An early warning method and system based on ecological data processing

Abstract

The present invention discloses a warning method and system based on ecological data processing, specifically related to the field of ecological data warning. By improving the data interpolation method, the present invention can more accurately predict the values of missing data points, and by adopting a more complex ecological model to reflect the dynamics and complexity of the ecosystem, the adaptive prediction ability of ecological meteorology is improved.

Description

Technical Field

[0001] The present invention relates to the field of ecological data early warning, and particularly to an early warning method and system based on ecological data processing. Background Art

[0002] Air pollution monitoring and management are key areas of environmental protection. These efforts mainly rely on remote sensing image data obtained from Earth observation satellites, as well as air pollutant concentrations, meteorological data, and environmental quality indicators collected by ground monitoring stations. Satellite remote sensing technology has been widely used because it can cover a vast area and provide data with continuous time series. Through the integration of these data and advanced data processing methods, tools are provided for atmospheric scientists to deeply understand the dynamics and change patterns of air pollution.

[0003] Advances in remote sensing technology, especially high-resolution imaging technology, have greatly improved the spatial resolution of data, enabling atmospheric scientists to more precisely monitor small-scale air pollution changes. The application of multi-spectral and hyper-spectral sensors has enhanced the ability to identify air pollutant types and concentrations, as well as the assessment of environmental quality.

[0004] In terms of data processing, existing technologies have been able to handle large-scale data sets, improving the efficiency of data processing through automated data processing workflows. This workflow includes steps such as data preprocessing, classification, feature extraction, and analysis. Data preprocessing involves operations such as denoising, calibration, and atmospheric correction to ensure data quality. Classification techniques are used to convert remote sensing image data into useful information, such as the distribution of air pollutants and the identification of emission sources. Feature extraction techniques further analyze the classification results and extract parameters useful for air pollution research, such as air quality indices, pollutant emissions, and emission rates.

[0005] Despite significant progress in existing technologies for air pollution data processing, there are still some challenges. First, the quality of remote sensing images may be affected by factors such as cloud cover, sensor noise, and atmospheric conditions, resulting in data discontinuity and incompleteness. Second, the distribution of ground monitoring stations may be uneven, especially in remote or inaccessible areas, which limits the spatial representativeness of the data. In addition, existing data processing methods, such as traditional spatial interpolation and temporal interpolation techniques, may not be able to fully handle the complexity and heterogeneity of the data, especially when facing large-scale and high-dimensional data sets, and face problems of low computational efficiency.

[0006] To overcome these challenges, the field of air pollution monitoring and management needs to further develop and improve remote sensing technology, ground monitoring networks, and data processing algorithms to improve the accuracy, representativeness, and processing efficiency of data, so as to better address air pollution problems and protect and improve environmental quality. Summary of the Invention

[0007] The object of the present invention is to provide a warning method and system based on ecological data processing, which can improve the adaptive prediction ability of ecological meteorology through the niche suitability model and atmospheric dynamic model constructed by machine learning technology.

[0008] The present invention is realized through the following technical solutions:

[0009] A warning method based on ecological data processing, comprising the following steps:

[0010] S1. Obtain multi-temporal and multi-spectral remote sensing image data through Earth observation satellites. The remote sensing image data is atmospheric remote sensing image, and obtain atmospheric parameter data. Correct the clouds, atmospheric scattering and absorption affecting the remote sensing image data through an atmospheric correction algorithm;

[0011] S2. Define spatial data points and their spatial positions, and fill in the missing values in the dataset through spatio-temporal data interpolation. Among them, the spatio-temporal data interpolation includes spatial data interpolation and temporal data interpolation. The spatial data interpolation is for the blank areas in the geospatial data, and the values of unknown points are predicted according to the Kriging method through known data points. The temporal data interpolation is for the missing data points in the time series, and the values of the missing points are predicted according to cubic spline interpolation through the known time series data;

[0012] S3. Extract features from the data after temporal data interpolation processing. Calculate the atmospheric quality index as environmental variable data through atmospheric parameters; establish a niche suitability model through the environmental variable data, and calculate the niche width and predict the niche suitability through the niche suitability model;

[0013] S4. Establish an atmospheric dynamic model through the atmospheric parameter data, and predict the atmospheric change trend through the atmospheric dynamic model;

[0014] S5. Use the niche width prediction result of the niche suitability model and the atmospheric change trend prediction result of the atmospheric dynamic model as features, enhance the data of the features through a generative adversarial network, and identify the atmospheric change pattern through a deep neural network.

[0015] Further, the step S1 specifically includes the following sub-steps:

[0016] S101. Perform boundary adaptability clipping on the remote sensing image data through GIS;

[0017] S102. Reduce the spatial deviation by adjusting the multi-temporal images;

[0018] S103. Correct the clouds, atmospheric scattering and absorption affecting the remote sensing image data through an atmospheric correction algorithm;

[0019] S104. Detect and process clouds and shadows by the pixel reflection value threshold method.

[0020] Further, the step S2 specifically includes the following sub-steps:

[0021] S201. Define the spatial data points Z = {z1, z2…z n} and their spatial positions S = {s1, s2…s n}, where the Z represents the spatial data points, the n represents the total number of known data points, and the S represents the spatial positions of the spatial data points;

[0022] S202. Identify the missing areas in the spatial data and perform spatial interpolation by Kriging method: where the Z(s0) represents the predicted value at the unknown position, the Z(s i ) represents the predicted value at the known position, the λ i represents the weight calculated according to the spatial correlation, and the n represents the total number of known data points;

[0023] S203. Identify the missing points in the time series and perform time interpolation by cubic spline interpolation: where the S(t) represents the interpolation function at the time point t, the t represents the time point, the a i represents the coefficients of the polynomial, the i ranges from 0, 1, 2, and 3, corresponding to the constant term, the first-order term, the second-order term, and the third-order term respectively, the c j represents the correlation coefficient related to the spline basis function B 3,j (t), the j represents the weight of the basis function, and the B 3,j (t) represents the cubic B-spline basis function, which is used to create a smooth curve between data points.

[0024] Further, perform feature extraction on the data after time data interpolation processing, calculate the air quality index through atmospheric parameters, and use it as environmental variable data, specifically including the following sub-steps:

[0025] S3011. Calculate the air quality index through atmospheric parameters, and its process is:

[0026] AQI = max(I p );

[0027]

[0028] where the I p represents the sub-index of pollutant p, the p represents the pollutant affecting the atmospheric quality, the C prepresents the actual concentration of pollutant p, and the B high and B low respectively represent the break point values of the pollutant concentration, and the I high and I low respectively represent the air quality indices corresponding to B high and B low ;

[0029] S3012. Extracting features from the time series of atmospheric parameters by wavelet transform: Among them, W f (a, b) represents the wavelet transform coefficient, ψ * represents the complex conjugate of the wavelet function, a represents the scaling parameter, b represents the translation parameter, f(t) represents the signal, and t represents the time point.

[0030] Furthermore, in step S3, an ecological niche suitability model is established through environmental variable data, and ecological niche suitability prediction is carried out through the ecological niche suitability model, which specifically includes the following sub-steps:

[0031] S3021. Estimating the ecological niche width through the standard deviation of environmental variable data: Among them, X represents the ecological niche width of atmospheric parameter i, and represents the standard deviation of atmospheric environmental variable i. The ecological niche width is used as an input parameter of the ecological niche suitability model;

[0032] S3022. According to the parameters of the maximum likelihood estimation logistic regression model, establish a logistic regression model for predicting the ecological niche suitability of the atmosphere: X = {x1, x2,..., x p}; among them, P(x) represents the atmospheric suitability probability under the given environmental variable x, β0, β1...β p represent the model parameters, and x1, x2...x p represent subsets of the ecological niche width X.

[0033] Furthermore, step S4 specifically includes the following sub-steps:

[0034] S401. Establish a mathematical model describing the atmosphere dynamics through the Lotka-Volterra equation:

[0035]

[0036]

[0037]

[0038] wherein, the dT, dH, and dP respectively represent the change rates of temperature T, humidity H, and atmospheric pressure P, and the F T , F H and F P respectively represent the influence functions of temperature, humidity, and atmospheric pressure on the atmospheric state;

[0039] S402. Predict the change trend of future atmospheric parameters through the Euler algorithm: wherein, the X(t + h) represents the predicted atmospheric parameters at the future time point t + h, the X(t) represents the actual atmospheric parameters at the time point t, and the represents the change rate of the atmospheric parameter vector [T, H, P,...] at time t n , and the h represents the time step.

[0040] Further, the step S5 specifically includes:

[0041] S501. Obtain the predicted results of the niche width of the niche fitness model and the predicted results of the atmospheric change trend of the atmospheric dynamic model, and construct a training set;

[0042] S502. Minimize the loss between the generator and the discriminator through a generative adversarial network to perform data augmentation on the features in the training set: wherein, the G represents the generator network, the D represents the discriminator network, the P data represents the probability distribution of real data, the P z represents the probability distribution of random noise, the V(D, G) represents the value function, the x represents the real sample, and the represents the expectation of the sample x under the real data distribution P data (x), the represents the expectation of the sample x under the latent space distribution z ∼ P z (z), the G z represents the output of the generator, and the z represents the sample drawn from the latent space distribution P z (z);

[0043] S503. Construct and train a deep learning model through the training set processed by data augmentation, and identify the atmospheric change pattern.

[0044] Further, a warning system based on ecological data processing is proposed, and this system is implemented based on any one of the foregoing warning methods based on ecological data processing, and includes:

[0045] A data acquisition module, which is used to obtain multi-temporal and multi-spectral remote sensing image data through an Earth observation satellite. The remote sensing image data is atmospheric remote sensing image, and atmospheric parameter data is obtained. The cloud layer, atmospheric scattering and absorption affecting the remote sensing image data are corrected through an atmospheric correction algorithm;

[0046] A data preprocessing module, which defines spatial data points and their spatial positions, and fills in missing values in the dataset through spatio-temporal data interpolation. Among them, the spatio-temporal data interpolation includes spatial data interpolation and temporal data interpolation. The spatial data interpolation is for blank areas in geospatial data, and the value of unknown points is predicted according to the Kriging method through known data points. The temporal data interpolation is for missing data points in a time series, and the value of missing points is predicted according to cubic spline interpolation through known time series data;

[0047] A feature extraction module, which is used to extract features from the data after temporal data interpolation processing, and calculates the atmospheric quality index as environmental variable data through atmospheric parameters;

[0048] A model construction module, which is used to establish an ecological niche suitability model and an atmospheric dynamic prediction model respectively;

[0049] An atmospheric change pattern prediction module, which uses the prediction result of the ecological niche width of the ecological niche suitability model and the prediction result of the atmospheric change trend of the atmospheric dynamic model as features, enhances the features through a generative adversarial network, and identifies the atmospheric change pattern through a deep neural network;

[0050] Among them, the model construction module specifically includes:

[0051] An ecological niche suitability prediction unit, which is used to establish an ecological niche suitability model through environmental variable data, and calculate the ecological niche width and predict the ecological niche suitability through the ecological niche suitability model;

[0052] An atmospheric dynamic model prediction unit, which is used to establish an atmospheric dynamic model through the prediction result of the ecological niche width and environmental variable data, and predict the atmospheric change trend through the atmospheric dynamic model.

[0053] Furthermore, the ecological niche suitability prediction unit specifically includes:

[0054] An ecological niche width prediction subunit, which is used to estimate the ecological niche width through the standard deviation of environmental variable data: Among them, X represents the ecological niche width of atmospheric parameter i, and represents the standard deviation of atmospheric environmental variable i. The ecological niche width is used as an input parameter of the ecological niche suitability model;

[0055] A model establishment subunit, configured to estimate parameters of a logistic regression model according to maximum likelihood and establish a logistic regression model for predicting the ecological niche suitability of the atmosphere: X = {x1, x2,..., x p}, where P(x) represents the atmospheric suitability probability under the given environmental variable x, β0, β1...β p represent model parameters, and x1, x2...x p represent subsets of the niche width X.

[0056] Further, the atmospheric dynamic model prediction unit specifically includes:

[0057] An atmospheric dynamic model construction subunit, configured to establish a mathematical model describing the atmosphere dynamics through the Lotka-Volterra equation:

[0058]

[0059]

[0060]

[0061] where dT, dH, and dP respectively represent the change rates of temperature T, humidity H, and atmospheric pressure P, and F T , F H and F P respectively represent the influence functions of temperature, humidity, and atmospheric pressure on the atmospheric state;

[0062] A change trend prediction subunit, configured to predict the change trend of future atmospheric parameters through the Euler algorithm: where X(t + h) represents the predicted atmospheric parameters at the future time point t + h, X(t) represents the actual atmospheric parameters at the time point t, and represents the change rate of the atmospheric parameter vector [T, H, P,...] at time t n , and h represents the time step.

[0063] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0064] By improving the data interpolation method, the present invention can more accurately predict the values of missing data points, and by adopting a more complex ecological model, it can reflect the dynamics and complexity of the ecosystem, and improve the adaptive prediction ability of ecological meteorology. Description of the Drawings

[0065] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, form a part of this application, and do not limit the embodiments of the present invention. In the drawings:

[0066] Figure 1 This is the method flow chart of an early warning method based on ecological data processing according to the present invention. Detailed implementation manners

[0067] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings, but the protection scope of the present invention is not limited to the following.

[0068] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail in combination with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention, that is, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Usually, the components of the embodiments of the present invention described and shown in the accompanying drawings herein can be arranged and designed in various different configurations.

[0069] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present invention. It should be noted that relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations.

[0070] Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "including a..." does not exclude the existence of additional identical elements in the process, method, article or device including the element.

[0071] The features and performance of the present invention will be further described in detail below in combination with embodiments.

[0072] An early warning method based on ecological data processing, as Figure 1 , includes the following steps:

[0073] S1. Obtain multi-temporal and multi-spectral remote sensing image data through an Earth observation satellite. The remote sensing image data is an atmospheric remote sensing image, and obtain atmospheric parameter data. Correct the clouds, atmospheric scattering and absorption affecting the remote sensing image data through an atmospheric correction algorithm;

[0074] S2. Define spatial data points and their spatial positions, and fill in the missing values in the dataset through spatio-temporal data interpolation. Among them, the spatio-temporal data interpolation includes spatial data interpolation and temporal data interpolation. The spatial data interpolation is to predict the values of unknown points for blank areas in geospatial data according to the Kriging method through known data points. The temporal data interpolation is to predict the values of missing data points in a time series according to cubic spline interpolation through known time series data;

[0075] S3. Extract features from the data after temporal data interpolation processing, calculate the air quality index as environmental variable data through atmospheric parameters; establish a niche suitability model through environmental variable data, and calculate the niche breadth and predict the niche suitability through the niche suitability model;

[0076] S4. Establish an atmospheric dynamic model through atmospheric parameter data, and predict the atmospheric change trend through the atmospheric dynamic model;

[0077] S5. Use the niche breadth prediction result of the niche suitability model and the atmospheric change trend prediction result of the atmospheric dynamic model as features, enhance the data of the features through a generative adversarial network, and identify the atmospheric change pattern through a deep neural network.

[0078] Further, the step S1 specifically includes the following sub-steps:

[0079] S101. Perform boundary adaptability cropping on remote sensing image data through GIS;

[0080] S102. Reduce spatial deviation by adjusting multi-temporal images;

[0081] S103. Correct clouds, atmospheric scattering and absorption affecting remote sensing image data through an atmospheric correction algorithm;

[0082] S104. Detect and process clouds and shadows through the pixel reflectance threshold method.

[0083] Further, the step S2 specifically includes the following sub-steps:

[0084] S201. Define spatial data points Z = {z1, z2…z n} and their spatial positions S = {s1, s2…s n}, where Z represents spatial data points, n represents the total number of known data points, and S represents the spatial positions of spatial data points;

[0085] S202. Identify the missing areas in the spatial data and perform spatial interpolation through the Kriging method: Wherein, Z(s0) represents the predicted value at the unknown position, and Z(s i ) represents the predicted value at the known position, the λ i represents the weight calculated according to spatial correlation, and n represents the total number of known data points;

[0086] S203. Identify missing points in the time series and perform time interpolation using cubic spline interpolation: Wherein, S(t) represents the interpolation function at time point t, t represents the time point, and a i Represents the coefficient of the polynomial, i ranges from 0, 1, 2 and 3, corresponding to the constant term, linear term, quadratic term and cubic term respectively, and the c j Represents the spline basis function B 3,j (t), the j represents the weight of the basis function, the B 3,j (t) represents the cubic B-spline basis function, which is used to create a smooth curve between data points.

[0087] Specifically, in the above-mentioned embodiments, spatial interpolation uses Kriging to estimate missing spatial data points. Based on the spatial correlation between data points, the values of unknown points are predicted using known data points. Furthermore, temporal interpolation addresses missing values in time series by filling in data using time series prediction models such as linear interpolation and polynomial interpolation.

[0088] Furthermore, in step S3, feature extraction is performed on the data after the time data interpolation processing, and the air quality index is calculated by atmospheric parameters as environmental variable data, which specifically includes the following sub-steps:

[0089] S3011. Calculate the air quality index using atmospheric parameters. The process is as follows:

[0090] AQI=max(I p );

[0091]

[0092] Wherein, the I p The sub-index of pollutant p represents the pollutant that affects the air quality. p represents the actual concentration of pollutant p, the B high and B low Respectively represent the breakpoint values of pollutant concentrations, the I high and I low Respectively represent the corresponding B high and B lowThe air quality index, where the break point values are used to determine the concentration boundaries of different AQI levels. Each pollutant has a series of break point values, which divide the pollutant concentration range into different intervals, and each interval corresponds to a specific AQI value according to; while B high and B low and its corresponding I high and I low are determined based on historical data, scientific research, health impact assessment, and environmental protection standards;

[0093] S3012. Wavelet transform extracts features from the time series of atmospheric parameters: Among them, the W f (a, b) represents the wavelet transform coefficient, the ψ * represents the complex conjugate of the wavelet function, the a represents the scaling parameter, the b represents the translation parameter, the f(t) represents the signal, specifically representing the sequence data of atmospheric parameters changing with time, and the t represents the time point;

[0094] S3013. Establish a feature set based on the features extracted by wavelet transform, other environmental parameters, and the calculated air quality index, and use the feature set as the environmental variable data env. The other environmental parameters at least include temperature, humidity, and atmospheric pressure.

[0095] Specifically, the air quality and pollutant concentration are evaluated by calculating the normalized pollutant index. Further, the multi-scale features in the time series data are extracted by wavelet transform.

[0096] Further, in step S3, an ecological niche suitability model is established through the environmental variable data, and the ecological niche suitability is predicted through the ecological niche suitability model, which specifically includes the following sub-steps:

[0097] S3021. Estimate the ecological niche width through the standard deviation of the environmental variable data: Among them, the X represents the ecological niche width of the atmospheric parameter i, and the represents the standard deviation of the atmospheric environmental variable i. The ecological niche width is used as the input parameter of the ecological niche suitability model;

[0098] S3022. According to the parameters of the maximum likelihood estimation logistic regression model, and establish a logistic regression model to predict the ecological niche suitability of the atmosphere: X = {x1, x2,..., x p}}, where the P(x) represents the atmospheric suitability probability under the given environmental variable x, the β0, β1...β p represent the model parameters, and the x1, x2...x p represent subsets of the ecological niche width X.

[0099] Specifically, for the above embodiments, by constructing the mathematical relationship between environmental variables and atmospheric state, the variability or adaptability of the atmosphere in the environmental variable space is quantified. The model utilizes environmental variables extracted from ground monitoring stations and meteorological data, such as temperature, humidity, atmospheric pressure, and pollutant concentration, etc. By calculating the standard deviation of the variables, the ecological niche breadth of the atmosphere is estimated. As a key indicator, the ecological niche breadth reflects the fluctuation range and adaptability of the atmospheric environment in the face of changes in environmental variables. Through the ecological niche suitability model, the suitability of the atmosphere under different environmental conditions is predicted, including not only the current atmospheric state but also potential change trends and ecological risks. The model uses statistical methods such as logistic regression to link environmental variables with atmospheric suitability and predict the response of the atmosphere to environmental changes.

[0100] Further, as a preferred implementation of this embodiment, a specific implementation process is proposed for the above maximum likelihood estimation. Among them, the maximum likelihood estimation is used to estimate the parameters of the logistic regression model: The represents the estimated value of the parameter after maximum likelihood estimation, and the y i represents the atmospheric observation value. In addition, the specific process of ecological niche suitability prediction is expressed as: DegradationWarning = f(P i (x)), where the f represents the warning function. Exemplarily, the warning level is determined according to the ecological niche suitability P i (x).

[0101] Further, the step S4 specifically includes the following sub-steps:

[0102] S401. By using the Lotka-Volterra equation, a mathematical model describing the atmosphere dynamics is established:

[0103]

[0104]

[0105]

[0106] where the dT, dH, and dP respectively represent the change rates of temperature T, humidity H, and atmospheric pressure P, and the F T , F H and F P respectively represent the influence functions of temperature, humidity, and atmospheric pressure on the atmospheric state;

[0107] S402. Predict the change trend of future atmospheric parameters through the Euler algorithm: Among them, the X(t + h) represents the predicted atmospheric parameters at the future time point t + h, the X(t) represents the actual atmospheric parameters at the time point t, and the represents the atmospheric parameter vector [T, H, P,...] at time t n of the rate of change, and the h represents the time step.

[0108] Specifically, the atmospheric dynamic model is used to simulate and predict the changes in the atmosphere in the ecosystem. Exemplarily, the interaction between environmental variables and atmospheric changes is simulated through the Lotka-Volterra equation.

[0109] Further, in the above embodiments, the parameters of the influence functions of temperature, humidity, and atmospheric pressure on the atmospheric state are determined by the nonlinear least squares method: The represents the estimated value of the parameters after the nonlinear least squares method, and the represents the concentration of atmospheric pollutants predicted by the model.

[0110] Further, the step S5 specifically includes:

[0111] S501. Obtain the predicted results of the niche width of the niche suitability model and the predicted results of the atmospheric change trend of the atmospheric dynamic model, and construct a training set;

[0112] S502. Minimize the loss between the generator and the discriminator through the generative adversarial network to perform data augmentation on the features in the training set: Among them, the G represents the generator network, the D represents the discriminator network, and the P data represents the probability distribution of the real data, the P z represents the probability distribution of the random noise, the V(D, G) represents the value function, the x represents the real sample, and the represents the expectation of the sample x under the real data distribution P data (x), and the represents the expectation of the sample x under the potential space distribution z ∼ P z (z), the G z represents the output of the generator, and the z represents the sample drawn from the potential space distribution P z (z);

[0113] S503. Construct and train a deep learning model through the training set processed by data augmentation, and identify the atmospheric change pattern.

[0114] Specifically, the deep learning model includes, but is not limited to, a convolutional neural network (CNN), a recurrent neural network (RNN), a long short-term memory network (LSTM), or a Transformer model. The niche width prediction results provide information on the adaptability of atmospheric parameters, while the atmospheric change trend prediction results provide dynamic information on the changes in atmospheric parameters over time. The niche width prediction results and the atmospheric change trend prediction results are used as input features for the deep learning model to help the model learn the complex patterns of the atmospheric state. The deep learning model identifies and predicts atmospheric change patterns, such as the development of pollution events and extreme weather events, based on the above features.

[0115] Further, as a preferred implementation of the above embodiment, for identifying the ecological degradation pattern after the prediction of the deep learning model, an optimized prediction method based on the Markov random field is proposed, which specifically includes the following steps:

[0116] S601. Integrate the atmospheric parameter features predicted by the deep learning model with the atmospheric environment variable data obtained through wavelet transform and AQI calculation;

[0117] S602. Define the energy function of the Markov random field according to the integrated feature data. The energy function includes node potential energy and the interaction term between adjacent nodes;

[0118] S603. Establish a Markov random field model based on the probability of the atmospheric state according to the energy function. Train the Markov random field model through the Gibbs sampling algorithm to obtain the probability distribution of the spatial data;

[0119] S604. Predict the atmospheric state through the Markov random field model, including but not limited to the change trend of the air quality index.

[0120] Further, the node potential energy and the interaction term between adjacent nodes are specifically expressed as:

[0121]

[0122] Among them, E(x) represents the point potential energy and the interaction term between adjacent nodes, U(x i ) represents the potential energy of node i, J ij represents the interaction between nodes i and j, and δ(·) represents the indicator function.

[0123] Further, the training of the Markov random field model by the Gibbs sampling algorithm is specifically expressed as:

[0124]

[0125] Among them, x i(t + 1) represents the state of node i at time t + 1, and the P(x i |x \i ) represents the probability distribution of node i given the states of all other nodes x \i , where the x \i represents the states of all nodes other than node i. The T represents the temperature parameter, and the E(x i |x \i ) represents the energy function of node i given the states of all other nodes x \i .

[0126] Specifically, the above embodiments are used to simulate the spatial distribution and correlation of air pollution, and generate spatial distribution samples of air pollution through Gibbs sampling. The spatial distribution samples are used to predict and evaluate the risk of air pollution.

[0127] Furthermore, an early warning system based on ecological data processing is proposed. This system is implemented based on the above-mentioned early warning method for ecological data processing, and includes:

[0128] A data acquisition module, which is used to obtain multi-temporal and multi-spectral remote sensing image data through an Earth observation satellite. The remote sensing image data is an atmospheric remote sensing image, and obtain atmospheric parameter data, and correct the clouds, atmospheric scattering and absorption affecting the remote sensing image data through an atmospheric correction algorithm;

[0129] A data preprocessing module, which defines spatial data points and their spatial positions, and fills in the missing values in the dataset through spatio-temporal data interpolation. Among them, the spatio-temporal data interpolation includes spatial data interpolation and temporal data interpolation. The spatial data interpolation is for blank areas in geospatial data, and the values of unknown points are predicted according to the Kriging method through known data points. The temporal data interpolation is for missing data points in a time series, and the values of the missing points are predicted according to cubic spline interpolation through known time series data;

[0130] A feature extraction module, which is used to extract features from the data after temporal data interpolation processing, and calculate the air quality index as environmental variable data through atmospheric parameters;

[0131] A model construction module, which is used to establish a niche suitability model and an atmospheric dynamic prediction model respectively;

[0132] An atmospheric change pattern prediction module, which uses the niche width prediction result of the niche suitability model and the atmospheric change trend prediction result of the atmospheric dynamic model as features, enhances the features through a generative adversarial network, and identifies the atmospheric change pattern through a deep neural network;

[0133] Among them, the model construction module specifically includes:

[0134] The niche suitability prediction unit is used to establish a niche suitability model through environmental variable data, and calculate the niche breadth and predict the niche suitability through the niche suitability model;

[0135] The atmospheric dynamic model prediction unit is used to establish an atmospheric dynamic model through the niche breadth prediction result and environmental variable data, and predict the atmospheric change trend through the atmospheric dynamic model.

[0136] Furthermore, the niche suitability prediction unit specifically includes:

[0137] The niche breadth prediction subunit is used to estimate the niche breadth through the standard deviation of environmental variable data: Among them, X represents the niche breadth of atmospheric parameter i, and represents the standard deviation of atmospheric environmental variable i. The niche breadth is used as an input parameter of the niche suitability model;

[0138] The model establishment subunit is used to estimate the parameters of the maximum likelihood estimation logistic regression model and establish a logistic regression model to predict the niche suitability of the atmosphere: X = {x1, x2,..., x p}, where P(x) represents the atmospheric suitability probability under the given environmental variable x, and β0, β1...β p represent the model parameters, and x1, x2...x[[ID=2,6]] p represent subsets of the niche breadth X.

[0139] Furthermore, the atmospheric dynamic model prediction unit specifically includes:

[0140] The atmospheric dynamic model construction subunit is used to establish a mathematical model describing the atmospheric dynamics through the Lotka-Volterra equation:

[0141]

[0142]

[0143]

[0144] Among them, dT, dH, and dP respectively represent the change rates of temperature T, humidity H, and atmospheric pressure P, and F T 、F H and F P respectively represent the influence functions of temperature, humidity, and atmospheric pressure on the atmospheric state;

[0145] A change trend prediction subunit for predicting the change trend of future atmospheric parameters through the Euler algorithm: wherein, X(t + h) represents the predicted atmospheric parameter at the future time point t + h, X(t) represents the actual atmospheric parameter at the time point t, and the represents the change rate of the atmospheric parameter vector [T, H, P,...] at time t n and h represents the time step.

[0146] The specific embodiments described above further elaborate on the purpose, technical solutions, and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. An early warning method based on ecological data processing, characterized in that, It includes the following steps: S1. Obtain multi-temporal and multi-spectral remote sensing image data through an Earth observation satellite. The remote sensing image data is atmospheric remote sensing image, and obtain atmospheric parameter data. Correct the clouds, atmospheric scattering and absorption that affect the remote sensing image data through an atmospheric correction algorithm; S2. Define spatial data points and their spatial positions, and fill in the missing values in the dataset through spatio-temporal data interpolation. Among them, the spatio-temporal data interpolation includes spatial data interpolation and temporal data interpolation. The spatial data interpolation is for the blank areas in the geospatial data, and the values of unknown points are predicted according to the Kriging method through known data points. The temporal data interpolation is for the missing data points in the time series, and the values of the missing points are predicted according to the cubic spline interpolation through the known time series data; S3. Extract features from the data after temporal data interpolation processing. Calculate the atmospheric quality index through atmospheric parameters as environmental variable data; establish a niche suitability model through the environmental variable data, and calculate the niche breadth and predict the niche suitability through the niche suitability model; S4. Establish an atmospheric dynamic model through the atmospheric parameter data, and predict the atmospheric change trend through the atmospheric dynamic model; S5. Use the niche breadth prediction result of the niche suitability model and the atmospheric change trend prediction result of the atmospheric dynamic model as features, enhance the data of the features through a generative adversarial network, and identify the atmospheric change pattern through a deep neural network.

2. The early warning method based on ecological data processing according to claim 1, wherein The step S1 specifically includes the following sub-steps: S101. Perform boundary adaptability cropping on the remote sensing image data through GIS; S102. Reduce the spatial deviation by adjusting the multi-temporal images; S103. Correct the clouds, atmospheric scattering and absorption that affect the remote sensing image data through an atmospheric correction algorithm; 3. The early warning method based on ecological data processing according to claim 1, characterized in that S104. Detect and process clouds and shadows through the pixel reflectance threshold method. S201. Define spatial data points and their spatial positions , where the represents spatial data points, the represents the total number of known data points, and the represents the spatial positions of the spatial data points; S202. Identify the missing regions in the spatial data and perform spatial interpolation by Kriging method: , where the represents the predicted value at the unknown location, and the represents the predicted value at the known location, and the represents the weight calculated based on spatial correlation, and the represents the total number of known data points; S203. Identify the missing points in the time series and perform time interpolation through cubic spline interpolation: ; wherein, the represents the interpolation function at time point t, the t represents the time point, and the represents the coefficients of the polynomial. The value range of i is 0, 1, 2, and 3, corresponding to the constant term, the first-order term, the second-order term, and the third-order term respectively. The represents the correlation coefficient related to the spline basis function . The j represents the weight of the basis function, and the represents the cubic B-spline basis function, which is used to create a smooth curve between data points.

4. The early warning method based on ecological data processing according to claim 1, characterized in that The step S2 specifically includes the following sub-steps: In the step S3, when extracting features from the data after temporal data interpolation processing and calculating the atmospheric quality index through atmospheric parameters as environmental variable data, it specifically includes the following sub-steps: ； ； Among them, the represents the sub-index of the pollutant , the represents the pollutant affecting the atmospheric quality, the represents the actual concentration of the pollutant , the and respectively represent the break point values of the pollutant concentration, the and respectively represent the corresponding and atmospheric quality indices; S3012. Extract features from the time series of atmospheric parameters by wavelet transform: , where the represents the wavelet transform coefficient, the represents the complex conjugate of the wavelet function, a represents the scaling parameter, b represents the translation parameter, and the represents the signal, and t represents the time point.

5. The early warning method based on ecological data processing according to claim 1, characterized in that, S3011. Calculate the atmospheric quality index through atmospheric parameters, and its process is: S3021. Estimate the niche breadth through the standard deviation of environmental variable data: , where the represents the niche breadth of atmospheric parameter i, and the represents the standard deviation of atmospheric environmental variable i. The niche breadth is used as an input parameter for the niche suitability model; S3022. Estimate the parameters of the logistic regression model according to the maximum likelihood, and establish a logistic regression model to predict the niche suitability of the atmosphere: ; { 、 、 、 }, where the represents the atmospheric suitability probability under the given environmental variable x, the represents the model parameter, and the represents the niche width subset.

6. The early warning method based on ecological data processing according to claim 1, characterized in that In the step S3, when establishing a niche suitability model through the environmental variable data and predicting the niche suitability through the niche suitability model, it specifically includes the following sub-steps: The step S4 specifically includes the following sub-steps: ； ； ； Among them, the , and respectively represent the change rates of temperature T, humidity H, and atmospheric pressure P, and the , and respectively represent the influence functions of temperature, humidity, and atmospheric pressure on the atmospheric state; S402. Predict the change trend of future atmospheric parameters through the Euler algorithm: , where the represents the predicted atmospheric parameters at the future time point , the represents the actual atmospheric parameters at the time point t, and the represents the atmospheric parameter vector at the time of the change rate, and the represents the time step.

7. The early warning method based on ecological data processing according to claim 1, characterized in that S401. Establish a mathematical model describing the atmospheric dynamics through the Lotka-Volterra equation: The step S5 specifically includes: S502. Data augmentation of features in the training set by minimizing the loss between the generator and the discriminator through a generative adversarial network: , where the represents the generator network, the represents the discriminator network, the represents the probability distribution of real data, the represents the probability distribution of random noise, the represents the value function, the represents the real sample, the represents the real data distribution of the expectation of the sample x under represents the latent space distribution of the expectation of the sample x under represents the generator output, the represents a sample drawn from the latent space distribution ; S501. Obtain the niche breadth prediction result of the niche suitability model and the atmospheric change trend prediction result of the atmospheric dynamic model, and construct a training set; 8. An early warning system based on ecological data processing, which is implemented based on an early warning method for ecological data processing according to any one of claims 1-7, characterized in that, S503. Construct and train a deep learning model through the training set processed by data enhancement, and identify the atmospheric change pattern. It includes: A data acquisition module, which is used to obtain multi-temporal and multi-spectral remote sensing image data through an Earth observation satellite. The remote sensing image data is atmospheric remote sensing image, and atmospheric parameter data is obtained. The cloud cover, atmospheric scattering and absorption affecting the remote sensing image data are corrected through an atmospheric correction algorithm; A data preprocessing module, which defines spatial data points and their spatial positions, and fills in missing values in the dataset through spatio-temporal data interpolation. Among them, the spatio-temporal data interpolation includes spatial data interpolation and temporal data interpolation. The spatial data interpolation is for blank areas in geospatial data, and the values of unknown points are predicted according to the Kriging method through known data points. The temporal data interpolation is for missing data points in a time series, and the values of missing points are predicted according to cubic spline interpolation through known time series data; A feature extraction module, which is used to extract features from the data after temporal data interpolation processing, and calculate the atmospheric quality index as environmental variable data through atmospheric parameters; A model construction module, which is used to establish an ecological niche suitability model and an atmospheric dynamic prediction model respectively; An atmospheric change pattern prediction module, which uses the ecological niche width prediction result of the ecological niche suitability model and the atmospheric change trend prediction result of the atmospheric dynamic model as features, enhances the data of the features through a generative adversarial network, and identifies the atmospheric change pattern through a deep neural network; Among them, the model construction module specifically includes: An ecological niche suitability prediction unit, which is used to establish an ecological niche suitability model through environmental variable data, and calculate the ecological niche width and predict the ecological niche suitability through the ecological niche suitability model; An atmospheric dynamic model prediction unit, which is used to establish an atmospheric dynamic model through the ecological niche width prediction result and environmental variable data, and predict the atmospheric change trend through the atmospheric dynamic model.

9. The early warning system based on ecological data processing according to claim 8, wherein The ecological niche suitability prediction unit specifically includes: The niche width prediction subunit is used to estimate the niche width by the standard deviation of environmental variable data: , where the represents the niche width of atmospheric parameter i, and the represents the standard deviation of atmospheric environmental variable i. The niche width is used as an input parameter for the niche suitability model; A model establishment subunit, configured to estimate parameters of a maximum likelihood estimation logistic regression model and establish a logistic regression model for predicting the niche suitability of the atmosphere: ; { , , , }, where the represents the atmospheric suitability probability under a given environmental variable x, the represents the model parameter, and the represents the niche width subset.

10. The early warning system based on ecological data processing according to claim 8, wherein, The atmospheric dynamic model prediction unit specifically includes: An atmospheric dynamic model construction subunit, which is used to establish a mathematical model describing the atmosphere dynamics through the Lotka-Volterra equation: ； ； ； Among them, the , and respectively represent the change rates of temperature T, humidity H, and atmospheric pressure P, and the , and respectively represent the influence functions of temperature, humidity, and atmospheric pressure on the atmospheric state; A change trend prediction subunit, which is used to predict the change trend of future atmospheric parameters through the Euler algorithm: ； wherein, the represents the predicted atmospheric parameters at a future time point , the represents the actual atmospheric parameters at time point t, the represents the atmospheric parameter vector at time with a rate of change, the represents the time step.

Images