African swine fever monitoring method and device, electronic equipment and storage medium

Abstract

The invention discloses an African swine fever monitoring method and device, electronic equipment and a storage medium, and relates to the technical field of livestock breeding epidemic disease monitoring. Physiological parameters, environmental parameters and flow information parameters of infected pigs are collected through the equipment; establishing a parameter matrix of each parameter according to a time sequence, analyzing the parameter matrix, establishing a judgment matrix, determining the weight of each parameter, and screening parameters having significant influence on ASF propagation to form an influence matrix; constructing a search space based on the influence matrix, and optimizing an adaptive coefficient of an ELM network through an ENAS search algorithm by taking the search space as a constraint to obtain an adaptive ELM network model; fusing the parameters into a feature vector by using an entropy weight method; and inputting the feature vector into an adaptive ELM network model to obtain a risk assessment value, dividing a risk area in combination with a risk threshold, and performing corresponding monitoring based on a preset risk area monitoring scheme. According to the invention, targeted monitoring of the risk area of the infected pig is realized.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of livestock breeding epidemic disease monitoring, and more particularly to an African swine fever monitoring method and device, an electronic device and a storage medium. BACKGROUND

[0002] African swine fever (ASF) has occurred, spread and prevailed in many countries worldwide. China is a large country in pig breeding and pork consumption, and the live pig output, inventory and pork consumption are all at the top of the world. The total annual import of breeding pigs and pork products is huge, and trade with many countries is frequent. Therefore, the direct and indirect losses caused by ASF are immeasurable.

[0003] African swine fever is an acute and highly contagious disease caused by African swine fever virus, with a mortality rate of 100%. The current monitoring method relies on fluorescent quantitative PCR detection, but due to the latent period of the virus and environmental factors, there is a high misjudgment rate and low efficiency of artificial judgment in risk areas. In the prior art, it is difficult for farms to integrate real-time physiological parameters of pigs, environmental data and flow information, which leads to the inability to accurately locate the risk area and the difficulty in achieving early warning and prevention and control.

[0004] Therefore, how to monitor the risk area of infected pigs is a problem that needs to be solved by those skilled in the art. SUMMARY

[0005] Therefore, the application provides an African swine fever monitoring method, device, electronic device and storage medium, which realizes targeted monitoring of the risk area of infected pigs.

[0006] In order to achieve the above purpose, the application adopts the following technical solutions:

[0007] An African swine fever monitoring method comprises the following steps:

[0008] Physiological parameters, environmental parameters and flow information parameters of infected pigs are collected based on RFID tags, infrared sensors and environmental monitoring equipment;

[0009] A parameter matrix of each parameter is established in chronological order, the parameter matrix is analyzed by using an analytic hierarchy process, a decision matrix is established, the weight of each parameter is determined, and parameters that have a significant impact on ASF transmission are screened to form an influence matrix;

[0010] An adaptive ELM network model is obtained by optimizing the adaptive coefficients of the ELM network through an ENAS search algorithm based on the search space.

[0011] The weights of the parameters are calculated by using an entropy weight method, and the parameters are fused into a feature vector.

[0012] Input the feature vector into the adaptive ELM network model to obtain a risk assessment value, divide a risk region in combination with a risk threshold value, and perform corresponding monitoring based on a preset risk region monitoring scheme.

[0013] Preferably, the physiological parameters include first physiological parameters in an acute phase and second physiological parameters in a latent phase of the infected pigs; the environmental parameters include first environmental parameters in an active area and second environmental parameters in an inactive area of the infected pigs in different seasons; the flow information parameters include first flow parameters in the acute phase and second flow parameters in the latent phase of the infected pigs; the first physiological parameters and the second physiological parameters both include body temperature, food intake, and activity amount; the first environmental parameters and the second environmental parameters both include temperature, humidity, and air quality; and the first flow parameters and the second flow parameters both include personnel flow records and material flow records.

[0014] Preferably, the method further comprises: performing time series mean square deviation calculation on the first physiological parameters in the acute phase of the infected pigs, and the formula is:

[0015]

[0016] wherein m is the sampling number, is the mean value of the acute phase parameters; a mean square deviation threshold value kσ is set, and parameters exceeding the mean square deviation threshold value are determined as third physiological parameters;

[0017] The first environmental parameters are subjected to seasonal adaptability analysis, the fluctuation range of parameters deviating from the seasonal normal value is calculated, and parameters deviating from the seasonal mean value are determined as third environmental parameters.

[0018] The personnel flow records and the material flow records entering the infected area ≥ 2 times per day without disinfection are determined as third flow parameters.

[0019] Preferably, the establishment of the decision matrix specifically comprises:

[0020] Parameter matrices of the same type of parameters are established according to time series, and the matrix elements are parameter values at different times;

[0021] A decision matrix A = [a ij ] n×n , which satisfies: wherein the elements in the decision matrix are a ij , a ij represents the importance of the i th parameter relative to the j th parameter, n is the total number of parameters, 1 ≤ i ≤ n, and 1 ≤ j ≤ n.

[0022] A consistency test coefficient of the decision matrix is judged, and the decision matrix is adjusted according to a judgment result until the consistency test coefficient of the decision matrix is less than a preset threshold; an influence matrix is formed by N parameters in the decision matrix in order.

[0023] Preferably, the constructing a search space based on the influence matrix specifically comprises:

[0024] The first physiological parameter, the first environmental parameter, the second physiological parameter and the second environmental parameter are taken as first search parameters, and the third physiological parameter, the third environmental parameter and the third flow information parameter are taken as second search parameters.

[0025] A first element in the influence matrix of the same type of first search parameter is taken as an origin, a distance between other elements and the origin is calculated as a radius to construct a first search circle, and a coverage range of the first search circle is taken as a first search space.

[0026] A maximum element in the influence matrix of the same type of second search parameter is taken as an edge point of a second search circle, a distance between other elements and the origin is calculated as a radius to construct the second search circle, and a coverage range of all the second search circles is taken as a second search space.

[0027] The influence matrix of the first physiological parameter, the second physiological parameter and the third physiological parameter corresponds to a first physiological search space, a second physiological search space and a third physiological search space; the influence matrix of the first environmental parameter, the second environmental parameter and the third environmental parameter corresponds to a first environmental search space, a second environmental search space and a third environmental search space; and the influence matrix of the third flow information parameter corresponds to a third flow search space.

[0028] Preferably, the acquiring the adaptive ELM network model specifically comprises:

[0029] The same number of parameter values as the number of ELM network channels P are acquired from each search space, and a difference value ΔD between the parameter values of the first and second physiological search spaces, the first and second environmental search spaces and the third physiological search space, the third environmental search space and the third flow search space is calculated, and a formula is as follows:

[0030]

[0031] Wherein v phys,i and v env,i are the first and second search space parameter values, v phy3,i , v env3,i , v flow3,i are the third search space parameter values.

[0032] The coefficient λ is adjusted by a step of 0.2, and an adaptive value is obtained by substituting the adaptive Simpson formula function.

[0033]

[0034] f(AD, l) = l AD + (1 - l) base;

[0035] wherein, l0, l1, l2 are coefficient values of continuous steps, f(AD, l) is a mapping function based on difference value and coefficient, base is a historical epidemic data reference value;

[0036] The adaptive value is used as an adaptive coefficient of the trained ELM network.

[0037] Preferably, it further comprises:

[0038] The monitoring information obtained by monitoring the risk area based on the preset risk area monitoring scheme within the monitoring duration is acquired.

[0039] The analysis result is obtained by analyzing the monitoring information, and the monitoring of the risk area is released in the case that the analysis result matches the monitoring release condition.

[0040] An African swine fever monitoring device comprises:

[0041] A parameter acquisition module acquires physiological parameters, environmental parameters and flow information parameters of infected pigs based on RFID tags, infrared sensors and environmental monitoring equipment;

[0042] A matrix establishment module establishes a parameter matrix of each parameter in a time sequence, analyzes the parameter matrix by using an analytic hierarchy process, establishes a decision matrix, determines the weight of each parameter, screens parameters that have a significant impact on the spread of ASF to form an impact matrix;

[0043] A model optimization module constructs a search space based on the impact matrix, optimizes the adaptive coefficient of the ELM network by using an ENAS search algorithm with the search space as a constraint, and acquires an adaptive ELM network model;

[0044] A vector fusion module calculates the weight of each parameter by using an entropy weight method, and fuses each parameter into a feature vector;

[0045] A monitoring module inputs the feature vector into the adaptive ELM network model to obtain a risk assessment value, divides a risk area in combination with a risk threshold, and performs corresponding monitoring based on a preset risk area monitoring scheme.

[0046] A computer device comprises a processor and a memory, and the processor is used to execute an African swine fever monitoring and early warning program stored in the memory.

[0047] A storage medium stores one or more programs that are executable by one or more processors.

[0048] Compared with the prior art, the technical scheme can provide a monitoring method and device for African swine fever, an electronic device and a storage medium, and the beneficial effects are as follows:

[0049] (1) Based on RFID tags, infrared sensors and environmental monitoring equipment, physiological parameters, environmental parameters and flow information parameters of infected pigs are collected, and multi-dimensional coverage of African swine fever monitoring data is realized. Physiological parameters include body temperature, food intake and activity of infected pigs in acute and latent periods, environmental parameters include temperature, humidity and air quality of infected pigs in active and inactive areas in different seasons, and flow information parameters include personnel flow records and material flow records of infected pigs in acute and latent periods. Compared with single parameter monitoring, this multi-parameter collection method can more comprehensively reflect various factors related to African swine fever infection and provide rich data support for accurate risk assessment.

[0050] (2) The parameter matrix of each parameter is established according to time sequence, the parameter matrix is analyzed by using the analytic hierarchy process, the decision matrix is established, the weight of each parameter is determined, the parameters with significant influence on ASF transmission are screened to form the influence matrix; important parameters are screened out to form the influence matrix, so that the subsequent model optimization and risk assessment can focus on key factors, and the accuracy and efficiency of monitoring are improved.

[0051] (3) The construction of the search space combines the characteristics of normal parameters and abnormal parameters, and the first search space and the second search space are constructed by different ways, which can better capture the rules and abnormalities in the data. The ENAS search algorithm optimizes the adaptive coefficients of the ELM network under the constraint of the search space, so that the model can adaptively adjust according to the actual data characteristics, improve the adaptability of the model to different breeding scenes and epidemic situations, and more accurately evaluate the risk of African swine fever. BRIEF DESCRIPTION OF DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced as follows. Obviously, the drawings in the following description are only embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor on the basis of the provided drawings.

[0053] Figure 1 The structural schematic diagram provided by the present application. DETAILED DESCRIPTION

[0054] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0055] The embodiment of the present invention discloses a method for monitoring African swine fever, such as Figure 1 Shown, including:

[0056] Based on RFID tags, infrared sensors and environmental monitoring equipment, the physiological parameters, environmental parameters and flow information parameters of infected pigs are collected;

[0057] Establish a parameter matrix for each parameter according to the time series, analyze the parameter matrix using the analytic hierarchy process, establish a decision matrix, determine the weight of each parameter, and select the parameters that have a significant impact on ASF propagation to form an impact matrix;

[0058] A search space is constructed based on the influence matrix. With the search space as a constraint, the adaptive coefficient of the ELM network is optimized through the ENAS search algorithm to obtain an adaptive ELM network model.

[0059] The entropy weight method is used to calculate the weight of each parameter and fuse each parameter into a feature vector;

[0060] The feature vector is input into the adaptive ELM network model to obtain the risk assessment value, which is then divided into risk areas based on the risk threshold, and corresponding monitoring is performed based on the preset risk area monitoring plan.

[0061] In a specific embodiment, the physiological parameters include the first physiological parameters of the infected pigs in the acute phase and the second physiological parameters of the incubation phase; the environmental parameters include the first environmental parameters of the infected pigs' activity areas and the second environmental parameters of the inactive areas in different seasons; the flow information parameters include the first flow parameters of the infected pigs in the acute phase and the second flow parameters of the incubation phase; the first physiological parameters and the second physiological parameters both include body temperature, food intake, and activity level; the first environmental parameters and the second environmental parameters both include temperature, humidity, and air quality; the first flow parameters and the second flow parameters both include personnel flow records and material flow records.

[0062] In a specific embodiment, the material flow record refers to items that have come into contact with pigs, such as brushes, food bowls, etc.

[0063] In a specific embodiment, the method further includes: calculating the time series mean square error of the first physiological parameter of the infected pigs in the acute phase, using the formula:

[0064]

[0065] wherein m is the number of sampling times, is the mean value of the acute phase parameters; set the mean square deviation threshold value as kσ, and determine the parameters exceeding the mean square deviation threshold value as the third physiological parameters;

[0066] perform seasonal adaptability analysis on the first environmental parameters, calculate the fluctuation amplitude of the parameters deviating from the seasonal normal value, and determine the parameters deviating from the seasonal mean value as the third environmental parameters;

[0067] record the personnel flow records of people who have not been disinfected entering the infection area ≥2 times / day and the material flow records entering the infection area as the third flow parameters.

[0068] In one specific embodiment, the fluctuation amplitude of the parameters deviating from the seasonal normal value is determined by the following parameters: temperature deviating from the seasonal normal value by ±5℃ or more; humidity >90% or <30%; air quality index exceeding the safety threshold by 2 times or more.

[0069] In one specific embodiment, establishing the decision matrix specifically includes:

[0070] establishing a parameter matrix of the same type of parameters in time sequence, and the matrix elements are the parameter values at different times;

[0071] constructing a decision matrix A=[a ij ] n×n , satisfying: wherein the elements in the decision matrix are a ij , a ij represents the importance of the i-th parameter relative to the j-th parameter, n is the total number of parameters, 1≤i≤n, and 1≤j≤n;

[0072] determine the consistency test coefficient of the decision matrix, and adjust the decision matrix according to the determination result until the consistency test coefficient of the decision matrix is less than a preset threshold value; and the first N parameters in the decision matrix form an influence matrix.

[0073] In one specific embodiment, constructing a search space based on the influence matrix specifically includes:

[0074] taking the first physiological parameters, the first environmental parameters, the second physiological parameters and the second environmental parameters as the first search parameters, and taking the third physiological parameters, the third environmental parameters and the third flow information parameters as the second search parameters;

[0075] taking the first element in the influence matrix of the same type of first search parameters as the origin, calculating the distance between other elements and the origin as the radius to construct a first search circle, and forming a first search space by the coverage range of the first search circle;

[0076] The maximum element in the influence matrix of the second search parameter of the same kind is taken as the edge point of the second search circle, the distance between other elements and the origin is calculated as the radius to construct the second search circle, and the coverage of all second search circles is combined to form the second search space;

[0077] The influence matrix of the first physiological parameter, the second physiological parameter, and the third physiological parameter corresponds to the first physiological search space, the second physiological search space, and the third physiological search space; the influence matrix of the first environmental parameter, the second environmental parameter, and the third environmental parameter corresponds to the first environmental search space, the second environmental search space, and the third environmental search space; and the influence matrix of the third flow information parameter corresponds to the third flow search space.

[0078] In the construction of the search space based on the influence matrix, only the third flow information parameter is used without using the first and second flow information parameters, so as to avoid the inclusion of normal flow data of non-infected pigs, prevent the dilution of low-risk information with high-risk features, and make the search space cover only the transmission key nodes related to infected pigs. In the optimization of the ENAS algorithm, the difference value calculation of the flow parameter directly focuses on the absolute value weight of the high-risk flow record, avoiding the redundant calculation of the first / second flow parameter.

[0079] In one specific embodiment, obtaining the adaptive ELM network model specifically includes:

[0080] The same number of parameter values as the number of ELM network channels P are obtained from each search space, and the difference value ΔD of the parameter values of the first and second physiological search spaces, the first and second environmental search spaces, and the third physiological search space, the third environmental search space, and the third flow search space is calculated, and the formula is:

[0081]

[0082] wherein v phys,i and v env,i are the first and second search space parameter values, v phy3,i , v env3,i , and v flow3,i are the third search space parameter values.

[0083] The adjustment coefficient λ is adjusted by a step of 0.2, and the adaptive value is obtained by substituting the adaptive Simpson formula function;

[0084]

[0085] f(ΔD, λ) = λ·ΔD + (1-λ)·base;

[0086] wherein λ0, λ1, and λ2 are the coefficient values of consecutive steps, f(ΔD, λ) is a mapping function based on the difference value and the coefficient, and base is a historical epidemic data reference value.

[0087] The adaptive value is used as an adaptive coefficient of the trained ELM network.

[0088] In one specific embodiment, the fusion feature vector includes: processing the influence parameters at the current moment using the entropy weight method to obtain the weights of each influence parameter; forming normalized sequences of the influence parameters at the current moment respectively; fusing all normalized sequences and forming a fusion matrix after multiplying each normalized sequence by the weight of the corresponding influence parameter; performing dimension reduction processing on the fusion matrix using the principal component analysis method to obtain a feature matrix; and flattening the feature matrix using a flatten function to convert it into a feature vector.

[0089] The feature vector is input into the adaptive ELM network model, so that the adaptive ELM network model learns the feature vector according to the adaptive coefficient corresponding to the channel, thereby outputting a risk assessment value; wherein the number of nodes of the input layer of the adaptive ELM network model is the same as the dimension of the feature vector, and the number of nodes of the output layer is M; the activation function is a Sigmiod function, a ReLU function or a Tanh function, and the number of hidden layer nodes is 25-200.

[0090] In one specific embodiment, it further includes:

[0091] Obtaining monitoring information obtained by monitoring the risk area within a monitoring duration based on a preset risk area monitoring scheme;

[0092] Analyzing the monitoring information to obtain an analysis result, and releasing the monitoring of the risk area in a case where the analysis result matches a monitoring release condition.

[0093] A monitoring device for African swine fever, comprising:

[0094] A parameter acquisition module acquires physiological parameters, environmental parameters and flow information parameters of infected pigs based on RFID tags, infrared sensors and environmental monitoring equipment;

[0095] A matrix establishment module establishes a parameter matrix of each parameter according to a time sequence, analyzes the parameter matrix using an analytic hierarchy process, establishes a judgment matrix, determines the weights of each parameter, and selects parameters that have a significant impact on the spread of ASF to form an influence matrix;

[0096] A model optimization module constructs a search space based on the influence matrix, optimizes the adaptive coefficients of the ELM network by an ENAS search algorithm with the search space as a constraint, and obtains an adaptive ELM network model;

[0097] A vector fusion module calculates the weights of each parameter using an entropy weight method, and fuses each parameter into a feature vector;

[0098] The monitoring module inputs the feature vector into the adaptive ELM network model to obtain a risk evaluation value, divides a risk region in combination with a risk threshold value, and performs corresponding monitoring based on a preset risk region monitoring scheme.

[0099] A computer device comprises a processor and a memory, and the processor is configured to execute an African swine fever monitoring and early warning program stored in the memory.

[0100] A storage medium stores one or more programs, and the one or more programs can be executed by one or more processors.

[0101] Parameters of temperature deviating from seasonal mean by more than ±5℃, humidity >90% or <30%, and NH3 concentration >30ppm are determined as the third environmental parameters;

[0102] The various embodiments are described in a progressive manner in the specification, and each embodiment focuses on the difference from other embodiments, and the same or similar parts between the various embodiments can be referred to each other. For the device disclosed by the embodiments, since it corresponds to the method disclosed by the embodiments, the description is relatively simple, and the related parts can be referred to the method part.

[0103] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to the embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for monitoring African swine fever, characterized in that: include: Based on RFID tags, infrared sensors and environmental monitoring equipment, the physiological parameters, environmental parameters and flow information parameters of infected pigs are collected; Establish a parameter matrix for each parameter according to the time series, analyze the parameter matrix using the analytic hierarchy process, establish a decision matrix, determine the weight of each parameter, and select the parameters that have a significant impact on ASF propagation to form an impact matrix; Constructing a search space based on the influence matrix, optimizing the adaptive coefficient of the ELM network through the ENAS search algorithm with the search space as a constraint, and obtaining an adaptive ELM network model; The entropy weight method is used to calculate the weight of each parameter and fuse each parameter into a feature vector; The characteristic vector is input into the adaptive ELM network model to obtain a risk assessment value, and the risk area is divided in combination with the risk threshold, and corresponding monitoring is performed based on a preset risk area monitoring plan.

2. The method for monitoring African swine fever according to claim 1, wherein: The physiological parameters include the first physiological parameters of the infected pigs in the acute phase and the second physiological parameters of the incubation phase; the environmental parameters include the first environmental parameters of the infected pigs' activity areas and the second environmental parameters of the inactive areas in different seasons; the flow information parameters include the first flow parameters of the infected pigs in the acute phase and the second flow parameters of the incubation phase; the first physiological parameters and the second physiological parameters both include body temperature, food intake, and activity level; the first environmental parameters and the second environmental parameters both include temperature, humidity, and air quality; the first flow parameters and the second flow parameters both include personnel flow records and material flow records.

3. The method for monitoring African swine fever according to claim 2, wherein: Also includes: The time series mean square error of the first physiological parameter of infected pigs in the acute phase was calculated using the following formula: Where m is the number of sampling times, is the mean value of the acute phase parameter; the mean square error threshold is set to kσ, and the parameter exceeding the mean square error threshold is determined as the third physiological parameter; Performing seasonal adaptability analysis on the first environmental parameter, calculating the fluctuation range of the parameter from the seasonal normal value, and determining the parameter that deviates from the seasonal mean as the third environmental parameter; The personnel flow records and material flow records entering the infected area without disinfection ≥2 times / day are determined as the third flow parameter.

4. The method for monitoring African swine fever according to claim 1, wherein: The establishment of the decision matrix specifically includes: Establish a parameter matrix of the same type of parameters according to the time series, and the matrix elements are the parameter values ​​at different times; Construct the decision matrix A = [a ij ] n×n ,satisfy: Among them, the element in the decision matrix is ​​a ij , a ij Indicates the importance of the i-th parameter relative to the j-th parameter, n is the total number of parameters, 1≤i≤n, 1≤j≤n; The consistency check coefficient of the decision matrix is ​​judged, and the decision matrix is ​​adjusted according to the judgment result until the consistency check coefficient of the decision matrix is ​​less than a preset threshold; the top N parameters in the decision matrix are formed into an influence matrix.

5. The method for monitoring African swine fever according to claim 3, wherein: The constructing of the search space based on the influence matrix specifically includes: using the first physiological parameter, the first environmental parameter, the second physiological parameter, and the second environmental parameter as first search parameters, and using the third physiological parameter, the third environmental parameter, and the third flow information parameter as second search parameters; Taking the first element in the influence matrix of the first search parameter of the same type as the origin, calculating the distance between other elements and the origin as the radius to construct a first search circle, and forming the coverage range of the first search circle into a first search space; The maximum element in the influence matrix of the same type of second search parameter is used as the edge point of the second search circle, the distance between other elements and the origin is calculated as the radius to construct the second search circle, and the coverage range of all the second search circles constitutes the second search space; Among them, the influence matrices of the first physiological parameter, the second physiological parameter and the third physiological parameter correspond to the first physiological search space, the second physiological search space and the third physiological search space; the influence matrices of the first environmental parameter, the second environmental parameter and the third environmental parameter correspond to the first environmental search space, the second environmental search space and the third environmental search space; the influence matrix of the third flow information parameter corresponds to the third flow search space.

6. The method for monitoring African swine fever according to claim 5, characterized in that: The obtaining of the adaptive ELM network model specifically includes: Obtain the same number of parameter values ​​as the number of ELM network channels P from each search space, and calculate the difference ΔD between the parameter values ​​of the first and second physiological search spaces, the first and second environmental search spaces, and the parameter values ​​of the third physiological search space, the third environmental search space, and the third flow search space. The formula is: where v phys,i and v env,i are the first and second search space parameter values, v phy3,i 、v env3,i 、v flow3,i is the parameter value of the third search space; Adjust the coefficient λ with a step size of 0.2 and substitute it into the adaptive Simpson formula function to obtain the adaptive value; f(ΔD,λ)=λ·ΔD+(1-λ)·base; Among them, λ0, λ1, and λ2 are the coefficient values ​​of continuous steps, f(ΔD,λ) is the mapping function based on the difference value and the coefficient, and base is the benchmark value of historical epidemic data; The adaptive value is used as the adaptive coefficient of the trained ELM network.

7. The method for monitoring African swine fever according to claim 1, wherein: Also includes: Acquire monitoring information obtained by monitoring the risk area within a monitoring period based on a preset risk area monitoring plan; The monitoring information is analyzed to obtain an analysis result, and when the analysis result matches a monitoring release condition, monitoring of the risk area is released.

8. A monitoring device for African swine fever, characterized in that: include: Parameter acquisition module, which collects physiological parameters, environmental parameters and flow information parameters of infected pigs based on RFID tags, infrared sensors and environmental monitoring equipment; The matrix building module builds the parameter matrix of each parameter according to the time series, uses the hierarchical analysis method to analyze the parameter matrix, builds the decision matrix, determines the weight of each parameter, and selects the parameters that have a significant impact on ASF propagation to form an impact matrix; A model optimization module constructs a search space based on the influence matrix, and optimizes the adaptive coefficient of the ELM network through the ENAS search algorithm with the search space as a constraint to obtain an adaptive ELM network model; The vector fusion module uses the entropy weight method to calculate the weight of each parameter and fuses each parameter into a feature vector; The monitoring module inputs the feature vector into the adaptive ELM network model to obtain a risk assessment value, divides the risk area into risk areas based on the risk threshold, and performs corresponding monitoring based on a preset risk area monitoring plan.

9. A computer device, characterized in that: include: A processor and a memory, wherein the processor is used to execute an African swine fever monitoring and early warning program stored in the memory to implement an African swine fever monitoring method as described in any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the African swine fever monitoring method described in any one of claims 1 to 7.

Images