A method for predicting the spread of pine wood nematode disease, electronic equipment and storage medium

Abstract

A method for predicting the spread of pine nematode disease, an electronic device and a storage medium belong to the technical field of pine nematode disease prediction. In order to improve the accuracy of pine nematode disease prediction, the present invention includes data collection, including historical disease occurrence data in diseased wood distribution areas and meteorological data and geographic information data in diseased wood distribution areas; preprocessing, and then constructing a training set, a validation set and a test set; constructing a pine nematode disease propagation prediction model based on the Kalman‑LSTM‑NGO combination, using the obtained training set to train the model, using the validation set to verify, tune and test the model, using the test set to predict the model, and obtaining the pine nematode disease propagation prediction results for visualization processing, which is used to help decision makers analyze fertilization, thinning, and felling measures based on the visualization processing results, and propose optimization suggestions including pine epidemic monitoring, control means and resource allocation. The present invention predicts accurately.

Description

Technical Field

[0001] The invention belongs to the technical field of pine wood nematode disease prediction, and specifically relates to a pine wood nematode disease propagation prediction method, electronic equipment and storage medium. Background Art

[0002] Pine wilt disease is a common forest disease. Traditional prediction and prevention methods rely on experience and observation, and have problems such as low accuracy and lack of timeliness. Remote sensing technology is used to monitor the pine wilt disease epidemic. Optical satellite remote sensing images obtained by remote sensing technology are analyzed and processed to quickly and timely monitor the distribution and location of dead trees in pine wilt disease epidemic areas. However, since optical satellite remote sensing images are greatly affected by weather and are also limited by spatial resolution or spectral resolution, it is difficult to accurately predict and prevent them. Deep learning, as an artificial intelligence technology, has powerful data processing and pattern recognition capabilities and is widely used in various fields. In the face of time series sensitive problems and tasks, RNN (such as LSTM) is usually more suitable, but when solving a specific problem, multiple different single models are generally used for various assumptions. When a single prediction model is used for prediction, the predicted value occasionally deviates greatly from the actual value, resulting in greater uncertainty in the prediction and a great impact on the prediction results. For pine wilt disease, an efficient and accurate prediction and prevention method is needed. Summary of the invention

[0003] The problem to be solved by the present invention is to improve the prediction accuracy of pine wood nematode disease, and propose a pine wood nematode disease propagation prediction method, electronic equipment and storage medium.

[0004] To achieve the above object, the present invention is implemented through the following technical solutions:

[0005] A method for predicting the spread of pine wood nematode disease, comprising the following steps:

[0006] S1. Data collection, including historical disease occurrence data in diseased tree distribution areas, meteorological data and geographic information data in diseased tree distribution areas;

[0007] S2. Preprocess the data collected in step S1, and then construct a training set, a validation set, and a test set;

[0008] S3. Construct a pine wood nematode disease propagation prediction model based on the Kalman-LSTM-NGO combination, and use the training set obtained in step S2 to train the model to obtain a trained pine wood nematode disease propagation prediction model based on the Kalman-LSTM-NGO combination;

[0009] S4. Use the validation set obtained in step S2 to verify, tune and test the pine wood nematode disease propagation prediction model based on the Kalman-LSTM-NGO combination trained in step S3 to obtain a tuned pine wood nematode disease propagation prediction model based on the Kalman-LSTM-NGO combination;

[0010] S5. Use the test set obtained in step S2 to predict the pine wood nematode disease propagation prediction model based on the Kalman-LSTM-NGO combination optimized in step S4 to obtain a pine wood nematode disease propagation prediction result;

[0011] S6. Visualize the prediction results of pine wood nematode disease spread obtained in step S5 to help decision makers analyze fertilization, thinning, and felling measures based on the visualization results, and put forward optimization suggestions including pine tree epidemic monitoring, control measures, and resource allocation.

[0012] Furthermore, the historical disease occurrence data of the diseased tree distribution area in step S1 includes location, epidemic situation, time and disease severity information data, and the meteorological data and geographic information data of the diseased tree distribution area include average annual temperature, average annual maximum temperature, average annual minimum temperature, humidity, average annual precipitation, normalized vegetation index, topography, soil type, and altitude.

[0013] Furthermore, the specific implementation method of step S2 includes the following steps:

[0014] S2.1. For the data collected in step S1, select the factors affecting pine wood nematode disease, including human activity factors and natural environment factors. The human activity factors include the total amount of pine wood transported in, the amount of pine wood transported in from the epidemic area, road density, and host insect control rate. The natural environment factors include the average temperature from June to August, the average precipitation from June to August, the average altitude, the host insect disaster area, and the slope;

[0015] S2.2. The Pearson correlation coefficient is used to perform correlation analysis on the influencing factors of pine wood nematode disease selected in step S1. The expression is:

[0016]

[0017] Among them, x represents the data of pine wood nematode disease, y represents the data of the influencing factors of pine wood nematode disease, ρ x,y represents the correlation coefficient, and N represents the number of variable values;

[0018] Based on correlation analysis, the influencing factors of pine wood nematode disease were compared through 10 rounds of iterations, and the average temperature from June to August, the average precipitation from June to August, the total amount of pine wood transported in, the amount of pine wood transported in from the epidemic area, the road density, and the host insect control rate were selected as the significant influencing factors of pine wood nematode disease.

[0019] S2.3. Extract the data of significant pine wood nematode disease influencing factors selected in step S2.2, and perform data preprocessing, including data standardization, missing value processing, and normalization processing;

[0020] S2.4. Use the data preprocessed in step S2.3 to construct a training set, a validation set, and a test set. Use 80% of the data in the 2017-2021 data set as the training set, 20% of the data as the validation set, and the 2022 data set as the test set.

[0021] Furthermore, the specific implementation method of step S3 includes the following steps:

[0022] S3.1. Use Kalman filter algorithm to perform noise reduction on the preprocessed data obtained in step S2;

[0023] S3.2. Construct a Kalman model, input the training set data after denoising in step S3.1 into the Kalman model for processing, and obtain the residual sequence of the Kalman model prediction value;

[0024] S3.2.1. Parameter initialization: Set the observation value at the start time to The expression is:

[0025]

[0026] P 0 / 0 =∑x0

[0027] Among them, m X0 and P 0 / 0 Respectively represent the expectation and variance of the initial state vector x0;

[0028] S3.2.2. Based on the initial state vector, use the prediction formula to make a prediction and get the estimated value The expression is:

[0029]

[0030] Wherein, H0, R0, L0 represent the first initialization value, the second initialization value, and the third initialization value of the state parameter;

[0031] S3.2.3. Calculate the error covariance P using the observed value at the start time of step S3.2.1, then update the Kalman gain coefficients Q and R, perform recursive operations, correct the estimated value, and update the error covariance until the iterative calculation is completed, and output the residual sequence of the Kalman model prediction value;

[0032] S3.3. Construct an LSTM neural network model, input the residual sequence of the Kalman model prediction value obtained in step S3.2 into the LSTM neural network model for training, update the LSTM neural network model parameters based on the Adam error back propagation method, and obtain the Kalman-LSTM neural network model;

[0033] S3.3.1. Construct an LSTM neural network model and use it to fit and predict the number of pine wood nematode disease cases. The LSTM neural network model unit consists of an input gate, a forget gate, and an output gate. The LSTM neural network model selectively allows information to pass through the gate structure to update or retain historical information. The expression of the LSTM neural network model is:

[0034] f t =σ(W f ·[h t-1 ,x t ]+b f )

[0035] i t =σ(W i ·[h t-1 ,x t ]+b i )

[0036]

[0037]

[0038] h t =o t tanh(C t )

[0039] Among them, f t represents the forget gate, σ represents the Sigmoid function, W f represents the weight of the forget gate, h t-1 represents the hidden state at time t-1, x t represents the input at time t, b f represents the bias of the forget gate, i t represents the input gate, W i represents the weight of the input gate, b i represents the bias of the input gate, represents the candidate state value of the input unit at time t, tanh represents the Tanh function, W c represents the weight of the output gate, b c represents the deviation of the output gate, C t and C t -1 represents the unit state update value at time t and time t-1, ht represents the hidden state at time t, o t represents the output gate;

[0040] S3.3.2. Input the residual sequence of the Kalman model prediction value obtained in step S3.2 into the LSTM neural network model for training, and update the LSTM neural network model parameters based on the Adam error back propagation method. The expression is:

[0041] m t =μ×m t-1 +(1-μ)×g t

[0042]

[0043]

[0044] Among them, m t represents the first-order moment estimate of the gradient at time t, m t-1 represents the first-order moment estimate of the gradient at time t-1, n t represents the second-order moment estimate of the gradient, n t-1 represents the second-order moment estimate of the gradient at time t-1, represents the gradient first-order moment estimate m t Correction, Represents the second-order moment estimate of the gradient n t Correction, g t represents the gradient value, v represents the second attenuation rate, △θ t represents the update value of the parameter, μ represents the first decay rate;

[0045] S3.4. Use the NGO algorithm to optimize the number of hidden layer neurons, learning rate and training times of the Kalman-LSTM neural network model obtained in step S3.3 to obtain a trained pine wood nematode disease transmission prediction model based on the Kalman-LSTM-NGO combination.

[0046] Furthermore, step S4 uses mean absolute error MAE and mean absolute error percentage MAPE as prediction error evaluation indicators;

[0047] The expression of mean absolute error MAE is:

[0048]

[0049] in, Represents the predicted value, y k represents the true value;

[0050] The expression of mean absolute error percentage MAPE is:

[0051]

[0052] Furthermore, the prediction results of pine wood nematode disease spread obtained in step S5 are visualized by marking the disease prediction areas on a national map using the geographic information system ArcGIS.

[0053] An electronic device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of a method for predicting the spread of pine wood nematode disease when executing the computer program.

[0054] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements a method for predicting the spread of pine wood nematode disease.

[0055] Beneficial effects of the present invention:

[0056] The invention discloses a method for predicting the spread of pine wood nematode disease. The method uses a combination of an LSTM long short-term memory network and a Kalman filter model to predict the spread of the disease. The Northern Goshawk Optimization (NGO) algorithm with strong global search capability and high algorithm robustness is used to optimize hyperparameters such as the number of neurons in the hidden layer of the Kalman-LSTM network, a learning rate, and the number of training times. A Kalman-LSTM model is established, and then a pine wood nematode disease prevention and control system is established, so that corresponding prevention and control decisions can be taken in an early stage, thereby better protecting forestry resources.

[0057] The method for predicting the spread of pine wood nematode disease described in the present invention has high accuracy and timeliness: the application of the method for predicting the spread of pine wood nematode disease improves the accuracy and timeliness of pine wood nematode disease prediction, and compared with traditional methods, it can more accurately predict the future spread trend of the disease.

[0058] The method for predicting the spread of pine wilt disease described in the present invention constructs a data-driven propagation model: the system uses a large amount of historical disease data for learning, thereby establishing an efficient prediction model. The prediction results are more reliable based on actual observation data. By constructing a propagation model of pine wilt disease, a scientific basis can be provided to help decision makers understand the disease spread trend, so as to better plan prevention and control measures.

[0059] The method for predicting the spread of pine wood nematode disease described in the present invention realizes interdisciplinary knowledge integration, combines knowledge from multiple fields such as machine learning, geographic information systems, ecology and forestry management, and combines expert knowledge and historical experience. It can comprehensively consider the complexity of the spread of pine wood nematode disease and provide more comprehensive prevention and control measures.

[0060] The method for predicting the spread of pine wood nematode disease described in the present invention not only provides prediction results, but also provides relevant prevention and control measures for decision makers. By analyzing the spread of the disease in real time, prevention and control work can be responded to and adjusted in a timely manner, thereby improving the efficiency of responding to the disease. This helps government agencies and forestry managers better protect forestry resources and reduce economic losses.

[0061] The method for predicting the spread of pine wood nematode disease described in the present invention can automatically learn and extract features, automatically adapt feature extraction according to data for processing diverse data types (such as images, time series data, text descriptions, etc.), and can monitor and analyze data in real time, so that the outbreak or spread of the disease can be detected more quickly. This helps to take timely control measures to reduce the impact of pests and diseases on forestry.

[0062] The method for predicting the spread of pine wood nematode disease described in the present invention is scalable: the deep learning model can be expanded and adjusted as needed to adapt to different types of pest prediction tasks. Different deep learning architectures can be constructed to cope with different types of pest challenges. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 The present invention provides a flow chart of a method for predicting the spread of pine wood nematode disease. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical solution and advantages of the present invention more clear, the present invention is further described in detail below in conjunction with the accompanying drawings and specific 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 specific embodiments described are only part of the embodiments of the present invention, rather than all of the specific embodiments. The components of the specific embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations, and the present invention can also have other embodiments.

[0065] Therefore, the following detailed description of the specific embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents the selected specific embodiments of the present invention. Based on the specific embodiments of the present invention, all other specific embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present invention.

[0066] In order to further understand the content, features and effects of the present invention, the following specific implementation methods are given as examples, and the attached Figure 1 The detailed instructions are as follows: Specific implementation method one:

[0068] A method for predicting the spread of pine wood nematode disease, comprising the following steps:

[0069] S1. Data collection, including historical disease occurrence data in diseased tree distribution areas, meteorological data and geographic information data in diseased tree distribution areas;

[0070] Furthermore, the historical disease occurrence data of the diseased tree distribution area in step S1 includes location, epidemic situation, time and disease severity information data, and the meteorological data and geographic information data of the diseased tree distribution area include average annual temperature, average annual maximum temperature, average annual minimum temperature, humidity, average annual precipitation, normalized vegetation index, topography, soil type, and altitude;

[0071] S2. Preprocess the data collected in step S1, and then construct a training set, a validation set, and a test set;

[0072] Furthermore, the specific implementation method of step S2 includes the following steps:

[0073] S2.1. For the data collected in step S1, select the factors affecting pine wood nematode disease, including human activity factors and natural environment factors. The human activity factors include the total amount of pine wood transported in, the amount of pine wood transported in from the epidemic area, road density, and host insect control rate. The natural environment factors include the average temperature from June to August, the average precipitation from June to August, the average altitude, the host insect disaster area, and the slope;

[0074] S2.2. The Pearson correlation coefficient is used to perform correlation analysis on the influencing factors of pine wood nematode disease selected in step S1. The expression is:

[0075]

[0076] Among them, x represents the data of pine wood nematode disease, y represents the data of the influencing factors of pine wood nematode disease, ρ x,y represents the correlation coefficient, and N represents the number of variable values;

[0077] Furthermore, in the Pearson correlation test, the p-value is generally 5% (also known as the significance level). If the p-value is far greater than 0.05, the correlation between the two variables is not obvious. And usually, the value of the correlation coefficient is discussed only when the p-value reaches a certain level. The correlation coefficient is used to measure linear correlation. The correlation becomes stronger as the absolute value of the correlation coefficient increases. That is to say, the closer the correlation coefficient is to 1 or -1, the stronger the correlation is, and the closer the correlation coefficient is to 0, the weaker the correlation is.

[0078] Based on correlation analysis, the influencing factors of pine wood nematode disease were compared through 10 rounds of iterations, and the average temperature from June to August, the average precipitation from June to August, the total amount of pine wood transported in, the amount of pine wood transported in from the epidemic area, the road density, and the host insect control rate were selected as the significant influencing factors of pine wood nematode disease.

[0079] S2.3. Extract the data of significant pine wood nematode disease influencing factors selected in step S2.2, and perform data preprocessing, including data standardization, missing value processing, and normalization processing;

[0080] Furthermore, there are large differences in the distribution of sample values, so it is necessary to preprocess the samples to eliminate the correlation between various attributes in order to obtain more satisfactory results. Common data preprocessing methods mainly include: data standardization, missing value processing, normalization, and clustering. Among them, normalization (Resaling) is to map the data to a specific area and separate different dimensions and quantity units. Commonly used mapping domains are [0, -1] and [-1, 1]. The advantages of normalization: its operation is simple, the concept is clear, it is easy to understand and operate, and it can improve the convergence speed and accuracy of the training model. In addition, in the training of the model, the standardization method is used to avoid problems such as gradient explosion and gradient disappearance, thereby enhancing the robustness of the model;

[0081] S2.4. Use the data preprocessed in step S2.3 to construct a training set, a validation set, and a test set, using 80% of the data in the 2017-2021 data set as the training set, 20% of the data as the validation set, and the 2022 data set as the test set;

[0082] S3. Construct a pine wood nematode disease propagation prediction model based on the Kalman-LSTM-NGO combination, and use the training set obtained in step S2 to train the model to obtain a trained pine wood nematode disease propagation prediction model based on the Kalman-LSTM-NGO combination;

[0083] Furthermore, the specific implementation method of step S3 includes the following steps:

[0084] S3.1. Use Kalman filter algorithm to perform noise reduction on the preprocessed data obtained in step S2;

[0085] Furthermore, the Kalman filter is an iterative method based on the state space model. It can not only predict stable sequences, but also unstable data. The main contents of the model are three aspects: first, by using state parameters to reflect the intrinsic characteristics of the system; second, establish a state equation that can reflect the current state transfer and change trend; third, by constructing an observation equation that can reflect the relationship between the state and the observed data. The Kalman filter algorithm requires real-time acquisition of relevant information and parameter updates, noise reduction of other interferences, prediction based on measured data, and continuous removal of interference to obtain the final correction. The algorithm combines state error and observation error to find the optimal estimate, and at the same time, through the update of real-time data, it can correct the prediction result of the previous step. In the presence of interference, this algorithm has high accuracy. The algorithm has a strong dependence on the model. The introduction of observation data will correct the estimated data to prevent the error of the estimated data from being too large, and use the parameters of the control input and the output variable to estimate the state quantity.

[0086] S3.2. Construct a Kalman model, input the training set data after denoising in step S3.1 into the Kalman model for processing, and obtain the residual sequence of the Kalman model prediction value;

[0087] S3.2.1. Parameter initialization: Set the observation value at the start time to The expression is:

[0088]

[0089] P 0 / 0 =∑x0

[0090] Among them, m X0 and P 0 / 0 Respectively represent the expectation and variance of the initial state vector x0;

[0091] S3.2.2. Based on the initial state vector, use the prediction formula to make a prediction and get the estimated value The expression is:

[0092]

[0093] Wherein, H0, R0, L0 represent the first initialization value, the second initialization value, and the third initialization value of the state parameter;

[0094] S3.2.3. Calculate the error covariance P using the observed value at the start time of step S3.2.1, then update the Kalman gain coefficients Q and R, perform recursive operations, correct the estimated value, and update the error covariance until the iterative calculation is completed, and output the residual sequence of the Kalman model prediction value;

[0095] Furthermore, when the statistical characteristics of the target initial state information are unclear, the following expression can be set:

[0096]

[0097] When the initial state of the Kalman filter meets the following formula, X is processed in the filter processing k and P k,k The estimates will all be unbiased:

[0098]

[0099] When the Kalman filter algorithm has a good initial value and a covariance matrix of noise, the Kalman filter can converge quickly. On the contrary, when the Kalman filter is used, the prediction result of the Kalman filter is very different from the actual situation, and the predicted value of the Kalman filter has a large discreteness. Therefore, in the Kalman filter algorithm, its initial vector assignment is often not 0. Introducing the collected observation data into the model for statistical processing can reduce the prediction deviation caused by the unclear statistical characteristics of the initial value, thereby reducing the prediction error caused by the insignificant initial value, thereby reducing the impact of this factor on the monitoring estimation results.

[0100] Furthermore, the determination of the initial value assumes that at the initial time t0, the estimated value of the dynamic target is The target estimate at time t1 is Then the target state estimate is:

[0101]

[0102] The initial value of the Kalman filter algorithm is

[0103] Furthermore, the initial state variance P 0,0 Since the initial value of the error variance generally requires accurate calculation of the observation vector to obtain the variance of the initial state vector, it is assumed that the error variance P of the initial state vector in the equation is 0,0 is a zero matrix.

[0104] S3.3. Construct an LSTM neural network model, input the residual sequence of the Kalman model prediction value obtained in step S3.2 into the LSTM neural network model for training, update the LSTM neural network model parameters based on the Adam error back propagation method, and obtain the Kalman-LSTM neural network model;

[0105] Furthermore, the main idea of ​​using the Kalman-LSTM combined prediction method to analyze, process and predict pine wood nematode disease is: considering that the modeling of the LSTM network and the selection of training samples have high requirements, the Kalman filter algorithm is used to reduce the noise of the observed data at the beginning, and it is transmitted as input data to the LSTM neural network for processing and analysis. Then the early data was studied and fitted using the LSTM network to observe the fitting effect. If the fitting result is not satisfactory, it is necessary to continuously adjust the weights and thresholds through back propagation. If the fitting effect meets the test, or the actual deviation is small, the model is used to predict the number of infections in the future period;

[0106] S3.3.1. Construct an LSTM neural network model and use it to fit and predict the number of pine wood nematode disease cases. The LSTM neural network model unit consists of an input gate, a forget gate, and an output gate. The LSTM neural network model selectively allows information to pass through the gate structure to update or retain historical information. The expression of the LSTM neural network model is:

[0107] f t =σ(W f ·[h t-1 ,x t ]+b f )

[0108] i t =σ(W i ·[h t-1 ,x t ]+b i )

[0109]

[0110]

[0111] h t =o t tanh(C t )

[0112] Among them, f t represents the forget gate, σ represents the Sigmoid function, W f represents the weight of the forget gate, h t-1represents the hidden state at time t-1, x t represents the input at time t, b f represents the bias of the forget gate, i t represents the input gate, W i represents the weight of the input gate, b i represents the bias of the input gate, represents the candidate state value of the input unit at time t, tanh represents the Tanh function, W c represents the weight of the output gate, b c represents the deviation of the output gate, C t and C t -1 represents the unit state update value at time t and time t-1, h t represents the hidden state at time t, o t represents the output gate;

[0113] S3.3.2. Input the residual sequence of the Kalman model prediction value obtained in step S3.2 into the LSTM neural network model for training, and update the LSTM neural network model parameters based on the Adam error back propagation method. The expression is:

[0114] m t =μ×m t-1 +(1-μ)×g t

[0115]

[0116]

[0117] Among them, m t represents the first-order moment estimate of the gradient at time t, m t-1 represents the first-order moment estimate of the gradient at time t-1, n t represents the second-order moment estimate of the gradient, n t-1 represents the second-order moment estimate of the gradient at time t-1, represents the gradient first-order moment estimate m t Correction, Represents the second-order moment estimate of the gradient n t Correction, g t represents the gradient value, v represents the second attenuation rate, △θ t represents the update value of the parameter, μ represents the first decay rate;

[0118] Furthermore, based on the infection of pine wood nematode disease, a prediction model was established using LSTM neural network, and the average deviation was used as the loss function during model training. The Adam model with adaptive moment estimation was used, which includes both the estimation of the first-order moment and the estimation of the second-order moment. Compared with other random optimization methods, the Adam method has fast operation speed, small memory usage, small parameter adjustment, and is suitable for solving problems with a large amount of data and parameters. Among them, the following issues should be considered when building the model: learning rate, number of iterations, time window period and number of network layers, number of hidden nodes and Dropout ratio. First, when the model is trained, it is not enough to iterate the required data once, and multiple iterations are required to enable the model to obtain better fitting ability. If the number of iterations is too small, the model cannot obtain correct prediction results because it cannot be fully learned. On the contrary, if the number of iterations is too large, the model will overfit due to excessive learning, affecting the prediction accuracy of the model on the test data set. Secondly, the number of hidden layers of the LSTM network is closely related to the extraction of sequence features. Too many or too few will affect the prediction accuracy of the sequence. In certain cases, increasing the number of layers can improve the prediction accuracy of the model, but increasing the number of layers too much may lead to computational redundancy, prolong the training time of the model, and the effect of improving the prediction accuracy is not obvious. Finally, there are some regulations on the number of nodes in the hidden layer. If there are too few hidden nodes in the network, then the association of hidden features cannot be obtained without complete understanding; when selecting the hidden layer, a rough area is determined based on experience, and then the most suitable position is determined after multiple experiments.

[0119] S3.4. The NGO algorithm is used to optimize the number of neurons in the hidden layer of the Kalman-LSTM neural network model obtained in step S3.3, the learning rate and the number of training times, and a trained pine wood nematode disease propagation prediction model based on the Kalman-LSTM-NGO combination is obtained;

[0120] Further, initialization, in the NGO algorithm, the northern goshawk population is represented by the following population matrix:

[0121]

[0122] Among them, X is the matrix of the northern goshawk population, X i is the position of the i-th northern goshawk, X i,j is the position of the i-th northern goshawk in the j-th dimension, N is the population size of the northern goshawk, and m is the dimension of the problem to be solved.

[0123] In the Northern Goshawk algorithm, the objective function value can be expressed by the objective function value vector:

[0124]

[0125] Among them, F is the objective function vector of the northern goshawk population, F i is the objective function value of the i-th northern goshawk.

[0126] Furthermore, in the prey identification and attack phase, the northern goshawk first randomly selects a prey as the hunting object and then launches an attack. Since the prey is randomly selected in the search space, the exploration ability of the NGO algorithm is enhanced. In order to determine the best area, the NGO algorithm conducts a comprehensive search of the search space in this phase.

[0127] The formula for the northern goshawk's prey selection and attack behavior is expressed as:

[0128] P i =X k ,i=1,2,...,N,k=1,2,...,i-1,i+1,N

[0129]

[0130] Among them, P i is the position of the i-th northern goshawk’s prey, F Pi is the objective function value of the location of the i-th northern goshawk’s prey, k is a random integer in the range [1, N] except i, is the new position of the i-th northern goshawk, is the new position of the i-th northern goshawk in the j-th dimension, F i new,P1 is the objective function value of the i-th northern goshawk after the update based on the first stage, r is a random number in the range of [0,1], and I is a random integer with a value of 1 or 2.

[0131] Furthermore, in the chasing and escaping phase, after the northern goshawk attacks the prey, the prey will try to escape. In the process of chasing the prey to the end, the northern goshawk will continue to chase the prey. The northern goshawk has a very high attack speed, which makes it almost impossible to miss the prey during the hunting process. The simulation of the northern goshawk's chasing and prey escape behavior can greatly improve the algorithm's local search ability in the search space. Assuming that the northern goshawk's hunting activities are close to an attack range of radius R, the formula for the chasing and escaping phase is expressed as:

[0132]

[0133]

[0134]

[0135] Among them, t is the current iteration number, T is the maximum iteration number, is the new position of the i-th northern goshawk, is the new position of the i-th northern goshawk in the j-th dimension, X i,j is the new position of the i-th northern goshawk in the j-th dimension after the second stage update, F i new,P2 is the objective function value of the i-th northern goshawk after the second stage update.

[0136] The NGO algorithm was used to optimize the three parameters of the Kalman-LSTM algorithm: the number of hidden layer nodes, the initial learning rate, and the L2 regularization coefficient. The Kalman-LSTM model was constructed using the diseased tree data to predict the number of cases. The maximum number of iterations of the Kalman-LSTM model was set to complete the model fitting and prediction.

[0137] S4. Use the validation set obtained in step S2 to verify, tune and test the pine wood nematode disease propagation prediction model based on the Kalman-LSTM-NGO combination trained in step S3 to obtain a tuned pine wood nematode disease propagation prediction model based on the Kalman-LSTM-NGO combination;

[0138] Furthermore, step S4 uses mean absolute error MAE and mean absolute error percentage MAPE as prediction error evaluation indicators;

[0139] The expression of mean absolute error MAE is:

[0140]

[0141] in, Represents the predicted value, y k represents the true value;

[0142] The expression of mean absolute error percentage MAPE is:

[0143]

[0144] S5. Use the test set obtained in step S2 to predict the pine wood nematode disease propagation prediction model based on the Kalman-LSTM-NGO combination optimized in step S4 to obtain a pine wood nematode disease propagation prediction result;

[0145] S6. Visualize the prediction results of pine wood nematode disease spread obtained in step S5 to help decision makers analyze measures on fertilization, thinning, and felling based on the visualization results, and propose optimization suggestions including pine tree epidemic monitoring, control measures, and resource allocation;

[0146] Furthermore, the prediction results of pine wood nematode disease spread obtained in step S5 are visualized by marking the disease prediction areas on a national map using the geographic information system ArcGIS.

[0147] Further specific prevention and control measures are provided, including measures in fertilization, thinning, and felling, including suggestions on pine epidemic monitoring, control measures, and optimization of resource allocation. According to the different situations of the prediction results and model analysis, corresponding priority recommendations are provided for different regions or areas with different disease severity to help decision makers reasonably formulate the urgency of prevention and control work. In order to maintain the accuracy of the recommendations, it is necessary to update the data regularly, retrain the model, and continuously improve the model and decision-making strategy based on the new data. This cyclic process aims to continuously improve the prediction and response strategies to ensure timely and accurate response to the epidemic. Specific implementation method 2:

[0149] An electronic device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of a method for predicting the spread of pine wood nematode disease described in a specific implementation mode 1 are implemented.

[0150] The computer device of the present invention may be a device including a processor and a memory, such as a single chip microcomputer including a central processing unit, etc. Moreover, the processor is used to implement the steps of the above-mentioned method for predicting the spread of pine wood nematode disease when executing the computer program stored in the memory.

[0151] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0152] The memory may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Specific implementation method three:

[0154] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements a method for predicting the spread of pine wood nematode disease as described in a first specific implementation manner.

[0155] The computer-readable storage medium of the present invention can be any form of storage medium that can be read by a processor of a computer device, including but not limited to non-volatile memory, volatile memory, ferroelectric memory, etc. A computer program is stored on the computer-readable storage medium. When the processor of the computer device reads and executes the computer program stored in the memory, the steps of the above-mentioned method for predicting the spread of pine wood nematode disease can be implemented.

[0156] The computer program includes computer program code, which may be in source code form, object code form, executable file or some intermediate form, etc. The computer readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer readable medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer readable media do not include electric carrier signals and telecommunication signals.

[0157] It should be noted that relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0158] Although the present application has been described above with reference to specific embodiments, various modifications may be made thereto and parts thereof may be replaced with equivalents without departing from the scope of the present application. In particular, as long as there is no structural conflict, the various features in the specific embodiments disclosed in the present application may be used in combination with each other in any manner, and the fact that these combinations are not exhaustively described in this specification is only for the sake of omitting space and saving resources. Therefore, the present application is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A method for predicting the spread of pine wood nematode disease, characterized in that: The steps include: S1. Data collection, including historical disease occurrence data in diseased tree distribution areas, meteorological data and geographic information data in diseased tree distribution areas; The historical disease occurrence data of the diseased tree distribution area in step S1 includes location, epidemic situation, time and disease severity information data, and the meteorological data and geographic information data of the diseased tree distribution area include average annual temperature, average annual maximum temperature, average annual minimum temperature, humidity, average annual precipitation, normalized vegetation index, topography, soil type and altitude; S2. Preprocess the data collected in step S1, and then construct a training set, a validation set, and a test set; The specific implementation method of step S2 includes the following steps: S2.

1. For the data collected in step S1, select the factors affecting pine wood nematode disease, including human activity factors and natural environment factors. The human activity factors include the total amount of pine wood transported in, the amount of pine wood transported in from the epidemic area, the road density and the host insect control rate. The natural environment factors include the average temperature from June to August, the average precipitation from June to August, the average altitude, the host insect infestation area and the slope; S2.

2. The Pearson correlation coefficient is used to perform correlation analysis on the influencing factors of pine wood nematode disease selected in step S1. The expression is: Among them, x represents the data of pine wood nematode disease, y represents the data of the influencing factors of pine wood nematode disease, ρ x,y represents the correlation coefficient, and N represents the number of variable values; Based on correlation analysis, the influencing factors of pine wood nematode disease were compared through 10 rounds of iterations, and the average temperature from June to August, the average precipitation from June to August, the total amount of pine wood transported in, the amount of pine wood transported in from the epidemic area, the road density and the host insect control rate were selected as the significant influencing factors of pine wood nematode disease. S2.

3. Extract the data of significant pine wood nematode disease influencing factors selected in step S2.2, and perform data preprocessing, including data standardization, missing value processing and normalization processing; S2.

4. Use the data preprocessed in step S2.3 to construct a training set, a validation set, and a test set, using 80% of the data in the 2017-2021 data set as the training set, 20% of the data as the validation set, and the 2022 data set as the test set; S3. Construct a pine wood nematode disease propagation prediction model based on the Kalman-LSTM-NGO combination, and use the training set obtained in step S2 to train the model to obtain a trained pine wood nematode disease propagation prediction model based on the Kalman-LSTM-NGO combination; The specific implementation method of step S3 includes the following steps: S3.

1. Use Kalman filter algorithm to perform noise reduction on the preprocessed data obtained in step S2; S3.

2. Construct a Kalman model, input the training set data after denoising in step S3.1 into the Kalman model for processing, and obtain the residual sequence of the Kalman model prediction value; S3.2.

1. Parameter initialization: Set the observation value at the start time to The expression is: P 0 / 0 =∑x0 Among them, m X0 and P 0 / 0 Respectively represent the expectation and variance of the initial state vector x0; S3.2.

2. Based on the initial state vector, use the prediction formula to make a prediction and get the estimated value The expression is: Wherein, H0, R0, L0 represent the first initialization value, the second initialization value, and the third initialization value of the state parameter; S3.2.

3. Calculate the error covariance P using the observed value at the start time of step S3.2.1, then update the Kalman gain coefficients Q and R, perform recursive operations, correct the estimated value, and update the error covariance until the iterative calculation is completed, and output the residual sequence of the Kalman model prediction value; S3.

3. Construct an LSTM neural network model, input the residual sequence of the Kalman model prediction value obtained in step S3.2 into the LSTM neural network model for training, update the LSTM neural network model parameters based on the Adam error back propagation method, and obtain the Kalman-LSTM neural network model; S3.3.

1. Construct an LSTM neural network model and use it to fit and predict the number of pine wood nematode disease cases. The LSTM neural network model unit consists of an input gate, a forget gate, and an output gate. The LSTM neural network model selectively allows information to pass through the gate structure to update or retain historical information. The expression of the LSTM neural network model is: f t =σ(W f ·[h t-1 ,x t ]+b f ) i t =σ(W i ·[h t-1 ,x t ]+b i ) h t =o t ·tanh(C t ) Among them, f t represents the forget gate, σ represents the Sigmoid function, W f represents the weight of the forget gate, h t-1 represents the hidden state at time t-1, x t represents the input at time t, b f represents the bias of the forget gate, i t represents the input gate, W i represents the weight of the input gate, b i represents the bias of the input gate, represents the candidate state value of the input unit at time t, tanh represents the Tanh function, W c represents the weight of the output gate, b c represents the deviation of the output gate, C t and C t -1 represents the unit state update value at time t and time t-1, h t represents the hidden state at time t, o t represents the output gate; S3.3.

2. Input the residual sequence of the Kalman model prediction value obtained in step S3.2 into the LSTM neural network model for training, and update the LSTM neural network model parameters based on the Adam error back propagation method. The expression is: m t =μ×m t-1 +(1-μ)×g t Among them, m t represents the first-order moment estimate of the gradient at time t, m t-1 represents the first-order moment estimate of the gradient at time t-1, n t represents the second-order moment estimate of the gradient at time t, n t-1 represents the second-order moment estimate of the gradient at time t-1, Represents the first-order moment estimate m of the gradient at time t t Correction, represents the second-order moment estimate of the gradient at time t n t Correction, g t represents the gradient value, v represents the second attenuation rate, △θ t represents the update value of the parameter, μ represents the first decay rate; S3.

4. The NGO algorithm is used to optimize the number of neurons in the hidden layer of the Kalman-LSTM neural network model obtained in step S3.3, the learning rate and the number of training times, and a trained pine wood nematode disease propagation prediction model based on the Kalman-LSTM-NGO combination is obtained; S4. Use the validation set obtained in step S2 to verify, tune and test the pine wood nematode disease propagation prediction model based on the Kalman-LSTM-NGO combination trained in step S3 to obtain a tuned pine wood nematode disease propagation prediction model based on the Kalman-LSTM-NGO combination; S5. Use the test set obtained in step S2 to predict the pine wood nematode disease propagation prediction model based on the Kalman-LSTM-NGO combination optimized in step S4 to obtain a pine wood nematode disease propagation prediction result; S6. Visualize the prediction results of pine wood nematode disease spread obtained in step S5 to help decision makers analyze fertilization, thinning and felling measures based on the visualization results, and put forward optimization suggestions including pine tree epidemic monitoring, control measures and resource allocation.

2. A method for predicting the spread of pine wood nematode disease according to claim 1, characterized in that: Step S4 uses mean absolute error MAE and mean absolute error percentage MAPE as prediction error evaluation indicators; The expression of mean absolute error MAE is: in, Represents the predicted value, y k represents the true value; The expression of mean absolute error percentage MAPE is:

3. A method for predicting the spread of pine wood nematode disease according to claim 2, characterized in that: The prediction results of pine wood nematode disease spread obtained in step S5 are visualized by marking the disease prediction areas on a national map using the geographic information system ArcGIS.

4. An electronic device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of a method for predicting the spread of pine wood nematode disease as claimed in any one of claims 1 to 3 when executing the computer program.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting the spread of pine wood nematode disease according to any one of claims 1 to 3 is implemented.

Images