weight parameter of variability, X t-k is the environmental state vector at time tk, k is the index in the time series, α1 is the exponential adjustment parameter used to control the sensitivity of variability, and α 2k is the tuning parameter for smoothing variability.

[0068] The sampling optimization module combines the environmental state vector and data variability analysis to formulate the optimal data collection strategy to achieve dynamic adjustment of data collection. The sampling decision function uses regression analysis and gating mechanism. The specific calculation formula is:

[0069]

[0070] Where h is the data collection decision output based on the current environment state vector and data variability, σ is the activation function, and β 1p and β 2q is the decision regression coefficient, which determines the impact of environmental state and data variability on data collection decisions. p and q are the degree indicators in polynomial regression.

[0071] The data collection decision output h ranges from 0 to 1. The closer h is to 1, the higher the current environmental state or data variability, and the more the sampling frequency needs to be increased; the closer h is to 0, the relatively stable environment or the lower the data variability, and the sampling frequency can be reduced. Specifically, a threshold is set based on the expert experience method to determine whether to perform sampling. If h is greater than the threshold, the sampling operation is performed to collect environmental data; otherwise, the sampling at the current moment is skipped.

[0072] S2. Sampling is performed according to the data collection strategy, and the collected data is mapped to a high-dimensional space, and then optimized using a multi-dimensional data optimization method based on tensor decomposition, and the data in the high-dimensional space is further restored to the original dimension to obtain the restored data features;

[0073] Sampling is performed according to the intelligently adjusted data collection strategy. Since the data of coastal ecosystems come from a variety of sensors and observation platforms, it is necessary to process and integrate multi-source data. The high-dimensional mapping module maps the collected multi-source data to a high-dimensional space, converts and optimizes them, and realizes effective data fusion. The high-order kernel method and multi-dimensional feature expansion technology are used for high-dimensional mapping. The kernel function is selected as the radial basis function, and polynomial expansion is added. The polynomial expansion is constructed by combining different powers and interaction terms of the input data, such as

[0074]

[0075] Among them, Y represents the high-dimensional mapping result, D t represents the data collected at time t according to the data collection strategy, P(D t ) is a polynomial expansion, C c is the center of the cth kernel function, κ c is the width parameter of the cth kernel function.

[0076] The tensor decomposition module uses a multidimensional data optimization method based on tensor decomposition to further refine and optimize data. Tensor decomposition can handle the complex structure of high-dimensional data, reveal the multidimensional association of data, and provide more in-depth analysis of data. Through multidimensional data analysis and tensor decomposition, data is further refined and optimized to extract key information. Using the generalized Tucker decomposition, the data tensor is decomposed into a core tensor and a series of pattern matrices. The tensor decomposition function is:

[0077] T rs (Y) = G × 1U (1) ×2U (2) ×3...× s U (S)

[0078] Among them, T rs (Y) represents the tensor decomposition of the s-th dimension of Y under mode r, G is the core tensor, × s represents tensor product, U (s) is the pattern matrix, S is the total number of dimensions of the tensor, r is the pattern index of the tensor decomposition, and s is the dimension index of the tensor. After the tensor is decomposed, the spatial transformation optimization is performed by minimizing the reconstruction error. The formula is as follows:

[0079]

[0080] Among them, Z represents the spatial transformation optimization result, that is, the feature representation, ψ r is the optimization parameter in mode r.

[0081] The restoration module uses deep learning and nonlinear inverse transformation to restore the fused data in high-dimensional space back to the original dimension, making the results easier to understand and explain, ensuring that the fused data can be applied to the assessment of coastal ecological security. A deep neural network is used for multi-level nonlinear mapping, and each layer of the network achieves a deeper feature transformation through a composite function. The restoration formula is:

[0082]

[0083] Among them, F represents the restored data features, tanh is the nonlinear activation function, ζ n is the weight of the nth layer, ζ nm is the weight of the mth neuron in the nth layer, ξ mnk is the weight of the mth neuron in the nth layer for the lth feature, Z l is the lth feature.

[0084] S3. Model the complexity and variability of the coastal environment based on the restored data characteristics, and make ecological trend predictions. Combine the environmental state vector, data variability analysis, and ecological trend prediction results to conduct risk assessment and formulate response strategies.

[0085] The Gaussian modeling module captures the multidimensional characteristics and key information of the coastal environment through the restored data feature F. In order to gain a deeper understanding of the potential patterns and dynamic changes contained in the environmental data, the Gaussian mixture model is used to model the complexity and variability of the coastal environment, capturing the combination of various environmental factors and their impact on the ecosystem. Using a mixture of Gaussian distributions to describe the probability density under different environmental conditions can more accurately predict the impact of environmental changes on the ecosystem. The specific formula is:

[0086]

[0087] Among them, H t It represents the result of modeling the complex environmental impact at time t, which is a comprehensive feature representation, η v is the mixing weight of the vth Gaussian component, ρ vw is the weight of the w-th dimension feature in the v-th Gaussian component, F is the restored data feature, μ vw is the mean of the w-th dimension feature in the v-th Gaussian component, ∑ vw is the covariance matrix of the w-th dimension feature in the v-th Gaussian component, and T is the transpose. By decomposing complex environmental data into multiple Gaussian components, it helps to understand and predict environmental changes in more detail.

[0088] The ecological prediction module conducts ecological trend prediction based on the results of complex environmental impact modeling. By applying deep learning technology and nonlinear functions, it predicts the future trend of coastal ecosystems and their responses when facing various environmental changes. The prediction formula is as follows:

[0089]

[0090] Among them, TS t Represents the prediction of future ecological trends. The sigm function and the hyperbolic tangent function tanh are used to introduce nonlinearity to capture more complex ecosystem dynamics. is the weight of the xth feature in the comprehensive feature representation, χ xy It is the weight parameter of the yth dimension of the xth feature in the comprehensive feature representation. It combines multiple environmental factors and historical data to predict possible ecological trends in the future.

[0091] The ecological security assessment module integrates environmental state vectors, data variability analysis, and ecological trend forecast results to conduct comprehensive risk assessment and response strategy formulation. The risk assessment formula is as follows:

[0092]

[0093] Among them, M t represents the result of risk assessment and response strategy, ω z is the weight of the zth factor, φ z It is a composite function used to combine the ecological trend prediction results TS t and the current environment state vector X t , to provide a comprehensive risk assessment perspective, θ z is the adjustment variability index V t Parameters that influence risk assessment.

[0094] In summary, a coastal ecological security assessment system and method have been completed.

[0095] The order of the embodiments of the invention is for description only and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0096] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0097] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such 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 invention, and should be included in the protection scope of the present invention.

Claims

1. A coastal zone ecological security assessment method, characterized in that: The following steps are involved: S1. Obtain the environmental state vector and model and estimate the environmental state. Then perform data variability analysis on the time series of the environmental state vector and use the data variability calculation method to obtain the data variability index. The specific calculation formula is: Among them, V t Represents the data variability index at time t; ω k is the weight parameter of variability; X t is the environmental state vector at time t; X t-k is the environmental state vector at time tk; k is the index in the time series; α1 is the exponential adjustment parameter; α 2k is the tuning parameter for smoothing variability; Combined with the environmental state vector and data variability analysis, a data collection strategy is formulated. The specific calculation formula is: Where h is the data collection decision output based on the current environment state vector and data variability; σ is the activation function; β 1p and β 2q is the decision regression coefficient; p and q are the degree indicators in polynomial regression; S2. Sampling is performed according to the data collection strategy, and the collected data is mapped to a high-dimensional space, and then optimized using a multi-dimensional data optimization method based on tensor decomposition, and the data in the high-dimensional space is further restored to the original dimension to obtain the restored data features; S3. Based on the restored data characteristics, the complex and variable characteristics of the coastal environment are modeled and the ecological trend is predicted. The prediction formula is as follows: Among them, TS t Represents the prediction of future ecological trends; sigm function and hyperbolic tangent function tanh are used to introduce nonlinearity; H t It represents the result of modeling the complex environmental impact at time t, which is a comprehensive feature representation; It is the feature of the yth dimension in the comprehensive feature representation; is the weight of the xth feature in the comprehensive feature representation; xy is the weight parameter of the yth dimension of the xth feature in the comprehensive feature representation; Then, the environmental state vector, data variability analysis, and ecological trend prediction results are integrated to conduct risk assessment and formulate response strategies. The risk assessment formula is as follows: Among them, M t Represents the results of risk assessment and response strategy; ω z is the weight of the zth environmental factor; φ z It is combined with the ecological trend forecast results TS t and the current environment state vector X t Composite function of θ z is the adjustment variability index V t Parameters that influence risk assessment.

2. The coastal zone ecological security assessment method according to claim 1, characterized in that: The S1 specifically includes: Obtain the environmental state vector, and use nonlinear system theory to model and estimate the environmental state through polynomial expansion; use a polynomial model to describe the evolution of the environmental state over time, where the coefficient of each term corresponds to the environmental state and the influence of the control input respectively.

3. The coastal zone ecological security assessment method according to claim 1, characterized in that: Said S1 further comprises: In the process of implementing the data collection strategy, a sampling decision function based on regression analysis and gating mechanism is introduced.

4. The coastal zone ecological security assessment method according to claim 1, characterized in that: The S2 specifically includes: In the process of mapping the collected data into high-dimensional space, high-order kernel methods and multi-dimensional feature expansion technology are used for high-dimensional mapping. The kernel function is selected as the radial basis function, and polynomial expansion is added; the polynomial expansion is constructed by combining different powers and interaction terms of the collected data.

5. The coastal zone ecological security assessment method according to claim 1, characterized in that: Said S2 further comprises: Use deep learning and nonlinear inverse transformations to restore data in high-dimensional space back to its original dimensions.

6. The coastal zone ecological security assessment method according to claim 5, characterized in that: Said S2 further comprises: In the process of restoring data in high-dimensional space, deep neural networks are used for multi-level nonlinear mapping.

7. The coastal zone ecological security assessment method according to claim 1, characterized in that: The S3 specifically includes: The Gaussian mixture model is used to model the complexity and variability of the coastal environment. The mixture of Gaussian distribution is used to describe the probability density under different environmental conditions to obtain the results of complex environmental impact modeling. The specific formula is: Among them, H t Represents the result of modeling the complex environmental impact at time t, which is a comprehensive feature representation; η v is the mixing weight of the vth Gaussian component; ρ vw is the weight of the w-th dimension feature in the v-th Gaussian component; F is the restored data feature; μ vw is the mean of the w-th dimension feature in the v-th Gaussian component; Σ vw is the covariance matrix of the w-th dimension features in the v-th Gaussian component, and T is the transpose.

8. The coastal zone ecological security assessment method according to claim 7, characterized in that: The S3 further includes: Based on the results of complex environmental impact modeling, ecological trend prediction is performed by applying deep learning technology and nonlinear functions.

9. A coastal zone ecological security assessment system, applied to the coastal zone ecological security assessment method as claimed in claim 1, characterized in that: Includes the following sections: Data acquisition module, variation analysis module, sampling optimization module, high-dimensional mapping module, tensor decomposition module, restoration module, Gaussian modeling module, ecological prediction module, ecological security assessment module; The data acquisition module is used to obtain the environmental state vector and use nonlinear system theory to model and estimate the environmental state through polynomial expansion, so that the data acquisition strategy can adapt to environmental changes; The data acquisition module is connected with the variation analysis module, the high-dimensional mapping module and the ecological safety assessment module through data transmission; The variation analysis module is used to determine whether the sampling frequency has increased by analyzing the variability of time series data; the variation analysis module is connected with the sampling optimization module and the ecological safety assessment module through data transmission; Sampling optimization module, which is used to formulate data collection strategies by combining environmental state vector and data variability analysis; The sampling optimization module is connected to the data acquisition module by means of data transmission; The high-dimensional mapping module is used to map the collected data into a high-dimensional space, and perform conversion and optimization to obtain a high-dimensional mapping result; The high-dimensional mapping module is connected to the tensor decomposition module through data transmission; A tensor decomposition module is used to refine and optimize high-dimensional mapping results using a multi-dimensional data optimization method based on tensor decomposition to obtain spatial conversion optimization results. The tensor decomposition module is connected to the restoration module by means of data transmission; The restoration module is used to restore the spatial transformation optimization results back to the original dimension using deep learning and nonlinear inverse transformation to obtain the restored data features. The restoration module is connected to the Gaussian modeling module by data transmission; Gaussian modeling module, which is used to model the complexity and variability of the coastal environment based on the restored data characteristics, capture the combination of various environmental factors and their impact on the ecosystem. The Gaussian modeling module is connected to the ecological prediction module through data transmission; Ecological prediction module, used to predict the future trend of coastal ecosystems by applying deep learning technology and nonlinear functions, and generate ecological trend prediction results. The ecological prediction module is connected to the ecological security assessment module through data transmission; The ecological security assessment module is used to integrate environmental state vectors, data variability analysis, and ecological trend prediction results to carry out risk assessment and response strategy formulation.

Images