Intelligent cattle herd health analysis and prevention and control system

By constructing a behavior network structure chart and health risk assessment, the problem of insufficient response ability of the Niuqun health monitoring system to identify complex behavior patterns and emergencies in the existing technology is solved, and real-time monitoring and disease prediction of Niuqun health management is realized. Improved accuracy.

CN118969274AInactive Publication Date: 2024-11-15第八师石河子市畜牧水产发展服务中心

Patent Information

Application Number
CN202411028474.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-30
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The existing intelligent analysis and prevention and control system of cattle herd health has limitations in deep data integration and advanced analysis, and lacks the ability to identify complex behavior patterns and instant response to sudden health problems, resulting in inefficient identification of early diseases and limited prevention capabilities. .

Method used

Using behavioral data collection module, behavioral network conversion module, behavioral abnormality detection module and health risk prediction module, the sensor collects the movement data of herd location, builds a behavioral network structure chart, detects behavioral abnormalities, combines health records and environmental data to predict health risks , form health risk assessment results.

Benefits of technology

It significantly improves the real-time monitoring capabilities of cattle herd health management, can identify potential health problems early, improve the accuracy of disease prediction and forward-looking decision-making, reduce economic losses, and enhance the effectiveness of ranch management.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118969274A_ABST
    Figure CN118969274A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of intelligent monitoring, in particular to an intelligent cattle health analysis, prevention and control system, which comprises a behavior data acquisition module, a data processing module, a data processing module, a data processing module, a data processing module and a data processing module, wherein the behavior data acquisition module is based on sensors at a plurality of positions in a pasture and is used for collecting position movement data of cattle, recording time information, linking the data with time labels and storing the data in a time sequence format; and obtaining behavior time sequence data. According to the method, the real-time monitoring capability of health management of the cattle herds is improved through a comprehensive data acquisition and analysis technology, a behavior network can be constructed and real-time anomaly detection can be carried out by adopting a data processing and analysis technology, and potential health problems are effectively mined from complex data through continuous analysis of behavior data, so that the health management efficiency of the cattle herds is improved. Early recognition is carried out on subtle changes of behavior modes, the accuracy of disease prediction and the perspectiveness of decision making are remarkably improved, errors of manual judgment are reduced through automatic data analysis, economic losses are effectively reduced, and the overall efficiency of pasture management is enhanced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of intelligent monitoring technology, and in particular to an intelligent analysis and prevention and control system for cattle health. Background Art

[0002] The field of intelligent monitoring technology covers the use of advanced data processing technology, artificial intelligence algorithms and sensor technology to achieve automated monitoring and management of the environment, objects or individual behaviors. This technology is often used in a variety of scenarios such as security monitoring, health management, and production quality control. By integrating video analysis, biometrics, geographic information systems (GIS) and Internet of Things (IoT) technologies, intelligent monitoring systems can collect and analyze data in real time, thereby improving response speed and decision-making efficiency. In the agricultural field, intelligent monitoring also involves the use of sensors to monitor the growing conditions of crops and livestock to optimize resource allocation and prevent the occurrence of diseases.

[0003] Among them, the intelligent analysis and prevention and control system for cattle health refers to the use of intelligent monitoring technology to continuously track and evaluate the health status of cattle. This system usually includes sensors, cameras and other biological monitoring tools that can collect physiological and behavioral data of cattle in real time. These data are then analyzed to detect early signs of disease, evaluate the quality of the feeding environment, or monitor the reproductive health of cattle. The main purpose is to improve pasture management efficiency, reduce the incidence of disease, ensure the quality of beef or dairy products, and thus improve the sustainability and economic benefits of agricultural production.

[0004] Existing technologies show significant limitations in deep data integration and advanced analysis, especially in agricultural livestock health monitoring. Existing systems often rely on basic data collection and simple monitoring techniques, lacking the ability to identify complex behavioral patterns and respond immediately to sudden health issues. This technical deficiency leads to inefficiency in identifying subtle changes in behavioral patterns, such as stress signals or early signs of disease. The lack of advanced algorithms and data integration capabilities limits the ability of traditional monitoring systems to prevent diseases, and failure to identify health risks in a timely manner may lead to the rapid development of diseases and increase the economic burden on ranches. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and propose an intelligent cattle health analysis and prevention and control system.

[0006] In order to achieve the above purpose, the present invention adopts the following technical solutions: The cattle health intelligent analysis and prevention and control system includes:

[0007] The behavior data collection module collects the location movement data of the cattle based on sensors at multiple locations in the ranch, records the time information, links the data with the time tag, stores it in a time series format, and obtains the behavior time series data;

[0008] The behavior network conversion module extracts the target time point from the behavior time series data, converts the behavior pattern of the target time point into a network node, connects the nodes according to the fit of the behavior, constructs a behavior association graph, and establishes a behavior network structure graph through network structure analysis;

[0009] The behavior anomaly detection module uses the node and link data in the behavior network structure diagram to detect mutations and irregular distribution in the behavior pattern, sets a dynamic threshold of the entropy value based on historical data, compares it with the real-time network entropy value, mines abnormal changes in behavior, and identifies them as behavior anomaly signals;

[0010] The health risk prediction module combines the abnormal behavior signal, calls the health records and environmental data of the cattle, predicts the health risk, calculates the health risk level according to the abnormal behavior frequency and severity, and forms a health risk assessment result.

[0011] As a further solution of the present invention, the steps of acquiring the behavior time series data are specifically as follows:

[0012] Based on the sensor array in the pasture, the position coordinates (x t ,y t ) and the corresponding timestamp t, using the formula:

[0013]

[0014] Calculate the distance change between consecutive time points to obtain the position change distance data of each cow;

[0015] Among them, x t ,y t represents the position coordinate at time t, x t+1 ,y t+1 represents the position coordinate at time t+1, Δd represents the distance of position change;

[0016] The position change distance data is associated with the timestamp using the formula:

[0017]

[0018] Calculate the moving speed and obtain the speed time series;

[0019] Among them, V represents the speed, Δt represents the time interval, k is the adjustment coefficient, and ∈ is a small quantity to avoid the denominator being zero;

[0020] The speed time series is optimized using exponential smoothing, using the formula:

[0021] S t =αV t +(1-α)St-1 +βcos(γt)

[0022] Adjust the periodic parameters, calculate the smoothed velocity time series, and generate behavior time series data;

[0023] Among them, S t represents the smoothed velocity time series, α is the smoothing coefficient, V t is the original speed value, and β and γ adjust the periodic changes.

[0024] As a further solution of the present invention, the step of obtaining the behavior association graph is specifically as follows:

[0025] Extract the data of the target time point from the behavior time series data, and use the data conversion function of the time point, using the formula:

[0026] N t =f(S t ,θ)

[0027] Convert to network nodes and generate a network node set;

[0028] Where f is the conversion function, N t is the set of network nodes after transformation, θ is the adjustment parameter;

[0029] Analyze each set of network nodes and decompose the node features into N ik and N jk , calculate the degree of fit between each pair of nodes, using the formula:

[0030]

[0031] Generate the fitness F between nodes ij , get the node connection data;

[0032] Among them, N ik 、N jk Respectively represent the values ​​of node i and node j on feature k, F ij represents the fit between nodes i and j, σ is the smoothing parameter, and n represents the total number of reference features;

[0033] Use the network structural analysis technique according to the node connection data, using the formula:

[0034]

[0035] Construct behavioral association graphs;

[0036] Among them, G represents the behavior association graph, a ij represents the connection weight, F ijis the degree of fit between nodes, and λ is the degree of fit magnification factor.

[0037] As a further solution of the present invention, the steps of obtaining the behavior network structure diagram are specifically as follows:

[0038] From the behavioral association matrix, the degree centrality of each node is calculated using the formula:

[0039]

[0040] Generate degree centrality scores;

[0041] Among them, G is composed of the relationship matrix A=[a ij ] definition, C d (i) represents the degree centrality of node i, a ij represents the connection weight between nodes, w ij the importance of multilateral adjustments;

[0042] Analyze the clustering characteristics between nodes using the formula:

[0043]

[0044] Generate the clustering coefficient of each node;

[0045] Among them, C c (i) represents the clustering coefficient of node i, T(i) represents the number of triangles involved by node i, and k i represents the number of connections of node i, ∈ represents the smoothing parameter;

[0046] Combining the degree centrality score and the clustering coefficient, the formula is adopted:

[0047]

[0048] Generate the score S(i) of each node and construct the behavior network structure graph H;

[0049] Among them, S(i) represents the score of node i, and α and β represent the coefficients that adjust the influence of centrality and clustering coefficient.

[0050] As a further solution of the present invention, the step of acquiring the abnormal behavior signal is specifically:

[0051] Node and link data are extracted from the behavior network structure diagram, and the real-time network entropy value is calculated using the node and link data using the formula:

[0052]

[0053] Calculate and generate real-time network entropy value;

[0054] Among them, Et represents the entropy value of the real-time network, p ij represents the transition probability from node i to node j, q ij represents the background or expected transition probability;

[0055] The dynamic threshold of entropy value is set according to historical data, using the formula:

[0056]

[0057] Calculate and generate dynamic entropy threshold;

[0058] Among them, T represents the dynamic threshold of entropy value, μ E represents the mean of historical entropy values, σ E Represents standard deviation, skew E represents skewness, Kurt E represents kurtosis, k represents adjustment coefficient, δ represents adjustment coefficient, and γ represents adjustment coefficient;

[0059] Compare the real-time network entropy value with the dynamic threshold T, using the formula:

[0060]

[0061] If R is greater than the preset target value, it is marked as an abnormal behavior signal;

[0062] Among them, R represents the difference between the entropy value and the threshold.

[0063] As a further solution of the present invention, the steps for obtaining the health risk assessment results are specifically as follows:

[0064] Combined with the abnormal behavior signal, the health record data of the herd is called and the formula is used:

[0065]

[0066] Calculate and generate preliminary health risk factors;

[0067] Among them, F represents the preliminary health risk factor, α i represents the adjustment factor, R i Represents abnormal behavior signal, H i represents the health record of the herd, θ and φ represent nonlinear adjustment indices;

[0068] Based on the preliminary health risk factors, combined with environmental data, the formula is used:

[0069] D=β·F+γ·E 0.5 +δ·log(1+E)

[0070] Calculate and generate adjusted health risk factors;

[0071] Among them, D represents the adjusted health risk factor, β represents the weight coefficient, γ represents the weight coefficient, δ represents the weight coefficient, and E represents the environmental data;

[0072] Based on the adjusted health risk factors, according to the abnormal frequency and severity of the behavior, the formula is used:

[0073] L=ξ·D×(1+freq ψ ×sev ω )

[0074] Calculate the health risk level and form the health risk assessment results;

[0075] Among them, L represents the health risk assessment result, ξ represents the adjustment coefficient, ψ represents the adjustment coefficient, ω represents the adjustment coefficient, freq represents the abnormal frequency, and sev represents the abnormal severity.

[0076] Compared with the prior art, the advantages and positive effects of the present invention are:

[0077] In the present invention, the real-time monitoring capability of cattle health management is improved through comprehensive data collection and analysis technology. The use of data processing and analysis technology can build a behavioral network and perform real-time anomaly detection. Through continuous analysis of behavioral data, potential health problems can be effectively mined from complex data. And through early identification of subtle changes in behavioral patterns and in-depth comparison with historical health records and environmental data, the accuracy of disease prediction and the foresight of decision-making are significantly improved. Automated data analysis reduces the error of manual judgment, significantly improves decision-making efficiency, effectively reduces economic losses and enhances the overall effectiveness of pasture management. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Figure 1 is a system flow chart of the present invention;

[0079] Figure 2 A flowchart of steps for obtaining behavior time series data of the present invention;

[0080] Figure 3 A flowchart of steps for obtaining a behavior association diagram of the present invention;

[0081] Figure 4 A flowchart of steps for obtaining a behavior network structure diagram of the present invention;

[0082] Figure 5 A flowchart of the steps for obtaining abnormal behavior signals of the present invention;

[0083] Figure 6 The figure is a flow chart of the steps for obtaining the health risk assessment results of the present invention. DETAILED DESCRIPTION

[0084] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction 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 intended to limit the present invention.

[0085] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0086] Embodiment 1

[0087] See also Figure 1 , the cattle health intelligent analysis and prevention and control system includes:

[0088] The behavior data collection module collects the location movement data of the cattle based on sensors at multiple locations in the ranch, records the time information, links the data with the time tag, stores it in a time series format, and obtains the behavior time series data;

[0089] The behavior network conversion module extracts the target time point from the behavior time series data, converts the behavior pattern of the target time point into a network node, connects the nodes according to the fit of the behavior, constructs a behavior association graph, and establishes a behavior network structure graph through network structure analysis;

[0090] The behavior anomaly detection module uses the node and link data in the behavior network structure diagram to detect mutations and irregular distribution in the behavior pattern. It sets a dynamic threshold of the entropy value based on historical data, compares it with the real-time network entropy value, mines abnormal changes in behavior, and identifies them as behavior anomaly signals.

[0091] The health risk prediction module combines behavioral abnormality signals, calls the health records and environmental data of the herd, predicts health risks, calculates the health risk level based on the abnormal frequency and severity of the behavior, and forms a health risk assessment result.

[0092] Behavioral time series data specifically include location movement data, time information, and time series format. The behavioral network structure diagram includes network nodes, node connections, and behavioral association diagram. Behavioral abnormal signals specifically refer to mutation signals, irregular distribution signals, and entropy value comparison results. Health risk assessment results include health record calls, environmental data analysis, and health risk levels.

[0093] See also Figure 2 ,The specific steps for obtaining behavioral time series data are:

[0094] Based on the sensor array in the pasture, the position coordinates (x t ,y t ) and the corresponding timestamp t, using the formula:

[0095]

[0096] Calculate the distance change between consecutive time points to obtain the position change distance data of each cow;

[0097] Among them, x t ,y t represents the position coordinate at time t, x t+1 ,y t+1 represents the position coordinate at time t+1, Δd represents the distance of position change;

[0098] Associate the location change distance data with the timestamp using the formula:

[0099]

[0100] Calculate the moving speed and obtain the speed time series;

[0101] Among them, V represents the speed, Δt represents the time interval, k is the adjustment coefficient, and ∈ is a small quantity to avoid the denominator being zero;

[0102] The speed time series is optimized using exponential smoothing, using the formula:

[0103] S t =αV t +(1-α)S t-1 +βcos(γt)

[0104] Adjust the periodic parameters, calculate the smoothed velocity time series, and generate behavior time series data;

[0105] Among them, S t represents the smoothed velocity time series, α is the smoothing coefficient, V t is the original speed value, and β and γ adjust the periodic changes.

[0106] formula:

[0107]

[0108] Parameter details:

[0109] x t ,y t : Position coordinates at time t.

[0110] x t+1 ,y t+1 : Position coordinates at time t+1.

[0111] Δd: The change in distance between time t and t+1.

[0112] Example derivation:

[0113] Suppose at time t, a cow is located at (x t ,y t )=(2,3). At time t+1, the cow moves to (x t+1 ,y t+1 )=(5,7). According to the formula, calculate the distance change:

[0114]

[0115] Result description:

[0116] The calculated distance change Δd=5 meters means that the cow moved 5 meters from time t to t+1.

[0117] formula:

[0118]

[0119] Parameter details:

[0120] V: Movement speed.

[0121] Δd: position change distance.

[0122] Δt: time interval.

[0123] k: adjustment factor (reference environmental factors including wind speed), assuming k = 1 for simplification.

[0124] ∈: small amount, to avoid the denominator being zero, take ∈=0.01.

[0125] Example derivation:

[0126] If we assume that the time interval Δt = 2 seconds, then the speed is calculated as follows:

[0127]

[0128] Result description:

[0129] The calculated speed V≈2.49 m / s means that the cow moves at approximately 2.49 m / s.

[0130] formula:

[0131] S t =αV t +(1-α)S t-1 +βcos(γt)

[0132] Parameter details:

[0133] S t : Smoothed velocity time series.

[0134] α: Smoothing coefficient, take α=0.5.

[0135] V t : Original speed value.

[0136] S t-1 : The velocity sequence value at the previous time point, assumed to be S t-1 =2m / s.

[0137] β, γ: Parameters for adjusting periodic changes, assuming β = 0.1, γ = 2π.

[0138] Example derivation:

[0139] If the current time t = 1 second, V t =2.49m / s, then calculate the smoothed speed:

[0140] S t =0.5·2.49+0.5·2+0.1·cos(2π·1)=1.245+1-0.1=

[0141] 2.145m / s

[0142] Result description:

[0143] The calculated smoothing speed S t =2.145m / s means that after smoothing, the speed of the cow is about 2.145 meters / second.

[0144] See also Figure 3 , the steps to obtain the behavior association graph are as follows:

[0145] Extract the data of the target time point from the behavioral time series data, use the data conversion function of the time point, and use the formula:

[0146] N t =f(S t ,θ)

[0147] Convert to network nodes and generate a network node set;

[0148] Where f is the conversion function, N t is the set of network nodes after transformation, θ is the adjustment parameter;

[0149] Analyze each set of network nodes and decompose the node features into N ik and N jk , calculate the degree of fit between each pair of nodes, using the formula:

[0150]

[0151] Generate the fitness F between nodes ij , get the node connection data;

[0152] Among them, N ik 、N jk Respectively represent the values ​​of node i and node j on feature k, F ij represents the fit between nodes i and j, σ is the smoothing parameter, and n represents the total number of reference features;

[0153] Use the network's structural analysis technique based on the node connection data, using the formula:

[0154]

[0155] Construct behavioral association graphs;

[0156] Among them, G represents the behavior association graph, a ij represents the connection weight, F ij is the degree of fit between nodes, and λ is the degree of fit magnification factor.

[0157] formula:

[0158] N t =f(S t ,θ)

[0159] in:

[0160] S t : Behavioral data in time series, assumed to be continuous signal strength.

[0161] θ: Adjustment parameter of data conversion, used to correct or adjust data.

[0162] f: conversion function, transforming the behavior data S t Convert to network node N t .

[0163] Derivation process:

[0164] Assume that at the target time t, the behavior data S t The value of is 200. The adjustment parameter θ is set to 1.5, which represents the multiplication factor of the signal strength. The conversion function f can be set to f(S t ,θ)=θ·S t .

[0165] Calculation example:

[0166] N t =1.5×200=300

[0167] Result analysis:

[0168] The calculated N t A value of 300 represents the node signal strength after multiplication adjustment.

[0169] formula:

[0170]

[0171] in:

[0172] N ik 、N jk : The value of nodes i and j on feature k.

[0173] σ: Smoothing parameter, used to control sensitivity.

[0174] n: The total number of features referenced.

[0175] Derivation process:

[0176] Assume there are two nodes with eigenvalues ​​N i1 =300 and N j1 =270, number of features n=1, smoothing parameter σ=10.

[0177] Calculation example:

[0178]

[0179] Result analysis:

[0180] The obtained F ij The value is 0.75, which means that the fit between nodes i and j in terms of features is 75%. A higher fit indicates that the two nodes are similar in terms of features.

[0181] formula:

[0182]

[0183] in:

[0184] G: Behavioral association diagram.

[0185] aij : Weight coefficient, indicating whether nodes i and j are connected.

[0186] F ij : The degree of fit between nodes.

[0187] λ: Fitting magnification factor.

[0188] Derivation process:

[0189] Assume there are two nodes, a ij =1 means the two nodes are connected, and the fitting degree F ij Calculated as 0.75, λ = 2

[0190] Calculation example:

[0191] G=1×(1-exp(-2×0.75))=1-exp(-1.5)≈0.77687

[0192] Result analysis:

[0193] The resulting G value is approximately 0.77687, which represents the strength of the connection between these two nodes in the network, with higher values ​​indicating stronger associations.

[0194] See also Figure 4 , the steps to obtain the behavior network structure diagram are as follows:

[0195] From the behavior association matrix, calculate the degree centrality of each node using the formula:

[0196]

[0197] Generate degree centrality scores;

[0198] Among them, G is composed of the relationship matrix A=[a ij ] definition, C d (i) represents the degree centrality of node i, a ij represents the connection weight between nodes, w ij the importance of multilateral adjustments;

[0199] Analyze the clustering characteristics between nodes using the formula:

[0200]

[0201] Generate the clustering coefficient of each node;

[0202] Among them, C c (i) represents the clustering coefficient of node i, T(i) represents the number of triangles involved by node i, and k i represents the number of connections of node i, ∈ represents the smoothing parameter;

[0203] Combining the degree centrality score and the clustering coefficient, the formula is used:

[0204]

[0205] Generate the score S(i) of each node and construct the behavior network structure graph H;

[0206] Among them, S(i) represents the score of node i, and α and β represent the coefficients that adjust the influence of centrality and clustering coefficient.

[0207] Assumed network conditions

[0208] Number of nodes m = 3

[0209] Connection weight a ij :

[0210] a 12 =1, a 13 =0.5, a 23 =1

[0211] The importance of edge w ij Assume that each side is equal and equal to 1

[0212] Degree centrality C d The formula for (i) is:

[0213]

[0214] Take computing the degree centrality of node 1 as an example.

[0215] Node 1:

[0216] C d (1) = a 12 × 12 +a 13 × 13 =1×1+0.5×1=1.5

[0217] This means that the total connection strength between node 1 and other nodes in the network is 1.5, reflecting the level of activity in the network.

[0218] Clustering coefficient C c The formula for (i) is:

[0219]

[0220] Assume ∈=0.01 to avoid division by zero, and calculate the clustering coefficient of node 1.

[0221] Assume that node 1 forms a triangle through nodes 2 and 3:

[0222] T(1) = 1 (node ​​1 participates in a triangle)

[0223] k 1 =2 (node ​​1 has two connections)

[0224] Node 1:

[0225]

[0226] The clustering coefficient is close to 1, which means that node 1 and its neighbors form a tight clustering structure.

[0227] The formula for the comprehensive score is:

[0228]

[0229] α=0.6 and β=0.4 will be used.

[0230] Node 1:

[0231]

[0232] The overall score of 1.299 reflects the importance of node 1 in the network, referring to the connection strength and local structural tightness.

[0233] See also Figure 5 , the specific steps for obtaining abnormal behavior signals are:

[0234] Extract node and link data from the behavior network structure diagram, and use the node and link data to calculate the real-time network entropy value using the formula:

[0235]

[0236] Calculate and generate real-time network entropy value;

[0237] Among them, E t represents the entropy value of the real-time network, p ij represents the transition probability from node i to node j, q ij represents the background or expected transition probability;

[0238] The dynamic threshold of entropy value is set according to historical data, using the formula:

[0239] T=μ E +k·σ E +δ·skew E +γ·kurt E

[0240] Calculate and generate dynamic entropy threshold;

[0241] Among them, T represents the dynamic threshold of entropy value, μ E represents the mean of historical entropy values, σ ERepresents standard deviation, skew E represents skewness, Kurt E represents kurtosis, k represents adjustment coefficient, δ represents adjustment coefficient, and γ represents adjustment coefficient;

[0242] Compare the real-time network entropy value with the dynamic threshold T, using the formula:

[0243]

[0244] If R is greater than the preset target value, it is marked as an abnormal behavior signal;

[0245] Among them, R represents the difference between the entropy value and the threshold.

[0246] formula:

[0247]

[0248] Parameter explanation and examples:

[0249] p ij : The probability of transition from node i to node j. For example, if there are 3 transitions from node i to node j and the total number of transitions is 100, then

[0250] q ij : Background or expected transition probability. Assume that the expected uniform probability is For n = 10 nodes,

[0251] Calculation process:

[0252] Assume there is a network where n=2 (two nodes) and the number of transfers is distributed as follows:

[0253] p 11 =0.1, p 12 =0.4

[0254] p 21 =0.3, p 22 =0.2

[0255] Background probability q ij =0.25 (because n=2)

[0256] Insert the formula to calculate:

[0257]

[0258] Using the natural logarithm to calculate E t .

[0259] Assuming the calculated E tis -0.05. This value indicates that the current network state is more irregular or chaotic than the expected state.

[0260] formula:

[0261] T=μ E +k·σ E +δ·skew E +γ·kurt E

[0262] Parameter explanation and examples:

[0263] μ E : The mean of historical entropy values, assumed to be 0.3.

[0264] σ E : The standard deviation of historical entropy values, assumed to be 0.05.

[0265] skew E : The skewness of the historical entropy value, assumed to be 0.1.

[0266] kurt E : The kurtosis of the historical entropy value, assumed to be 0.2.

[0267] k, δ, γ: adjustment coefficients, set to 1, 1, 1 respectively.

[0268] Calculation process:

[0269] Plug in the given values ​​to calculate T:

[0270] T=0.3+1×0.05+1×0.1+1×0.2=0.65T=0.65 means that if the real-time entropy E t If it is greater than this value, an exception occurs.

[0271] formula:

[0272]

[0273] Parameter explanation and examples:

[0274] E t =-0.5

[0275] T=0.65

[0276] Calculation process:

[0277] Insert the formula to calculate R:

[0278]

[0279] R=1.15 indicates that the entropy value difference exceeds a predetermined threshold value, which is identified as abnormal behavior.

[0280] See also Figure 6 , the specific steps for obtaining health risk assessment results are:

[0281] Combined with abnormal behavior signals, the health record data of the herd is called and the formula is used:

[0282]

[0283] Calculate and generate preliminary health risk factors;

[0284] Among them, F represents the preliminary health risk factor, α i represents the adjustment factor, R i Represents abnormal behavior signal, H i represents the health record of the herd, θ and φ represent nonlinear adjustment indices;

[0285] Based on preliminary health risk factors and combined with environmental data, the formula is used:

[0286] D=β·F+γ·E 0.5 +δ·log(1+E)

[0287] Calculate and generate adjusted health risk factors;

[0288] Among them, D represents the adjusted health risk factor, β represents the weight coefficient, γ represents the weight coefficient, δ represents the weight coefficient, and E represents the environmental data;

[0289] Based on the adjusted health risk factors, according to the abnormal frequency and severity of the behavior, the formula is used:

[0290] L=ξ·D×(1+freq ψ ×sev ω )

[0291] Calculate the health risk level and form the health risk assessment results;

[0292] Among them, L represents the health risk assessment result, ξ represents the adjustment coefficient, ψ represents the adjustment coefficient, ω represents the adjustment coefficient, freq represents the abnormal frequency, and sev represents the abnormal severity.

[0293] formula:

[0294]

[0295] parameter:

[0296] α i : The weight of each data point, assumed to be 1.0.

[0297] R i : Abnormal behavior signal, assumed value is 2.

[0298] H i : Health record data of the herd, assumed value is 3.

[0299] θ, φ: nonlinear adjustment exponents, assumed to be 2 and 2 respectively.

[0300] Derivation process:

[0301] Assume there are three data points, ie, n=3.

[0302] F=1.0×2 2 ×3 2 +1.0×2 2 ×3 2 +1.0×2 2 ×3 2 =3×4×9=108

[0303] Therefore, the value of F is 108, which means that in this setting, the sum of the health risk factors is 108, implying a high accumulation of health risks.

[0304] formula:

[0305] D=β·F+γ·E 0.5 +δ·log(1+E)

[0306] parameter:

[0307] β, Y, δ: weight coefficients, assumed to be 1.0.

[0308] F: 108.

[0309] E: Environmental data, assumed to be 4.

[0310] Derivation process:

[0311] D=1.0×108+1.0×4 0.5 +1.0×log(1+4)

[0312] D=108+2+log(5)≈1 10.70

[0313] The results represent adjusted health risk factors with reference to environmental influences, where environmental factors contribute to the total risk estimate via square root and logarithmic functions.

[0314] formula:

[0315] L=ξ·D×(1+freq ψ ×sev ω )

[0316] parameter:

[0317] ξ: adjustment coefficient, assumed to be 1.0.

[0318] D: 110.70.

[0319] freq: abnormal frequency, assumed to be 2.

[0320] sev: abnormality severity, assumed to be 3.

[0321] ψ, ω: adjustment exponents, assumed to be 1 and 1 respectively.

[0322] Derivation process:

[0323] L = 1.0 × 110.70 × (1 + 2 1 ×3 1 )

[0324] L=110.70×(1+6)=110.70×7=774.9

[0325] The results represent the health risk level, referring to the abnormal frequency and severity of the behavior. A value of 774.9 indicates a high health risk level, suggesting that action is needed.

[0326] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. Intelligent analysis and prevention system of cattle health, It is characterized in that The system comprises: The behavior data collection module collects the location movement data of the cattle based on sensors at multiple locations in the ranch, records the time information, links the data with the time tag, stores it in a time series format, and obtains the behavior time series data; The behavior network conversion module extracts the target time point from the behavior time series data, converts the behavior pattern of the target time point into a network node, connects the nodes according to the fit of the behavior, constructs a behavior association graph, and establishes a behavior network structure graph through network structure analysis; The behavior anomaly detection module uses the node and link data in the behavior network structure diagram to detect mutations and irregular distribution in the behavior pattern, sets a dynamic threshold of the entropy value based on historical data, compares it with the real-time network entropy value, mines abnormal changes in behavior, and identifies them as behavior anomaly signals; The health risk prediction module combines the abnormal behavior signal, calls the health records and environmental data of the cattle, predicts the health risk, calculates the health risk level according to the abnormal behavior frequency and severity, and forms a health risk assessment result.

2. The cattle health intelligent analysis and prevention and control system according to claim 1, It is characterized in that The steps for obtaining the behavior time series data are specifically as follows: Based on the sensor array in the pasture, the position coordinates (x t ,y t ) and the corresponding timestamp t, using the formula: Calculate the distance change between consecutive time points to obtain the position change distance data of each cow; Among them, x t ,y t represents the position coordinate at time t, x t+1 ,y t+1 represents the position coordinate at time t+1, Δd represents the distance of position change; The position change distance data is associated with the timestamp using the formula: Calculate the moving speed and obtain the speed time series; Among them, V represents the speed, Δt represents the time interval, k is the adjustment coefficient, and ∈ is a small quantity to avoid the denominator being zero; The speed time series is optimized using exponential smoothing, using the formula: S t =αV t +(1-α)S t-1 +βcos(γt) Adjust the periodic parameters, calculate the smoothed velocity time series, and generate behavior time series data; Among them, S t represents the smoothed velocity time series, α is the smoothing coefficient, V t is the original speed value, and β and γ adjust the periodic changes.

3. The cattle health intelligent analysis and prevention and control system according to claim 2, It is characterized in that The steps for obtaining the behavior association graph are specifically as follows: Extract the data of the target time point from the behavior time series data, and use the data conversion function of the time point, using the formula: N t =f(S t ,i) Convert to network nodes and generate a network node set; Where f is the conversion function, N t is the set of network nodes after transformation, θ is the adjustment parameter; Analyze each set of network nodes and decompose the node features into N ik and N jk , calculate the degree of fit between each pair of nodes, using the formula: Generate the fitness F between nodes ij , get the node connection data; Among them, N ik 、N jk Respectively represent the values ​​of node i and node j on feature k, F ij represents the fit between nodes i and j, σ is the smoothing parameter, and n represents the total number of reference features; Use the network structural analysis technique according to the node connection data, using the formula: Construct behavioral association graphs; Among them, G represents the behavior association graph, a ij represents the connection weight, F ij is the degree of fit between nodes, and λ is the degree of fit magnification factor.

4. The cattle health intelligent analysis and prevention and control system according to claim 3, It is characterized in that The steps for obtaining the behavior network structure diagram are specifically as follows: From the behavioral association matrix, the degree centrality of each node is calculated using the formula: Generate degree centrality scores; Among them, G is composed of the relationship matrix A=[a ij ] definition, C d (i) represents the degree centrality of node i, a ij represents the connection weight between nodes, w ij the importance of multilateral adjustments; Analyze the clustering characteristics between nodes using the formula: Generate the clustering coefficient of each node; Among them, C c (i) represents the clustering coefficient of node i, T(i) represents the number of triangles involved by node i, and k i represents the number of connections of node i, ∈ represents the smoothing parameter; Combining the degree centrality score and the clustering coefficient, the formula is adopted: Generate the score S(i) of each node and construct the behavior network structure graph H; Among them, S(i) represents the score of node i, and α and β represent the coefficients that adjust the influence of centrality and clustering coefficient.

5. The cattle health intelligent analysis and prevention and control system according to claim 4, It is characterized in that The steps of obtaining the abnormal behavior signal are specifically as follows: Node and link data are extracted from the behavior network structure diagram, and the real-time network entropy value is calculated using the node and link data using the formula: Calculate and generate real-time network entropy value; Among them, E t represents the entropy value of the real-time network, p ij represents the transition probability from node i to node j, q ij represents the background or expected transition probability; The dynamic threshold of entropy value is set according to historical data, using the formula: T=μ E +k·s E +δ·skew E +g·kurt E Calculate and generate dynamic entropy threshold; Among them, T represents the dynamic threshold of entropy value, μ E represents the mean of historical entropy values, σ E Represents standard deviation, skew E represents skewness, Kurt E represents kurtosis, k represents adjustment coefficient, δ represents adjustment coefficient, and γ represents adjustment coefficient; Compare the real-time network entropy value with the dynamic threshold T, using the formula: If R is greater than the preset target value, it is marked as an abnormal behavior signal; Among them, R represents the difference between the entropy value and the threshold.

6. The cattle health intelligent analysis and prevention and control system according to claim 5, It is characterized in that The steps for obtaining the health risk assessment results are specifically as follows: Combined with the abnormal behavior signal, the health record data of the herd is called and the formula is used: Calculate and generate preliminary health risk factors; Among them, F represents the preliminary health risk factor, α i represents the adjustment factor, R i Represents abnormal behavior signal, H i represents the health record of the herd, θ and φ represent nonlinear adjustment indices; Based on the preliminary health risk factors and combined with environmental data, the formula is used: D=β·F+γ·E 0.5 +δ·log(1+E) Calculate and generate adjusted health risk factors; Among them, D represents the adjusted health risk factor, β represents the weight coefficient, γ represents the weight coefficient, δ represents the weight coefficient, and E represents the environmental data; Based on the adjusted health risk factors, according to the abnormal frequency and severity of the behavior, the formula is used: L=ξ·D×(1+freq ψ ×sev ω ) Calculate the health risk level and form the health risk assessment results; Among them, L represents the health risk assessment result, ξ represents the adjustment coefficient, ψ represents the adjustment coefficient, ω represents the adjustment coefficient, freq represents the abnormal frequency, and sev represents the abnormal severity.

Citation Information

Patent Citations

  • Health index evaluating method and system for achieving animal behavior monitoring

    CN104077550A

  • Method, device and terminal for analyzing network relationship between individual movement and ecological space

    CN115082250A

  • Cattle sales digital management system and method based on Internet of Things

    CN117437079A

  • Sheep behavior identification method and system based on spatio-temporal information

    CN117496420A

  • Living cattle supervision and evaluation system

    CN117581809A

Cited By

  • Pig farm anomaly detection system based on multi-sensor data fusion

    CN120123958A

  • Pig farm anomaly detection system based on multi-sensor data fusion

    CN120123958B

  • AI-TV platform cross-end cooperative processing method based on multi-modal interaction

    CN121524956A