Poultry breeding monitoring management system based on big data

Abstract

The invention discloses a poultry breeding monitoring management system based on big data, and relates to the technical field of poultry breeding supervision, the poultry breeding monitoring management system monitors the environment in a poultry house in real time, immediately triggers a behavior monitoring module when abnormality is detected, generates a specific environmental abnormality score, and combines visual monitoring and infrared thermal imaging to realize the monitoring and management of poultry breeding. According to the method, individual behavior characteristics of each poultry are effectively extracted, an individual behavior data stream is generated, individual potential health risks are identified, a health archive is adopted to accumulate long-term behaviors and environmental data of each poultry, and a time sequence analysis model is utilized to identify individual behavior modes and health change trends. Possible health risks are predicted based on historical data and models, management perspectiveness and disease prevention and control efficiency are effectively improved, meanwhile, abnormal features with the most influence are automatically recognized through comprehensive risk scoring and contribution degree analysis and are pushed to management personnel, the health archive of each poultry is updated in real time, and the management efficiency is improved. And the overall operation efficiency of the farm is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of poultry breeding supervision, and in particular to a poultry breeding monitoring and management system based on big data. Background Art

[0002] Currently, the most common poultry farming monitoring method mainly collects data through environmental sensors and video monitoring, monitors the temperature and humidity, carbon dioxide concentration and activity status of poultry groups in the poultry house, and identifies environmental changes with the help of threshold alarm mechanisms; however, most of these environmental sensors and video monitoring systems operate independently and can only perform surface analysis of environmental and health information. They cannot achieve individual health monitoring and cannot effectively adapt to environmental changes. In large-scale farms, it is difficult to capture early abnormal behaviors or health risks of individual poultry in a timely manner.

[0003] In order to improve the accuracy of environmental monitoring, some breeding plans have added odor detection systems in poultry houses to monitor changes in the concentration of harmful gases such as ammonia, and use historical experience to associate some environmental variables with health conditions. In addition, some video surveillance systems have introduced behavior recognition models to monitor poultry group activities and assist in assessing overall health status. Such plans have certain effects in group health monitoring and abnormal warning, but because it is impossible to track the health of individuals, it is impossible to conduct a comprehensive analysis of environmental, odor and behavioral data, and it is difficult to trace the source of the problem. Therefore, when the gas level in the environment changes, it is difficult to determine the specific affected individuals, ignore the individual's early abnormal behavior, and it is difficult to meet the response speed and targeted needs of poultry health monitoring. Therefore, a poultry breeding monitoring and management system based on big data is urgently needed to solve such problems. Summary of the invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] The present invention provides a poultry farming monitoring and management system based on big data to solve the problem that in traditional poultry farming monitoring and management solutions, the monitoring of environmental data and behavioral data is relatively scattered, lacks linkage analysis, and is difficult to extract valuable health monitoring information from the overall data; at the same time, traditional monitoring solutions lack big data support.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] On the one hand, the present invention provides a poultry farming monitoring and management system based on big data, which comprises:

[0008] An environmental monitoring system is used to monitor the environmental data in the poultry house and conduct preliminary analysis on it. The environmental monitoring system includes a sensor network and an edge computing module, and uses a sensor network to obtain environmental data;

[0009] The behavior analysis module is triggered by environmental warnings and uses visual technology to monitor and preliminarily screen the behavior status of individual poultry and identify individuals with abnormal behavior. The behavior analysis module includes an intelligent visual monitoring unit and a behavior benchmark comparison unit.

[0010] The intelligent visual monitoring unit uses a camera to scan the poultry house and extract individual behavioral characteristics, including feeding frequency, movement status, and posture changes, to establish a behavioral data stream for each poultry. In the intelligent visual monitoring unit, the behavioral characteristics collected by the camera are combined with the individual body temperature monitored by the infrared thermal imaging sensor for joint analysis.

[0011] Health profile library, which generates, updates and stores individual health profiles of poultry based on behavioral characteristics and environmental data, and uses in-depth analysis of individual health profiles to identify potential health trends and risks;

[0012] The early warning response management module notifies the management personnel when a health risk is identified in the health record library for an individual poultry.

[0013] Furthermore, in the environmental monitoring system, the collected environmental data include the air quality, temperature and humidity, and individual body temperature in the poultry house;

[0014] The sensor network is composed of temperature and humidity, carbon dioxide, ammonia and infrared thermal imaging sensors; the edge computing module is deployed in the sensor network, integrating the environmental data monitored by the sensors and performing preliminary analysis. When abnormal environmental data is detected, an environmental warning is issued to trigger the behavior analysis module to start monitoring.

[0015] Furthermore, the behavioral baseline comparison unit compares the behavioral data stream collected by the intelligent visual monitoring unit with the historical health baseline of each poultry, identifies individuals that deviate from the baseline, i.e., individuals with abnormal behavior, and generates a risk score.

[0016] Furthermore, the health archive integrates the behavioral characteristics, environmental data, and historical health data of the poultry, dynamically generating and updating the health profile of each poultry;

[0017] In the health archive, based on time series analysis and anomaly detection models, behavioral characteristics and environmental data are analyzed to identify individuals' abnormal behavior patterns and health trends, and to determine whether there are potential diseases or health risks.

[0018] Further, the warning response management module includes an abnormal warning unit and a push response unit;

[0019] The abnormal warning unit generates warning information for individual abnormalities, including risk sources, abnormal feature descriptions, and potential environmental factors;

[0020] The push response unit pushes the warning information to the management personnel; and records the management personnel's handling measures in the health archive as historical health data.

[0021] On the other hand, the present invention provides a poultry farming monitoring and management method based on big data, comprising:

[0022] Step S1, poultry house data monitoring,

[0023] The environmental monitoring system in the poultry house monitors the environmental data of the poultry house, including air quality, temperature and humidity, and individual body temperature. The edge computing module integrates and preliminarily analyzes the collected data. When the environmental data exceeds the set threshold, the behavior analysis module is triggered to monitor the individual health status of the poultry.

[0024] Step S2, behavioral feature collection and preliminary screening,

[0025] After the behavior analysis module is triggered to start monitoring, the camera is activated to scan each poultry in the poultry house, extracting individual behavior characteristics, including feeding frequency, activity level, posture changes, and individual body temperature, and establishing a behavior data stream for each poultry;

[0026] Compare the collected behavioral characteristics with the historical health benchmark of each poultry, conduct preliminary screening for individuals that deviate from the benchmark, generate risk scores and mark high-risk individuals, which are recorded in the individual health records;

[0027] Step S3, individual health record data analysis,

[0028] Integrate poultry behavioral characteristics, environmental data and historical health data to dynamically generate and update each poultry's health profile and record the latest risk score;

[0029] Based on environmental data and individual behavioral characteristics, time series analysis and anomaly detection models are used to identify abnormal behavior patterns and potential health trends of poultry, and determine whether there are diseases or health risks;

[0030] Step S4, individual abnormal warning,

[0031] Based on the individual health record data analysis results of step S3, when it is detected that an individual has a disease or health risk, specific abnormal warning information is generated, including the source of the risk, abnormal feature description and influencing environmental factors; the warning information is pushed to the management personnel.

[0032] Furthermore, in step S1, the poultry house data monitoring method is:

[0033] The sensor network data is combined into a unified acquisition matrix. Let the vector of data collected at time t be X t :X t =[Tt ,H t ,C t ,A t ,W t ], where T t Indicates the temperature in the poultry house, H t is humidity, C t represents the carbon dioxide concentration, A t Indicates the ammonia concentration, W t is the mean body temperature of the individual, and the vector X t Normalize the parameters in: Among them, max(X) and min(X) are the maximum and minimum values ​​of each parameter in the historical data respectively;

[0034] Build environment anomaly score S t Quantify the abnormality of the current environment and define w T 、w H 、w C 、w A 、w W is the weight of each parameter, then S t =w T ·T' t +w H ·H' t +w C ·C' t +w A ·A' t +w W ·W' t , where T' t , H' t , C' t , A' t , W' t is the normalized temperature, humidity, carbon dioxide, ammonia concentration and individual body temperature value, w T +w H +w C +w A +w W =1;

[0035] Average environmental score calculated based on historical data and standard deviation σ S , set the dynamic threshold S th :

[0036] Among them, k is the sensitivity coefficient, which is used to adjust the sensitivity of the threshold to fluctuations. If S t >S th , the environment is judged to be abnormal. When S t >S thWhen the behavior analysis module is triggered, individual health status monitoring is started.

[0037] Furthermore, in step S2, the behavior characteristics collection and preliminary screening method is:

[0038] The camera cooperates with the deep learning model to analyze the image data in the poultry house in real time, extracting the following characteristics of each poultry: feeding frequency, activity level, posture changes and individual body temperature characteristics. Within the monitoring time interval Δt, the number of times the poultry enters the feeding area is detected, and the feeding frequency f is calculated. i : Among them, n i represents the number of detections of poultry i in the feeding area, and Δt is the time interval;

[0039] Track the movement trajectory of poultry and calculate the total distance d moved per unit time i As a measure of activity level: Among them, a i represents the activity level of poultry i, d i is the moving distance of poultry i within Δt;

[0040] Analysis of the frequency of changes in standing and resting postures of poultry i : Among them, m i represents the number of posture switching times of poultry i within time Δt;

[0041] Extract the body temperature w of poultry i based on infrared thermal imaging data i , as one of the health indicators;

[0042] The above characteristics were compared with the historical health baseline of each bird {f i0 ,a i0 ,p i0 ,w i0} to compare, calculate the degree of feature deviation, and generate a risk score. The degree of deviation of each behavioral feature It is expressed as:

[0043] Among them, f i0 、a i0 、p i0 、w i0 The historical baseline values ​​of the feeding frequency, activity level, posture changes, and body temperature of poultry i;

[0044] Combining all the deviations, calculate the total risk score R i Quantifying individual health risks:

[0045] Among them, α, β, γ, δ are the weight coefficients of each feature, α+β+γ+δ=1; if R iIf the set health threshold is exceeded, poultry i is marked as a high-risk individual and recorded in the health record.

[0046] Furthermore, in step S3, the individual health record data is analyzed in the following manner:

[0047] Set poultry i in the time period [t1,t n ] and the environmental data are integrated into the feature matrix D i , used to update health records in real time:

[0048]

[0049] Among them, f(t) represents the frequency of eating at time t, a(t) represents the activity level, p(t) represents the frequency of posture changes, w(t) represents the individual body temperature, e(t) is the environmental anomaly score, and the matrix D i Automatic updates over time;

[0050] Use time series models to analyze the behavior patterns of poultry in different periods of time and identify potential health change trends. For each behavioral feature, use the autoregressive moving average model ARIMA to predict the future state and detect deviations. For the future feeding frequency of feature f(t), the predicted value It is expressed as:

[0051] Among them, φ k is the autoregressive coefficient, θ is the noise term, reflecting the impact of accidental factors on behavior, and k is the lag order, which is determined according to the periodicity of the behavior data. The model is applied to a(t), p(t), and w(t) to predict their respective future states.

[0052] Based on the calculation of the residual value r of the behavioral characteristics f 、r a 、r p 、r w , quantifying the deviation between the predicted behavior and the actual behavior, the residual value formula is:

[0053] Combine the residual values ​​of each feature to generate a health risk score R: R = αr f +βr a +γr p +δr w +ηe(t), where α, β, γ, δ, and η are feature weights, and R is the health risk score. If R exceeds the threshold, it is marked as abnormal, indicating potential health risks.

[0054] Furthermore, in step S4, the individual abnormal warning method is:

[0055] According to the current health risk score R of poultry i i , determine whether it exceeds the risk threshold R th To trigger an early warning: ΔR = R i -R th , where R i R is the current health risk score of poultry i. th is the set threshold for health risk, ΔR represents the health score deviation, and when ΔR>0, the system triggers an early warning;

[0056] The total risk score was decomposed to analyze the contribution of each characteristic (eating frequency, activity level, posture change and body temperature) to abnormality C j , identify the specific source of the abnormal feature and define the contribution as Among them, r j is the residual value of the jth feature (such as eating frequency, activity level, etc.), indicating the difference between the actual value and the predicted value, w j is the weight of the feature, reflecting its impact on the poultry health score, C j represents the contribution of feature j to the overall health risk score;

[0057] The contribution of each feature C j After normalization, the feature comprehensive score S is generated i , used to highlight key abnormal features when the system generates warning information: Among them, n is the number of features, S i represents the comprehensive abnormality score of poultry i, which is used to mark the influence of key features in abnormal description, and k is the summation index;

[0058] When ΔR>0 is detected, an early warning message is generated, including the following:

[0059] Risk source: Based on contribution C j Identify the main abnormal features,

[0060] Abnormal feature description: Comprehensive score S i High-impact features provided,

[0061] And influencing environmental factors: Based on the environmental data recorded in the archives (such as temperature, humidity and gas concentration), the potential health impacts are analyzed and attached to the early warning information, which is pushed to the management system in real time.

[0062] The beneficial effects of the present invention are:

[0063] The present invention monitors the environment in the poultry house in real time for data fusion and analysis. When an abnormality is detected, the behavior monitoring module is immediately triggered and a specific environmental abnormality score is generated. By combining visual monitoring and infrared thermal imaging, the individual behavior characteristics of each poultry are accurately extracted, and an individualized behavior data stream is generated, which effectively identifies individual potential health risks and reduces the risk of early missed detection of individual poultry diseases.

[0064] The present invention accumulates the long-term behavior and environmental data of each poultry through a health archive, and uses a time series analysis model to identify individual behavior patterns and health change trends. It can predict possible health risks based on historical data and models, effectively improving the foresight of management and the efficiency of disease prevention and control.

[0065] The present invention automatically identifies the most influential abnormal features through comprehensive risk scoring and contribution analysis, and pushes them to management personnel. At the same time, it updates the health records of each poultry in real time, records abnormal conditions and treatment measures, forms a closed-loop management of data, and reduces the frequency and burden of intervention by management personnel.

[0066] The present invention, when generating an abnormal warning, will mark the main risk sources, abnormal characteristics and environmental influencing factors, so that management personnel can quickly judge and accurately locate the source of the problem, make emergency treatment more targeted, avoid unnecessary intervention, reduce management difficulty, and improve the overall operation efficiency of the farm. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0068] Figure 1 This is a schematic diagram of the structure of the poultry breeding monitoring and management system of the present invention;

[0069] Figure 2 It is a schematic diagram of the flow chart of the poultry breeding monitoring and management method of the present invention. DETAILED DESCRIPTION

[0070] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0071] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0072] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0073] Example 1, reference Figure 1 This embodiment provides a poultry farming monitoring and management system based on big data, including:

[0074] An environmental monitoring system is used to monitor the environmental data in the poultry house and conduct preliminary analysis on it. The environmental monitoring system includes a sensor network and an edge computing module, and uses a sensor network to obtain environmental data;

[0075] In the environmental monitoring system, the environmental data collected include air quality, temperature and humidity, and individual body temperature in the poultry house;

[0076] The sensor network is composed of temperature and humidity, carbon dioxide, ammonia and infrared thermal imaging sensors. The edge computing module is deployed in the sensor network to integrate the environmental data monitored by the sensors and perform preliminary analysis. When abnormal environmental data is detected, an environmental warning is issued to trigger the behavior analysis module to start monitoring.

[0077] The behavior analysis module is triggered by environmental warnings and uses visual technology to monitor and preliminarily screen the behavior status of individual poultry and identify individuals with abnormal behavior. The behavior analysis module includes an intelligent visual monitoring unit and a behavior benchmark comparison unit.

[0078] The intelligent visual monitoring unit uses a camera to scan the poultry house and extract individual behavioral characteristics, including feeding frequency, movement status, and posture changes, to establish a behavioral data stream for each poultry. In the intelligent visual monitoring unit, the behavioral characteristics collected by the camera are combined with the individual body temperature monitored by the infrared thermal imaging sensor for joint analysis.

[0079] The behavioral benchmark comparison unit compares the behavioral data stream collected by the intelligent visual monitoring unit with the historical health benchmark of each poultry, identifies individuals that deviate from the benchmark, i.e., abnormal behavior individuals, and generates a risk score;

[0080] Health profile library, which generates, updates and stores individual health profiles of poultry based on behavioral characteristics and environmental data, and uses in-depth analysis of individual health profiles to identify potential health trends and risks;

[0081] The health record library integrates the behavioral characteristics, environmental data and historical health data of poultry, dynamically generates and updates the health record of each poultry;

[0082] In the health archive, based on time series analysis and anomaly detection models, behavioral characteristics and environmental data are analyzed to identify abnormal behavior patterns and health trends of individuals and determine whether there are potential diseases or health risks;

[0083] The early warning response management module notifies management personnel when a health risk is identified in the health archive for an individual poultry;

[0084] The early warning response management module includes an abnormal early warning unit and a push response unit;

[0085] The abnormal warning unit generates warning information for individual abnormalities, including risk sources, abnormal feature descriptions, and potential environmental factors;

[0086] The push response unit pushes the warning information to the management personnel; and records the management personnel's handling measures in the health archive as historical health data.

[0087] Example 2, reference Figure 2 This embodiment provides a poultry farming monitoring and management method based on big data, comprising the following steps:

[0088] Step S1, poultry house data monitoring,

[0089] The environmental monitoring system in the poultry house monitors the environmental data of the poultry house, including air quality, temperature and humidity, and individual body temperature. The edge computing module integrates and preliminarily analyzes the collected data. When the environmental data exceeds the set threshold, the behavior analysis module is triggered to monitor the individual health status of the poultry.

[0090] In step S1, the poultry house data monitoring method is:

[0091] The sensor network data is combined into a unified acquisition matrix. Let the vector of data collected at time t be X t :X t =[T t ,H t ,C t ,A t ,W t ], where T t Indicates the temperature in the poultry house, H t is humidity, C t represents the carbon dioxide concentration, A t Indicates the ammonia concentration, W t is the mean body temperature of the individual, and the vector X t Normalize the parameters in: Among them, max(X) and min(X) are the maximum and minimum values ​​of each parameter in the historical data respectively;

[0092] Build environment anomaly score S t Quantify the abnormality of the current environment and define w T 、w H 、w C 、w A 、w W is the weight of each parameter, then S t =w T ·T' t +w H ·H' t +w C ·C' t +w A ·A' t +w W ·W' t , where T' t , H' t , C' t , A' t , W' t is the normalized temperature, humidity, carbon dioxide, ammonia concentration and individual body temperature value, w T +w H +w C +w A +w W =1;

[0093] Average environmental score calculated based on historical data and standard deviation σ S , set the dynamic threshold S th :

[0094] Among them, k is the sensitivity coefficient, which is used to adjust the sensitivity of the threshold to fluctuations. If S t >S th , the environment is judged to be abnormal. When S t >S th When the behavior analysis module is triggered, individual health status monitoring is started;

[0095] Specifically, Through the sensor network, the temperature and humidity, carbon dioxide, ammonia concentration, and individual body temperature data in the poultry house are monitored in real time. The edge computing module is used for data fusion and preliminary analysis. Once an abnormality is detected, an early warning is triggered to dynamically reflect changes in environmental conditions, quickly identify environmental factors that affect poultry health, and be aware of fluctuations in environmental conditions at the first time, reducing potential threats to poultry health.

[0096] Step S2, behavioral feature collection and preliminary screening,

[0097] After the behavior analysis module is triggered to start monitoring, the camera is activated to scan each poultry in the poultry house, extracting individual behavior characteristics, including feeding frequency, activity level, posture changes, and individual body temperature, and establishing a behavior data stream for each poultry;

[0098] Compare the collected behavioral characteristics with the historical health benchmark of each poultry, conduct preliminary screening for individuals that deviate from the benchmark, generate risk scores and mark high-risk individuals, which are recorded in the individual health records;

[0099] In step S2, the behavior characteristics collection and preliminary screening method is:

[0100] The camera cooperates with the deep learning model to analyze the image data in the poultry house in real time, extracting the following characteristics of each poultry: feeding frequency, activity level, posture changes and individual body temperature characteristics. Within the monitoring time interval Δt, the number of times the poultry enters the feeding area is detected, and the feeding frequency f is calculated. i : Among them, n i represents the number of detections of poultry i in the feeding area, and Δt is the time interval;

[0101] Track the movement of poultry and calculate the total distance d moved per unit time i As a measure of activity level: Among them, a i represents the activity level of poultry i, d i is the moving distance of poultry i within Δt;

[0102] Analysis of the frequency of changes in standing and resting postures of poultry i : Among them, m i represents the number of posture switching times of poultry i within time Δt;

[0103] Extract the body temperature w of poultry i based on infrared thermal imaging data i , as one of the health indicators;

[0104] The above characteristics were compared with the historical health baseline of each bird {f i0 ,a i0 ,p i0 ,w i0} to compare, calculate the degree of feature deviation, and generate a risk score. The degree of deviation of each behavioral feature It is expressed as:

[0105] Among them, f i0 、a i0 、p i0 、w i0 The historical baseline values ​​of the feeding frequency, activity level, posture changes, and body temperature of poultry i;

[0106] Combining all the deviations, calculate the total risk score R i Quantifying individual health risks:

[0107] Among them, α, β, γ, δ are the weight coefficients of each feature, α+β+γ+δ=1; if R i If the set health threshold is exceeded, poultry i is marked as a high-risk individual and recorded in the health record;

[0108] Specifically, Combining intelligent vision modules with infrared thermal imaging technology, the system can extract the eating frequency, activity level, posture changes and body temperature characteristics of each poultry in real time, generate a dynamic behavioral data stream, and perform preliminary screening while collecting behavioral characteristics. The poultry behavior data is compared with its historical health benchmark, the deviation is quantified, and real-time risk scores are generated and high-risk individuals are marked. This can quickly identify poultry individuals with abnormal behavior and help breeding managers detect health risks at an early stage.

[0109] Step S3, individual health record data analysis,

[0110] Integrate poultry behavioral characteristics, environmental data and historical health data to dynamically generate and update each poultry's health profile and record the latest risk score;

[0111] Based on environmental data and individual behavioral characteristics, time series analysis and anomaly detection models are used to identify abnormal behavior patterns and potential health trends of poultry, and determine whether there are diseases or health risks;

[0112] In step S3, the individual health record data is analyzed in the following manner:

[0113] Set poultry i in the time period [t1,t n ] and the environmental data are integrated into the feature matrix D i , used to update health records in real time:

[0114]

[0115] Among them, f(t) represents the frequency of eating at time t, a(t) represents the activity level, p(t) represents the frequency of posture changes, w(t) represents the individual body temperature, e(t) is the environmental anomaly score, and the matrix D i Automatic updates over time;

[0116] Use time series models to analyze the behavior patterns of poultry in different periods of time and identify potential health change trends. For each behavioral feature, use the autoregressive moving average model ARIMA to predict the future state and detect deviations. For the future feeding frequency of feature f(t), the predicted value It is expressed as:

[0117] Among them, φ k is the autoregressive coefficient, θ is the noise term, reflecting the impact of accidental factors on behavior, and k is the lag order, which is determined according to the periodicity of the behavior data. The model is applied to a(t), p(t), and w(t) to predict their respective future states.

[0118] Based on the calculation of the residual value r of the behavioral characteristics f 、r a 、r p 、r w , quantifying the deviation between the predicted behavior and the actual behavior, the residual value formula is:

[0119] Combine the residual values ​​of each feature to generate a health risk score R: R = αr f +βr a +γr p +δr w +ηe(t), where α, β, γ, δ, and η are feature weights, and R is the health risk score. If R exceeds the threshold, it is marked as abnormal, indicating potential health risks;

[0120] Specifically, The health archive can dynamically generate, update and store the behavioral characteristics and environmental data of poultry. Based on time series analysis and anomaly detection models, the behavioral patterns and health change trends of poultry can be identified from the data.

[0121] Step S4, individual abnormal warning,

[0122] Based on the individual health record data analysis results of step S3, when it is detected that an individual has a disease or health risk, specific abnormal warning information is generated, including the source of the risk, abnormal feature description and influencing environmental factors; and the warning information is pushed to the management personnel;

[0123] In step S4, the individual abnormal warning method is:

[0124] According to the current health risk score R of poultry i i , determine whether it exceeds the risk threshold R th To trigger an early warning: ΔR = R i -R th , where R i R is the current health risk score of poultry i. th is the set threshold for health risk, ΔR represents the health score deviation, and when ΔR>0, the system triggers an early warning;

[0125] The total risk score was decomposed to analyze the contribution of each characteristic (eating frequency, activity level, posture change and body temperature) to abnormality C j , identify the specific source of the abnormal feature and define the contribution as Among them, r j is the residual value of the jth feature (such as eating frequency, activity level, etc.), indicating the difference between the actual value and the predicted value, w j is the weight of the feature, reflecting its impact on the poultry health score, C j represents the contribution of feature j to the overall health risk score;

[0126] The contribution of each feature C j After normalization, the feature comprehensive score S is generated i , used to highlight key abnormal features when the system generates warning information: Among them, n is the number of features, S i represents the comprehensive abnormality score of poultry i, which is used to mark the influence of key features in abnormal description, and k is the summation index;

[0127] When ΔR>0 is detected, an early warning message is generated, including the following:

[0128] Risk source: Based on contribution C j Identify the main abnormal features,

[0129] Abnormal feature description: Comprehensive score S i High-impact features provided,

[0130] And influencing environmental factors: Based on the environmental data recorded in the archives (such as temperature, humidity and gas concentration), the potential health impacts are analyzed and attached to the early warning information, which is pushed to the management system in real time;

[0131] Specifically, the early warning response management module can generate early warning information in a timely manner when an anomaly is discovered and push it to management personnel. Through multi-feature contribution analysis, the abnormal features are scored and normalized to identify the most influential abnormal indicators.

[0132] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A poultry farming monitoring and management system based on big data, characterized by: include, An environmental monitoring system is used to monitor the environmental data in the poultry house and conduct preliminary analysis on it. The environmental monitoring system includes a sensor network and an edge computing module, and uses a sensor network to obtain environmental data; The behavior analysis module is triggered by environmental warnings and uses visual technology to monitor and preliminarily screen the behavior status of individual poultry and identify individuals with abnormal behavior; The behavior analysis module includes an intelligent visual monitoring unit and a behavior benchmark comparison unit; The intelligent visual monitoring unit uses a camera to scan the poultry house and extract individual behavioral characteristics, including feeding frequency, movement status, and posture changes, to establish a behavioral data stream for each poultry; In the intelligent visual monitoring unit, the behavioral features collected by the camera are combined with the individual body temperature monitored by the infrared thermal imaging sensor for joint analysis; Health profile library, which generates, updates and stores individual health profiles of poultry based on behavioral characteristics and environmental data, and uses in-depth analysis of individual health profiles to identify potential health trends and risks; The early warning response management module notifies the management personnel when a health risk is identified in the health record library for an individual poultry.

2. A poultry farming monitoring and management system based on big data according to claim 1, characterized in that: In the environmental monitoring system, the environmental data collected include air quality, temperature and humidity, and individual body temperature in the poultry house; The sensor network is composed of temperature and humidity, carbon dioxide, ammonia and infrared thermal imaging sensors; the edge computing module is deployed in the sensor network, integrating the environmental data monitored by the sensors and performing preliminary analysis. When abnormal environmental data is detected, an environmental warning is issued to trigger the behavior analysis module to start monitoring.

3. A poultry farming monitoring and management system based on big data according to claim 2, characterized in that: The behavioral baseline comparison unit compares the behavioral data stream collected by the intelligent visual monitoring unit with the historical health baseline of each poultry, identifies individuals that deviate from the baseline, i.e., individuals with abnormal behavior, and generates a risk score.

4. The poultry breeding monitoring and management system based on big data according to claim 3 is characterized in that: The health record library integrates the behavioral characteristics, environmental data and historical health data of poultry, dynamically generates and updates the health record of each poultry; In the health archive, based on time series analysis and anomaly detection models, behavioral characteristics and environmental data are analyzed to identify individuals' abnormal behavior patterns and health trends, and to determine whether there are potential diseases or health risks.

5. The poultry breeding monitoring and management system based on big data according to claim 4 is characterized in that: The early warning response management module includes an abnormal early warning unit and a push response unit; The abnormal warning unit generates warning information for individual abnormalities, including risk sources, abnormal feature descriptions, and potential environmental factors; The push response unit pushes the warning information to the management personnel; The management personnel’s handling measures will be recorded in the health archive as historical health data.

6. A poultry farming monitoring and management method based on big data, characterized in that: A poultry farming monitoring and management system based on big data according to any one of claims 1 to 5, The following steps are included: Step S1, poultry house data monitoring, The environmental monitoring system in the poultry house monitors the environmental data of the poultry house, including air quality, temperature and humidity, and individual body temperature. The edge computing module integrates and preliminarily analyzes the collected data. When the environmental data exceeds the set threshold, the behavior analysis module is triggered to monitor the individual health status of the poultry. Step S2, behavioral feature collection and preliminary screening, After the behavior analysis module is triggered to start monitoring, the camera is activated to scan each poultry in the poultry house, extracting individual behavior characteristics, including feeding frequency, activity level, posture changes, and individual body temperature, and establishing a behavior data stream for each poultry; Compare the collected behavioral characteristics with the historical health benchmark of each poultry, conduct preliminary screening for individuals that deviate from the benchmark, generate risk scores and mark high-risk individuals, which are recorded in the individual health records; Step S3, individual health record data analysis, Integrate poultry behavioral characteristics, environmental data and historical health data to dynamically generate and update each poultry's health profile and record the latest risk score; Based on environmental data and individual behavioral characteristics, time series analysis and anomaly detection models are used to identify abnormal behavior patterns and potential health trends of poultry, and determine whether there are diseases or health risks; Step S4, individual abnormal warning, Based on the individual health record data analysis results of step S3, when it is detected that the individual has a disease or health risk, specific abnormal warning information is generated, including the source of the risk, abnormal feature description and influencing environmental factors; Push warning information to management personnel.

7. The poultry farming monitoring and management method based on big data according to claim 6 is characterized in that: In step S1, the poultry house data monitoring method is: The sensor network data is combined into a unified acquisition matrix. Let the vector of data collected at time t be X t :X t =[T t ,H t ,C t ,A t ,W t ], where T t Indicates the temperature in the poultry house, H t is humidity, C t represents the carbon dioxide concentration, A t Indicates the ammonia concentration, W t is the mean body temperature of the individual, and the vector X t Normalize the parameters in: Among them, max(X) and min(X) are the maximum and minimum values ​​of each parameter in the historical data respectively; Build environment anomaly score S t Quantify the abnormality of the current environment and define w T 、w H 、w C 、w A 、w W is the weight of each parameter, then S t =w T ·T' t +w H ·H' t +w C ·C' t +w A ·A' t +w W ·W' t , where T' t , H' t , C' t , A' t , W' r is the normalized temperature, humidity, carbon dioxide, ammonia concentration and individual body temperature value, w T +w H +w C +w A +w W =1; Average environmental score calculated based on historical data and standard deviation σ S , set the dynamic threshold S th : Among them, k is the sensitivity coefficient, which is used to adjust the sensitivity of the threshold to fluctuations. If S t >S th , the environment is judged to be abnormal. When S t >S th When the behavior analysis module is triggered, individual health status monitoring is started.

8. The poultry farming monitoring and management method based on big data according to claim 7 is characterized in that: In step S2, the behavior characteristics collection and preliminary screening method is: The camera works with the deep learning model to analyze the image data in the poultry house in real time, extracting the following characteristics for each poultry: feeding frequency, activity level, posture changes, and individual body temperature characteristics. Within the monitoring time interval Δt, the number of times the poultry enters the feeding area is detected, and the feeding frequency is calculated. Among them, n i represents the number of detections of poultry i in the feeding area, and Δt is the time interval; Track the movement trajectory of poultry and calculate the total distance d moved per unit time i As a measure of activity level: Among them, a i represents the activity level of poultry i, d i is the moving distance of poultry i within Δt; Analyze the frequency of standing and resting posture changes in poultry Among them, m i represents the number of posture switching times of poultry i within time Δt; Extract the body temperature w of poultry i based on infrared thermal imaging data i , as one of the health indicators; The above characteristics were compared with the historical health baseline of each bird {f i0 ,a i0 ,p i0 ,w i0 } to compare, calculate the degree of feature deviation, and generate a risk score. The degree of deviation of each behavioral feature It is expressed as: Among them, f i0 、a i0 、p i0 、w i0 The historical baseline values ​​of the feeding frequency, activity level, posture changes, and body temperature of poultry i; Combining all the deviations, calculate the total risk score R i Quantifying individual health risks: Among them, α, β, γ, δ are the weight coefficients of each feature, α+β+γ+δ=1; if R i If the set health threshold is exceeded, poultry i is marked as a high-risk individual and recorded in the health record.

9. The poultry farming monitoring and management method based on big data according to claim 8, characterized in that: In step S3, the individual health record data is analyzed in the following manner: Set poultry i in the time period [t1,t n ] and the environmental data are integrated into the feature matrix D i , used to update health records in real time: Among them, f(t) represents the frequency of eating at time t, a(t) represents the activity level, p(t) represents the frequency of posture changes, w(t) represents the individual body temperature, e(t) is the environmental anomaly score, and the matrix D i Automatic updates over time; Use time series models to analyze the behavior patterns of poultry at different times and identify potential health trends. For each behavior feature, use the autoregressive moving average model ARIMA to predict the future state and detect deviations. For the future feeding frequency of feature f(t), the predicted value It is expressed as: Among them, φ k is the autoregressive coefficient, θ is the noise term, reflecting the impact of accidental factors on behavior, and k is the lag order, which is determined according to the periodicity of the behavior data. The model is applied to a(t), p(t), and w(t) to predict their respective future states. Based on the calculation of the residual value r of the behavioral characteristics f 、r a 、r p 、r w , quantifying the deviation between the predicted behavior and the actual behavior, the residual value formula is: Combine the residual values ​​of each feature to generate a health risk score R: R = αr f +βr a +γr p +δr w +ηe(t), where α, β, γ, δ, and η are feature weights, and R is the health risk score. If R exceeds the threshold, it is marked as abnormal, indicating potential health risks.

10. A poultry farming monitoring and management method based on big data according to claim 9, characterized in that: In step S4, the individual abnormal warning method is: According to the current health risk score R of poultry i i , determine whether it exceeds the risk threshold R th To trigger an early warning: ΔR = R i -R th , where R i R is the current health risk score of poultry i. th is the set threshold for health risk, ΔR represents the health score deviation, and when ΔR>0, the system triggers an early warning; Decompose the total risk score and analyze the contribution of each feature to the anomaly C j , identify the specific source of the abnormal feature and define the contribution as Among them, r j is the residual value of the jth feature, indicating the difference between the actual value and the predicted value, w j is the weight of the feature, reflecting its impact on the poultry health score, C j represents the contribution of feature j to the overall health risk score; The contribution of each feature C j After normalization, the feature comprehensive score is generated Among them, n is the number of features, S i represents the comprehensive abnormality score of poultry i, k is the summation index; when ΔR>0 is detected, an early warning message is generated.

Images