Pastoral in-situ stress-free full life cycle production performance measurement system

Abstract

The present invention relates to the field of life cycle monitoring technology, and specifically to an in-situ stress-free full-life cycle production performance measurement system in pastoral areas. In the present invention, through comprehensive analysis of life cycle data, the growth stages of animals are accurately divided, and real-time monitoring of production performance in each stage is achieved. By collecting key data of animals, it is determined whether there is performance fluctuation and a production performance fluctuation identification result is generated. The life cycle stages are accurately divided through a hidden Markov model, which provides more targeted decision-making support for breeding management and enhances the stability and efficiency of overall production. The production performance is dynamically monitored through a convolutional neural network, and the fluctuation of production efficiency is judged according to a standard range, so that the monitoring of the growth process, production performance and health status of the animals can be continuously updated, and health abnormalities and trends of declining production efficiency can be identified in a timely manner, and corresponding adaptive adjustments can be made according to environmental changes.

Description

Technical Field

[0001] The present invention relates to the technical field of life cycle monitoring, and in particular to an in-situ stress-free full-life cycle production performance measurement system in pastoral areas. Background Art

[0002] The field of life cycle monitoring technology aims to comprehensively track and evaluate the performance of individuals and groups throughout their life cycle. By collecting data in real time and analyzing the performance of animals at different growth stages, it helps evaluate the long-term impact of environmental factors, feeding conditions and management methods on animal health and production capacity.

[0003] The purpose of the pastoral in-situ stress-free full life cycle production performance measurement system is to accurately measure the production performance of animals at each stage from birth to death through stress-free environmental monitoring, improve pastoral breeding efficiency, optimize animal production performance, avoid stress reactions in animals caused by human interference, ensure the authenticity and stability of data, and at the same time provide accurate production performance evaluation and real-time monitoring to help breeders reduce production costs and improve production efficiency. Through precise data support, it provides a scientific basis for feeding management and environmental adjustment, improves the overall benefits of pastoral breeding, and can provide innovative solutions for related industries, promote the industry to develop in a more efficient and sustainable direction, and thus bring considerable economic benefits and social value.

[0004] Existing technologies are not comprehensive enough for real-time judgment of production performance monitoring and health status during the life cycle. They are limited to monitoring of a single stage and certain specific indicators, and rely on sampling and testing at regular and fixed time points. They cannot reflect the dynamic performance of animals in different growth stages in real time, resulting in the inability to capture small fluctuations in animal health and production capacity in a timely manner in long-term tracking and evaluation, missing the best time for intervention, lacking accurate data analysis for different growth stages, and unable to conduct timely production performance evaluation based on actual data fluctuations. When production efficiency is low, it is impossible to accurately determine the specific cause, and it is impossible to conduct a comprehensive evaluation based on environmental factors and health status. The animal production performance monitoring system relies on Reliance on a single data source and a lack of comprehensive analysis of multi-dimensional data results in inaccurate division of life cycle stages, poor adaptability to environmental changes, inability to dynamically adjust to climate change and changes in feeding conditions, and inability to effectively determine the adaptation of animals through environmental adaptability analysis, leading to delayed diagnosis of animal health problems and unnecessary losses. In addition, existing technologies have problems with data processing lags. Due to the low frequency of data collection and delays in the processing process, real-time feedback on animal production performance cannot be achieved. It is difficult to determine through real-time data whether the animals are in a state of stress and whether there is a potential risk of decreased production capacity, resulting in an inability to make adjustments in the shortest time. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose an in-situ stress-free full-life cycle production performance measurement system for pastoral areas.

[0006] In order to achieve the above-mentioned purpose, the present invention adopts the following technical solution: the pastoral in-situ stress-free full life cycle production performance measurement system includes:

[0007] Life cycle stage identification module: This module collects animal weight, growth rate, and activity patterns, filters the data based on the life cycle, and uses a hidden Markov model to compare weight changes and activity frequency hour by hour to determine whether the data meets the criteria for different life cycle stages. Based on the set weight and growth rate thresholds, the module performs data classification operations to generate life cycle stage division results.

[0008] Dynamic production performance monitoring module: Based on the life cycle stage division results, animal weight, feed conversion rate and daily weight gain are collected. A convolutional neural network is used to compare the data according to the set standard range to determine whether the data exceeds the predetermined range. The monitoring data is updated item by item, and the production efficiency of each stage is gradually calculated to generate the production performance fluctuation identification results.

[0009] Health status monitoring module: Based on the production performance fluctuation identification results, combined with physiological data, it determines whether the data exceeds the normal physiological range, performs comparative analysis to determine whether abnormal fluctuations occur, detects the degree of deviation between health indicators and standard thresholds in real time, performs health status confirmation operations, and generates health status monitoring results;

[0010] Environmental adaptability analysis module: Based on the health status monitoring results, environmental data is collected and correlated with the animal health data item by item. By comparing the correlation between changes in environmental factors and changes in health status one by one, the environmental adaptability is calculated and the environmental adaptability analysis results are generated;

[0011] Risk warning and response module: Based on the health status monitoring results and environmental adaptability analysis results, combined with climate change and feeding adjustment data, and using historical fluctuation data to generate warnings, risk judgment is made by setting risk thresholds, and warning processing is executed when deviations from standard thresholds occur, generating health and production risk warning results;

[0012] Farming strategy adjustment module: Based on the health status monitoring results and health and production risk warning results, combined with current production performance data, it determines whether the feeding management plan needs to be adjusted. By analyzing health and production fluctuations, the feeding strategy is adjusted in real time to generate the breeding strategy adjustment results;

[0013] This system is applicable to a variety of pasture environments. In plateau pastures, the system can dynamically adjust feeding strategies based on the impact of low-oxygen environments on animal health. Through real-time monitoring of animal physiological indicators and environmental adaptability analysis, it can optimize stocking density, feed ratios, rest and activity arrangements. In tropical pastures, the system can optimize feed ratios, drinking water systems and indoor ventilation conditions in real time based on the impact of high temperature and high humidity environments on animal production efficiency, reduce stress responses caused by high temperatures, and improve animals' feed conversion rates and daily weight gain. At the same time, through continuous monitoring of the environment and production performance, the system can promptly identify potential risks of climate change to animal health and proactively take adjustment measures.

[0014] As a further solution of the present invention, the life cycle stage identification module includes a data collection submodule, a data screening submodule and a life cycle stage judgment submodule, wherein:

[0015] Data collection submodule: Based on the animal's weight, growth rate, and activity pattern data, it regularly collects the animal's weight and activity frequency. It uses sensors to record the animal's weight and activity pattern data every hour, obtains animal-related activity information through monitoring equipment, records and saves the data hourly, and generates collected data.

[0016] Data screening submodule: Based on the collected data, a hidden Markov model is used, assuming that different stages of the animal life cycle can correspond to different hidden states, each hidden state has a certain observation probability distribution, and through training on historical data, the state transition probability and the observation probability under each state of each life cycle stage are learned, and the data that meets the conditions is screened out by comparing with the set standard range. Data that fails to pass the standard screening is regarded as irrelevant data and is eliminated to generate screening data;

[0017] Life cycle stage judgment submodule: Based on the screening data, by comparing with the set weight and growth rate thresholds hourly, check whether the animal's weight changes and growth rate meet the standards, identify whether the animal is in the correct life cycle stage, calibrate the specific standards for each stage, and generate life cycle stage division results.

[0018] As a further solution of the present invention, the hidden Markov model is according to the formula:

[0019]

[0020] Where: P(X t |X t-1 ,W) represents the observation data X at the previous time point at time t t-1 and weight coefficient W, calculate the probability of the state transferring to the current time point t, X trepresents the observation data at time point t, including the animal's weight, growth rate and activity data, X t-1 represents the observed data at time point t-1, W represents the standard range of the life cycle stage, P(X t |S i ,W) represents a given hidden state S i and weight coefficient W when observing data X t The generation probability, S i represents the hidden state, representing the life cycle stage of the animal, P(S i |X t-1 ,W) represents the time from time point t-1 to time point t, given the observation data X t-1 and weight coefficient W, the hidden state S i The probability of transition, α i Represents the hidden state S i The weight coefficient is , and N represents the number of hidden states.

[0021] As a further solution of the present invention, the production performance dynamic monitoring module includes a production data acquisition submodule, a data comparison submodule and a production efficiency calculation submodule, wherein:

[0022] Production data collection submodule: Based on the life cycle stage division results, the animal's weight, feed conversion rate and daily weight gain data are collected item by item. The animal's weight changes, feed consumption and daily weight gain are recorded through real-time monitoring equipment to ensure the data is complete and timely, and generate collected production data;

[0023] Data comparison submodule: Based on the collected production data, a convolutional neural network is used, with animal weight, feed conversion rate and daily weight gain as input features, to construct a time series data matrix, and local feature extraction is performed on the changing trends of the features in different time periods. After feature extraction, the feature dimension is reduced through a pooling layer, and then a fully connected layer is used to determine whether each indicator meets the standard range. Through the labeled historical production performance data, the characteristic patterns under normal and abnormal conditions are learned, and then the deviations between the actual values ​​and target values ​​of animal weight, feed conversion rate and daily weight gain are compared item by item to identify whether the data deviates from the normal range, and the data that meets the standard is screened out to generate a comparison result;

[0024] Production efficiency calculation submodule: Based on the comparison results, the production efficiency of each life cycle stage is calculated step by step. By updating the data, the production efficiency of each stage is evaluated, and each data is weighted and integrated to calculate the final production efficiency value and generate the production performance fluctuation identification result.

[0025] As a further solution of the present invention, the convolutional neural network is based on the formula:

[0026]

[0027] Where: z(t,W,β,b) represents the output result after convolution, which is used to reflect the matching between data features and target standards; x(τ) represents the input data under the time and position index τ, including the collected weight, growth rate and daily weight gain data; w(t-τ) represents the convolution kernel, which is used to extract data features; τ represents the time and position index during the convolution process, which is used to calculate the input data item by item; W represents the weight coefficient; β represents the adjustment coefficient, which is used to adjust the relative importance between weight, growth rate and daily weight gain; b represents the bias term, which is used to balance the convolution results.

[0028] As a further solution of the present invention, the health status monitoring module includes a production data comparison submodule, a physiological data analysis submodule and a health status confirmation submodule, wherein:

[0029] Production data comparison submodule: Based on the production performance fluctuation identification results, collect animal weight, daily weight gain and feed consumption data, and compare them with the set normal physiological range to determine whether the data exceeds the physiological standard range. By comparing and screening the data item by item, the production data comparison results are generated;

[0030] Physiological data analysis submodule: Based on the production data comparison results, monitor the health indicators item by item, detect the deviation between the health data and the set standard threshold, analyze whether there is abnormal physiological fluctuation, and generate physiological status fluctuation results;

[0031] Health status confirmation submodule: Based on the physiological status fluctuation results, compare the deviation between the health data and the standard threshold, perform the health status confirmation operation, verify whether the health status is abnormal, combine historical data to evaluate the health trend, and generate health status monitoring results.

[0032] As a further solution of the present invention, the analysis determines whether abnormal physiological fluctuations occur. The definition of abnormal physiological fluctuations is based on the deviation between the health indicators in the animal's body and the preset normal physiological range. The health indicator data of the animal is collected regularly, and the standard range of each health indicator is set according to the normal health status of the animal population. At the same time, the health indicators collected in real time are compared with the set standard thresholds. If an indicator deviates from the normal range and exceeds the set tolerance, it is determined to be an abnormal physiological fluctuation.

[0033] As a further solution of the present invention, the environmental adaptability analysis module includes an environmental data acquisition submodule, a correlation analysis submodule and a fitness calculation submodule, wherein:

[0034] Environmental data collection submodule: Based on the health status monitoring results, collect data related to the environment, record environmental changes through real-time monitoring equipment, and generate environmental data;

[0035] Correlation analysis submodule: Based on the health status monitoring results and environmental data, the correlation between environmental factors and health status is analyzed item by item, and the correlation results of environmental factors are generated by comparing the relationship between each environmental change and health fluctuation;

[0036] Fitness calculation submodule: Based on the environmental factor association results, calculate the environmental fitness, quantify the impact of environmental changes on animal health, and generate environmental adaptability analysis results.

[0037] As a further solution of the present invention, the risk warning and response module includes a data acquisition submodule, a warning generation submodule and a warning processing submodule, wherein:

[0038] Data collection submodule: Based on the health status monitoring results and environmental adaptability analysis results, collect climate change data, feeding adjustment data and historical fluctuation data, monitor the environment and production data in real time through sensors, record climate, humidity, temperature and feeding adjustment information item by item, and generate collected environment and production data;

[0039] Warning generation submodule: Based on the collected environment and production data, by comparing the current health status with historical fluctuation data, combined with climate change and environmental data, it is determined whether the data exceeds the set risk threshold, and compared with the set risk standards to generate risk warning results;

[0040] Early warning processing submodule: Based on the risk early warning results, check whether there is any deviation from the standard threshold, perform early warning processing operations, mark the deviated data and take processing measures to generate health and production risk early warning results.

[0041] As a further solution of the present invention, the breeding strategy adjustment module includes a data analysis submodule, a strategy judgment submodule and an adjustment execution submodule, wherein:

[0042] Data analysis submodule: Based on the health status monitoring results and health and production risk warning results, combined with current production performance data, analyzes health and production fluctuation data, and generates health and production fluctuation analysis results by comparing production efficiency and health status fluctuations item by item;

[0043] Strategy judgment submodule: Based on the health and production fluctuation analysis results, it determines whether the feeding management plan needs to be adjusted. By comparing the current feeding conditions and production status, it evaluates whether the feed ratio and feeding method should be adjusted, and generates the strategy adjustment demand results;

[0044] Adjustment execution submodule: Based on the strategy adjustment demand results, execute the feeding strategy adjustment operation, update the feeding management plan in real time by modifying the feed ratio and adjusting the animal feeding environment, and generate the feeding strategy adjustment results.

[0045] Compared with the prior art, the advantages and positive effects of the present invention are:

[0046] 1. This invention comprehensively collects and analyzes animal lifecycle data, accurately dividing lifecycle stages. Based on the animal's key data, the animal's growth stage can be dynamically determined. This allows for real-time monitoring and timely evaluation of production performance at each growth stage, quickly identifying fluctuations in growth. When animal production performance fluctuates, fluctuation identification results are generated based on set thresholds, providing managers with specific feedback on production performance changes. Furthermore, stress-free data collection technology is employed to ensure data authenticity and stability.

[0047] 2. The present invention uses a hidden Markov model to finely divide life cycle stages, taking into account unobservable potential states and transition probabilities during animal growth, and can more accurately reflect the actual growth process of animals. The hidden Markov model infers the probability distribution of animals at different life cycle stages based on the observed data at each moment. By continuously updating the division of life cycle stages, dynamic tracking and adjustment of the animal's growth process are achieved, improving the accuracy and timeliness of data processing.

[0048] 3. In the present invention, the production performance of animals is dynamically monitored through a convolutional neural network, and the production performance indicators of animals are analyzed in real time. Based on the preset standard range, fluctuations in production efficiency are detected in a timely manner. When abnormal fluctuations are monitored, potential health problems and trends of declining production performance are identified, and early warnings are issued. The growth data of animals is continuously updated, and by learning and comparing historical data, future health trends and production fluctuations are predicted, providing accurate decision-making support for breeding managers. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 is a system flow chart of the present invention;

[0050] Figure 2 is a flow chart of the present invention;

[0051] Figure 3 Schematic diagram of the system framework of the present invention. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to 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.

[0053] See also Figure 1 The present invention provides a technical solution: a pastoral area in-situ stress-free full life cycle production performance measurement system comprising:

[0054] Life cycle stage identification module: This module collects animal weight, growth rate, and activity patterns, filters the data based on the life cycle, and uses a hidden Markov model to compare weight changes and activity frequency hour by hour to determine whether the data meets the criteria for different life cycle stages. Based on the set weight and growth rate thresholds, the module performs data classification operations to generate life cycle stage division results.

[0055] Dynamic Production Performance Monitoring Module: Based on the results of life cycle stage division, animal weight, feed conversion rate, and daily weight gain are collected. A convolutional neural network is used to compare the data with the set standard range to determine whether the data exceeds the predetermined range. The monitoring data is updated item by item, and the production efficiency of each stage is gradually calculated to generate the production performance fluctuation identification results.

[0056] Health status monitoring module: Based on the results of production performance fluctuation identification and combined with physiological data, it determines whether the data exceeds the normal physiological range, conducts comparative analysis to determine whether abnormal fluctuations occur, and detects the degree of deviation between health indicators and standard thresholds in real time. It performs health status confirmation operations and generates health status monitoring results;

[0057] Environmental adaptability analysis module: Based on the health status monitoring results, environmental data is collected and correlated with animal health data item by item. By comparing the correlation between changes in environmental factors and changes in health status one by one, the environmental adaptability is calculated and the environmental adaptability analysis results are generated;

[0058] Risk warning and response module: Based on health status monitoring results and environmental adaptability analysis results, combined with climate change and feeding adjustment data, and using historical fluctuation data to generate warnings, risk assessments are made by setting risk thresholds, and warnings are processed when deviations from standard thresholds are detected, generating health and production risk warning results;

[0059] Farming strategy adjustment module: Based on health status monitoring results and health and production risk warning results, combined with current production performance data, it determines whether the feeding management plan needs to be adjusted. By analyzing health and production fluctuations, the feeding strategy is adjusted in real time to generate the breeding strategy adjustment results;

[0060] This system is applicable to a variety of pasture environments. In plateau pastures, the system can dynamically adjust feeding strategies based on the impact of low-oxygen environments on animal health. Through real-time monitoring of animal physiological indicators and environmental adaptability analysis, it can optimize stocking density, feed ratios, rest and activity arrangements. In tropical pastures, the system can optimize feed ratios, drinking water systems and indoor ventilation conditions in real time based on the impact of high temperature and high humidity environments on animal production efficiency, reduce stress responses caused by high temperatures, and improve animals' feed conversion rates and daily weight gain. At the same time, through continuous monitoring of the environment and production performance, the system can promptly identify potential risks of climate change to animal health and proactively take adjustment measures.

[0061] See also Figure 3 ,The life cycle stage identification module includes a data collection submodule, a data screening submodule and a life cycle stage judgment submodule, among which:

[0062] Data collection submodule: Based on the animal's weight, growth rate, and activity pattern data, it regularly collects the animal's weight and activity frequency. It uses sensors to record the animal's weight and activity pattern data every hour, obtains animal-related activity information through monitoring equipment, records and saves the data hourly, and generates collected data.

[0063] Data screening submodule: Based on the collected data, the hidden Markov model is used. It is assumed that different stages of the animal life cycle can correspond to different hidden states. Each hidden state has a certain observation probability distribution. Through training on historical data, the state transition probability and the observation probability under each state of each life cycle stage are learned. The data that meets the conditions is screened out by comparing with the set standard range. The data that cannot pass the standard screening is regarded as irrelevant data and is eliminated to generate the screened data.

[0064] Life cycle stage determination submodule: Based on the screening data, the module compares the animal's weight change and growth rate with the set weight and growth rate thresholds hour by hour to check whether they meet the standards, identify whether the animal is in the correct life cycle stage, calibrate the specific standards for each stage, and generate the life cycle stage division results;

[0065] Data Collection Submodule: Based on animal weight, growth rate, and activity pattern data, sensors are used to regularly collect animal weight and activity frequency. The sensors record animal weight data every hour and obtain animal-related activity information through monitoring equipment. The collected data includes weight, activity frequency, and activity pattern. The monitoring equipment records the animal's dynamic behavior information at a high frequency, accurately obtaining hourly data points. Each collected data point is stored in real time in a storage unit and initially formatted. The configured collection program ensures that data is accurately collected within the set time interval. Each data collection will generate collected data, including the weight value, activity frequency value, and activity pattern at the corresponding time point within each time period.

[0066] Data screening submodule: Based on the collected data, the hidden Markov model is used to screen the collected animal weight, growth rate and activity data. The state space of the hidden Markov model is defined as different life cycle stages, each stage corresponds to a state, and the observation symbol is the collected data, namely weight change, activity frequency and growth rate. The hidden Markov model screens the collected data according to the preset life cycle stage standards, uses the Baum-Welch algorithm to train the model parameters, and performs probability estimation for the state selection at each moment, and uses the Viterbi algorithm to decode and obtain the most likely life cycle stage. By performing hourly comparative analysis of weight, growth rate and activity data, data that meets the set life cycle standards is screened out, and irrelevant noise data is removed. The generated screening data includes valid data sets and removes invalid data that does not meet the conditions to ensure the accuracy and validity of the data;

[0067] Life cycle stage determination submodule: Based on the filtered data, an hourly comparison algorithm is used to compare the filtered data with the set weight and growth rate thresholds. The thresholds used are standard values ​​set in advance according to different life cycle stages. The absolute error method is used to calculate the deviation between weight change and growth rate and the standard threshold. When the deviation exceeds the preset standard threshold, the data is marked as abnormal and further screened and classified to determine whether the data meets the standard range of the life cycle stage. The deviation value of all weight change and growth rate data from the threshold will be used to determine whether the animal is in the correct life cycle stage. Through hourly comparison and threshold verification of data, classification operations are performed to reasonably classify the data. The specific standards for each stage are calibrated based on the classification results, and the life cycle stage division results are finally generated. The life cycle stage of each animal is calibrated in detail to ensure the accuracy and rationality of the life cycle stage.

[0068] Hidden Markov model, according to the formula:

[0069]

[0070] Where: P(X t |X t-1 ,W) represents the observation data X at the previous time point at time t t-1 and weight coefficient W, calculate the probability of the state transferring to the current time point t, X t represents the observation data at time point t, including the animal's weight, growth rate and activity data, X t-1 represents the observed data at time point t-1, W represents the standard range of the life cycle stage, P(X t |S i ,W) represents a given hidden state S i and weight coefficient W when observing data X t The generation probability, S i represents the hidden state, representing the life cycle stage of the animal, P(S i |X t-1 ,W) represents the time from time point t-1 to time point t, given the observation data X t-1 and weight coefficient W, the hidden state S i The probability of transition, α i Represents the hidden state S i The weight coefficient of , N represents the number of hidden states;

[0071] Execution process: First, at each time point t, based on the observation data X at the previous time point t-1 and weight coefficient W to calculate the state transition probability P(X t |X t-1 ,W), ensure the validity of the data according to the set life cycle stage standard range, introduce the weight coefficient W, adjust the weight, growth rate and activity data at different stages of the screening criteria, optimize the accuracy of the data, and then calculate the given hidden state S i and weight coefficient W, the observed data X t The generation probability P(X t |S i ,W), ensuring that the growth data of animals at different life cycle stages meet the preset standards, and then calculating P(S i |X t-1 ,W), infer the hidden state S based on the observation data at time point t-1 and the weight coefficient W i The transition probability, that is, the conversion process of the life cycle stage, the final hidden state S i The weight coefficient α i Further adjust the data weights in the screening process, and ensure that important data within the life cycle has a greater influence in the overall data screening, effectively screen out data that meets the requirements at each life cycle stage, ensure the scientificity and accuracy of the data, and improve the performance of the in-situ measurement system in pastoral areas.

[0072] See also Figure 3 The production performance dynamic monitoring module includes a production data acquisition submodule, a data comparison submodule, and a production efficiency calculation submodule, among which:

[0073] Production data collection submodule: Based on the results of life cycle stage division, the weight, feed conversion rate and daily weight gain data of animals are collected item by item. The weight changes, feed consumption and daily weight gain of animals are recorded through real-time monitoring equipment to ensure the data is complete and timely, and generate and collect production data;

[0074] Data comparison submodule: Based on the collected production data, a convolutional neural network is used, with animal weight, feed conversion rate, and daily weight gain as input features to construct a time series data matrix. Local feature extraction is performed on the changing trends of these features in different time periods. After feature extraction, the feature dimension is reduced through a pooling layer, and then a fully connected layer is used to determine whether each indicator meets the standard range. Through the labeled historical production performance data, the characteristic patterns under normal and abnormal conditions are learned. Then, the deviation between the actual values ​​and the target values ​​of animal weight, feed conversion rate, and daily weight gain are compared item by item to identify whether the data deviates from the normal range, and the data that meets the standard is screened out to generate comparison results.

[0075] Production efficiency calculation submodule: Based on the comparison results, the production efficiency of each life cycle stage is calculated step by step. By updating the data, the production efficiency of each stage is evaluated, and each data is weighted and integrated to calculate the final production efficiency value and generate the production performance fluctuation identification result;

[0076] Production data collection submodule: Based on the results of life cycle stage division, the animal's weight, feed conversion rate and daily weight gain data are collected item by item. Real-time monitoring equipment is used to record changes in animal weight, feed consumption and daily weight gain. Weight data is collected hourly through sensors, feed consumption is recorded through metering devices, and daily weight gain is recorded synchronously with activity monitoring equipment. During data collection, the equipment executes the collection task at the set time interval, and the data is saved every hour to ensure data integrity and timeliness. The collected production data is generated, including the collected weight value, feed conversion rate, daily weight gain and timestamp information every hour;

[0077] Data comparison submodule: Based on the collected production data, a convolutional neural network is used to compare and analyze the data. The input layer of the convolutional neural network receives animal weight, feed conversion rate, and daily weight gain data. The convolution layer is used to perform convolution operations on the data, and the ReLU activation function is used to handle nonlinear relationships. The pooling layer is used to compress features and reduce data dimensions. The final fully connected layer is compared item by item with the preset standard range. The difference between the actual data and the target data is calculated through the mean square error. The model is optimized through the cross-entropy loss function to screen out data that meets the standards and generate comparison results, including the data set that meets the standards and excluding data points with large deviations.

[0078] Production efficiency calculation submodule: Based on the comparison results, the weighted average method is used to gradually calculate the production efficiency of each life cycle stage. First, the single production efficiency is calculated based on the comparison results of body weight, feed conversion rate and daily weight gain. The data are weighted by the weighting coefficient. The weight is set according to the contribution of each data to the production efficiency. The weighted sum calculation method is used to integrate the weighted data and finally calculate the production efficiency value. Through the data update process, the calculation results are adjusted in real time to generate the production performance fluctuation identification results, which include the production efficiency fluctuations in each life cycle stage for subsequent analysis.

[0079] Convolutional neural network, according to the formula:

[0080]

[0081] Where: z(t, W, β, b) represents the output result after convolution, which is used to reflect the matching between data features and target standards; x(τ) represents the input data under the time and position index τ, including the collected weight, growth rate and daily weight gain data; w(t-τ) represents the convolution kernel, which is used to extract data features; τ represents the time and position index during the convolution process, which is used to calculate the input data item by item; W represents the weight coefficient; β represents the adjustment coefficient, which is used to adjust the relative importance between weight, growth rate and daily weight gain; b represents the bias term, which is used to balance the convolution results;

[0082] Execution process: First, the input data x(τ) represents the weight, growth rate and daily weight gain data collected at different time points and locations. The convolution kernel w(t-τ) is used to perform weighted summation with the input data to complete the preliminary feature extraction. Then the convolution result z(t,W,β,b) is further adjusted by introducing the weight coefficient W, adjustment coefficient β and bias term b to reflect the different importance of data at different life cycle stages. The weight coefficient W is used to indicate the relative importance of data at different life cycle stages. The adjustment coefficient β controls the relative weights between weight, growth rate and daily weight gain. The bias term b ensures the balance of the convolution operation results and effectively screens the data at different life cycle stages of animals. The production performance of animals is accurately evaluated under stress-free conditions in the pastoral area to ensure that data results that meet the preset standards are generated.

[0083] See also Figure 3 The health status monitoring module includes a production data comparison submodule, a physiological data analysis submodule, and a health status confirmation submodule, among which:

[0084] Production data comparison submodule: Based on the results of production performance fluctuation identification, collect animal weight, daily weight gain and feed consumption data, and compare them with the set normal physiological range to determine whether the data exceeds the physiological standard range. By comparing and screening the data item by item, the production data comparison results are generated;

[0085] Physiological data analysis submodule: Based on the comparison results of production data, it monitors health indicators item by item, detects the deviation between health data and set standard thresholds, analyzes whether there are abnormal physiological fluctuations, and generates physiological status fluctuation results;

[0086] Health status confirmation submodule: Based on the physiological status fluctuation results, the deviation between health data and standard thresholds is compared, and health status confirmation operations are performed. By verifying whether the health status is abnormal, health trends are evaluated in combination with historical data, and health status monitoring results are generated;

[0087] Production data comparison submodule: Based on the results of production performance fluctuation identification, animal weight, daily weight gain and feed consumption data are collected, and the collected data are compared with the set normal physiological range using a one-by-one comparison algorithm. The difference between the data value and the standard threshold is compared item by item. The absolute error method is used to calculate the error of each data item, and the mean square error is used to evaluate the deviation between each data item and the set standard to determine whether the data exceeds the physiological standard range. By screening the data, the data that meets the standard is distinguished from the data that exceeds the standard range. The production data comparison results are generated, including the comparison of each data item and marking the data points with large deviations;

[0088] Physiological data analysis submodule: Based on the comparison results of production data, health indicators are monitored item by item. A dynamic threshold detection algorithm is used to compare the set standard threshold with the real-time collected health data. Kalman filtering is used to smooth the data to reduce noise interference. The sliding window technology is used to periodically update the historical data to ensure the monitoring accuracy of health indicators. Whether there are abnormal physiological fluctuations, the deviation value of each data is calculated and compared with the set health threshold to generate physiological status fluctuation results. The results indicate whether there are abnormal fluctuations for each health indicator and give the specific fluctuation value;

[0089] Health status confirmation submodule: Based on the results of physiological status fluctuations, compare the deviation between health data and standard thresholds, adopt status assessment algorithm, evaluate health trends through historical data comparison and analysis, use weighted moving average method to calculate the weighted average value of each health data, combine the deviation value and historical fluctuations to confirm the health status, perform health status confirmation operation, judge whether the current data exceeds the set threshold, and combine historical health data to evaluate whether the health status is abnormal, generate health status monitoring results, including health status monitoring information, indicate whether there are health abnormalities, and issue early warnings for abnormal trends.

[0090] Analyze whether abnormal physiological fluctuations occur. The definition of abnormal physiological fluctuations is based on the deviation between the health indicators in the animal's body and the preset normal physiological range. By regularly collecting the health indicator data of the animal and setting the standard range of each health indicator based on the normal health status of the animal population, the health indicators collected in real time are compared with the set standard thresholds. If an indicator deviates from the normal range and exceeds the set tolerance, it is determined to be an abnormal physiological fluctuation.

[0091] See also Figure 3 The environmental adaptability analysis module includes an environmental data acquisition submodule, a correlation analysis submodule, and a fitness calculation submodule, wherein:

[0092] Environmental data collection submodule: Based on the health status monitoring results, collects data related to the environment, records environmental changes through real-time monitoring equipment, and generates environmental data;

[0093] Correlation analysis submodule: Based on health status monitoring results and environmental data, it analyzes the correlation between environmental factors and health status item by item, and generates environmental factor correlation results by comparing the relationship between each environmental change and health fluctuation;

[0094] Fitness calculation submodule: Based on the correlation results of environmental factors, calculate environmental fitness, quantify the impact of environmental changes on animal health, and generate environmental adaptability analysis results;

[0095] Environmental data collection submodule: Based on health status monitoring results, it collects data related to the environment and uses real-time monitoring equipment to record environmental changes. Sensors collect environmental data every hour. The equipment automatically obtains environmental information within each time period through the configured collection program. The collected data points include timestamps, temperature values, humidity values, air pressure, and light intensity. The collected data points are saved and formatted in real time in the storage unit to generate environmental data, including the environmental information collected every hour, and provide the raw data basis for subsequent analysis.

[0096] Correlation Analysis Submodule: Based on health status monitoring results and environmental data, the Pearson correlation coefficient algorithm is used to analyze the correlation between environmental factors and health status. First, the units of health data and environmental data are standardized. Then, the correlation between each environmental factor and the corresponding health data is calculated. The linear correlation of each data is evaluated using the Pearson correlation coefficient. The calculated result is the correlation coefficient r, which represents the degree of correlation between the two. By comparing the relationship between each environmental change and health fluctuation item by item, the environmental factor correlation result is generated, which includes the specific correlation between each environmental data and health fluctuation.

[0097] Fitness calculation submodule: Based on the correlation results of environmental factors, the weighted fitness calculation method is used to calculate the environmental fitness. First, according to the correlation results between environmental factors and health status, a weight is assigned to each environmental factor. The weight is dynamically adjusted according to the Pearson correlation coefficient value. The higher the correlation coefficient, the greater the weight. The weighted sum calculation method is used to quantify the impact of environmental changes on animal health. By calculating the comprehensive score of environmental fitness, the environmental adaptability analysis results are generated, which include the specific fitness score of each environmental factor on animal health.

[0098] See also Figure 3 The risk warning and response module includes a data acquisition submodule, a warning generation submodule, and a warning processing submodule, among which:

[0099] Data collection submodule: Based on health status monitoring results and environmental adaptability analysis results, it collects climate change data, feeding adjustment data, and historical fluctuation data. It uses sensors to monitor environmental and production data in real time, records climate, humidity, temperature, and feeding adjustment information item by item, and generates collected environmental and production data.

[0100] Warning generation submodule: Based on the collected environmental and production data, by comparing the current health status with historical fluctuation data, combined with climate change and environmental data, it determines whether the data exceeds the set risk threshold, compares it with the set risk standards, and generates risk warning results;

[0101] Early warning processing submodule: Based on the risk early warning results, it checks whether there are any deviations from the standard threshold, performs early warning processing operations, marks the deviated data and takes processing measures to generate health and production risk early warning results;

[0102] Data collection submodule: Based on health status monitoring results and environmental adaptability analysis results, it collects climate change data, feeding adjustment data, and historical fluctuation data. Sensors are used to monitor environmental and production data in real time, and climate, humidity, temperature, and feeding adjustment information are recorded item by item. Sensors collect climate data and feeding adjustment data every hour, including feed type, feed amount, and changes in environmental parameters. All data is centrally stored in the data collection system to ensure data accuracy and real-time updates. Environmental and production data are generated, including information on each environmental and production variable, and provide basic data required for subsequent analysis;

[0103] Early warning generation submodule: Based on the collected environmental and production data, the current health status is compared and analyzed with historical fluctuation data through a comparison algorithm. A dynamic threshold judgment algorithm is used to set the normal fluctuation range of the health status based on the historical fluctuation data and compare it with the current data. The sliding window algorithm is used to calculate the difference between the current data and the historical data. In combination with climate change and environmental data, a risk threshold is set and the data is compared. When the data exceeds the set threshold range, a risk judgment operation is performed to generate a risk warning result, including whether each data item exceeds the preset risk threshold and identifies potential risk data;

[0104] Early warning processing submodule: Based on the risk warning results, a deviation detection algorithm is used to check whether there is any deviation from the standard threshold. First, the current data is compared with the historical fluctuation data, and the degree of data deviation is calculated through the Z-score standardization method. If the data exceeds the set standard threshold, the early warning processing operation is executed, and the deviated data is marked through the outlier marking algorithm. Further processing measures are taken, including adjusting the feeding plan and improving environmental conditions, to generate health and production risk warning results, including marked risk data and subsequent processing measures information.

[0105] See also Figure 3 The breeding strategy adjustment module includes a data analysis submodule, a strategy judgment submodule and an adjustment execution submodule, among which:

[0106] Data analysis submodule: Based on the health status monitoring results and health and production risk warning results, combined with the current production performance data, analyzes the health and production fluctuation data. By comparing the production efficiency and health status fluctuations item by item, the health and production fluctuation analysis results are generated;

[0107] Strategy judgment submodule: Based on the results of health and production fluctuation analysis, it determines whether the feeding management plan needs to be adjusted. By comparing the current feeding conditions and production status, it evaluates whether the feed ratio and feeding method should be adjusted, and generates the results of the strategy adjustment needs;

[0108] Adjustment execution submodule: Based on the results of strategy adjustment requirements, it executes feeding strategy adjustment operations, updates feeding management plans in real time by modifying feed ratios and adjusting animal feeding environments, and generates feeding strategy adjustment results;

[0109] Data Analysis Submodule: Based on health status monitoring results and health and production risk warning results, combined with current production performance data, a comparative analysis algorithm is used to analyze health and production fluctuation data. First, health status fluctuation and production efficiency data are converted to a unified scale through standardization. The Pearson correlation coefficient is used to calculate the correlation between health fluctuation and production efficiency. The difference between production efficiency and health status fluctuation is compared item by item. The significance of the fluctuation is determined through dynamic threshold comparison. The health and production fluctuation analysis results are generated, including the specific relationship between health and production fluctuations and the analyzed fluctuation amplitude.

[0110] Strategy judgment submodule: Based on the results of health and production fluctuation analysis, a multi-condition judgment algorithm is used to determine whether the feeding management plan needs to be adjusted. First, the current health status is compared with production efficiency fluctuations through condition screening. A weighted evaluation method is used to assess the impact of current feeding conditions on health and production status. Combined with historical fluctuation data, a decision tree algorithm is used to analyze whether to adjust feed ratios and feeding methods. The result of the strategy adjustment demand is generated, including whether feeding condition adjustments are needed and specific suggestions for adjustments.

[0111] Adjustment execution submodule: Based on the results of strategy adjustment requirements, the optimal adjustment algorithm is used to execute the feeding strategy adjustment operation. First, the optimal feed ratio is determined through the linear programming model. Combined with the results of health and production fluctuation analysis, the feed ratio is dynamically adjusted. The simulated annealing algorithm is used to optimize the adjustment plan of the animal feeding environment. The feeding management plan is updated in real time to generate the feeding strategy adjustment results, including the adjusted feed ratio, feeding environment changes and execution details.

[0112] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change and modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the scope of protection of the technical solution of the present invention.

Claims

1. The in-situ stress-free full life cycle production performance measurement system for pastoral areas is characterized by: The system comprises: Life cycle stage identification module: This module collects animal physiological data, compares weight changes and activity frequency using a hidden Markov model, and classifies the data based on weight and growth rate thresholds to generate life cycle stage classification results. Dynamic production performance monitoring module: Based on the life cycle stage division results, body weight, feed conversion rate and daily weight gain data are collected, and a convolutional neural network is used to compare the data with the standard range to determine whether the data exceeds the range and generate a production performance fluctuation identification result; the convolutional neural network is based on the formula: in: Represents the output result after convolution, which is used to reflect the matching between data features and target standards. Represents time and location index The input data under includes the collected weight, growth rate and daily weight gain data, Represents the convolution kernel, which is used to extract data features. Represents the time and position index in the convolution process, which is used to calculate the input data item by item. represents the weight coefficient, represents the adjustment factor, which is used to adjust the relative importance of body weight, growth rate and daily weight gain. Represents the bias term, which is used to balance the convolution results; Health status monitoring module: Based on the production performance fluctuation identification results, combined with physiological data, it is determined whether it exceeds the normal physiological range, and a health status confirmation operation is performed to generate a health status monitoring result; Environmental adaptability analysis module: collects environmental data based on the health status monitoring results, performs correlation analysis, and generates environmental adaptability analysis results; Risk warning and response module: Based on the health status monitoring results and environmental adaptability analysis results, combined with climate change and feeding adjustment data, risk thresholds are set to make risk judgments and generate health and production risk warning results; Breeding strategy adjustment module: Based on the health status monitoring results and health and production risk warning results, combined with production performance data, it determines whether to adjust the feeding plan, adjusts the strategy in real time, and generates breeding strategy adjustment results.

2. The pastoral area in-situ stress-free full life cycle production performance measurement system according to claim 1 is characterized in that: The life cycle stage identification module includes a data collection submodule, a data screening submodule and a life cycle stage judgment submodule, wherein: Data collection submodule: Based on the animal's weight, growth rate, and activity pattern data, it regularly collects the animal's weight and activity frequency. It uses sensors to record the animal's weight and activity pattern data every hour, obtains animal-related activity information through monitoring equipment, records and saves the data hourly, and generates collected data. Data screening submodule: Based on the collected data, a hidden Markov model is used, assuming that different stages of the animal life cycle correspond to different hidden states, each hidden state has a corresponding observation probability distribution, and through training on historical data, the state transition probability and the observation probability under each state of each life cycle stage are learned, and the data that meets the conditions is screened out by comparing with the set standard range. Data that fails to pass the standard screening is regarded as irrelevant data and is eliminated to generate screening data; Life cycle stage judgment submodule: Based on the screening data, by comparing with the set weight and growth rate thresholds hourly, check whether the animal's weight changes and growth rate meet the standards, identify whether the animal is in the correct life cycle stage, calibrate the specific standards for each stage, and generate life cycle stage division results.

3. The pastoral area in-situ stress-free full life cycle production performance measurement system according to claim 1 is characterized in that: The hidden Markov model is based on the formula: in: Indicates at a point in time , based on the observation data at the previous time point and weight coefficient , the calculation state is transferred to the current time point The probability of Indicates a time point Observational data on animals, including weight, growth rate and activity data, Indicates a time point The observation data on A standard range representing a life cycle phase, Represents a given hidden state and weight coefficient Time observation data The probability of generating represents the hidden state, representing the life cycle stage of the animal, Indicates from the time point At the time , given the observation data and weight coefficient In the case of The probability of transfer, Represents the hidden state The weight coefficient of Represents the number of hidden states.

4. The pastoral area in-situ stress-free full life cycle production performance measurement system according to claim 1 is characterized in that: The production performance dynamic monitoring module includes a production data acquisition submodule, a data comparison submodule and a production efficiency calculation submodule, wherein: Production data collection submodule: Based on the life cycle stage division results, the animal's weight, feed conversion rate and daily weight gain data are collected item by item. The animal's weight changes, feed consumption and daily weight gain are recorded through real-time monitoring equipment to ensure the data is complete and timely, and generate collected production data; Data comparison submodule: Based on the collected production data, a convolutional neural network is used, with animal weight, feed conversion rate and daily weight gain as input features, to construct a time series data matrix, and local feature extraction is performed on the changing trends of the features in different time periods. After feature extraction, the feature dimension is reduced through a pooling layer, and then a fully connected layer is used to determine whether each indicator meets the standard range. Through the labeled historical production performance data, the characteristic patterns under normal and abnormal conditions are learned, and then the deviations between the actual values ​​and target values ​​of animal weight, feed conversion rate and daily weight gain are compared item by item to identify whether the data deviates from the normal range, and the data that meets the standard is screened out to generate a comparison result; Production efficiency calculation submodule: Based on the comparison results, the production efficiency of each life cycle stage is calculated step by step. By updating the data, the production efficiency of each stage is evaluated, and each data is weighted and integrated to calculate the final production efficiency value and generate the production performance fluctuation identification result.

5. The pastoral area in-situ stress-free full life cycle production performance measurement system according to claim 1 is characterized in that: The health status monitoring module includes a production data comparison submodule, a physiological data analysis submodule and a health status confirmation submodule, wherein: Production data comparison submodule: Based on the production performance fluctuation identification results, collect animal weight, daily weight gain and feed consumption data, and compare them with the set normal physiological range to determine whether the data exceeds the physiological standard range. By comparing and screening the data item by item, the production data comparison results are generated; Physiological data analysis submodule: Based on the production data comparison results, monitor the health indicators item by item, detect the deviation between the health data and the set standard threshold, analyze whether there is abnormal physiological fluctuation, and generate physiological status fluctuation results; Health status confirmation submodule: Based on the physiological status fluctuation results, compare the deviation between the health data and the standard threshold, perform the health status confirmation operation, verify whether the health status is abnormal, combine historical data to evaluate the health trend, and generate health status monitoring results.

6. The pastoral area in-situ stress-free full life cycle production performance measurement system according to claim 5 is characterized in that: The analysis determines whether abnormal physiological fluctuations occur. The definition of abnormal physiological fluctuations is based on the deviation between the health indicators in the animal's body and the preset normal physiological range. The health indicator data of the animal is collected regularly, and the standard range of each health indicator is set according to the normal health status of the animal population. At the same time, the health indicators collected in real time are compared with the set standard thresholds. If an indicator deviates from the normal range and exceeds the set tolerance, it is determined to be an abnormal physiological fluctuation.

7. The pastoral area in-situ stress-free full life cycle production performance measurement system according to claim 1 is characterized in that: The environmental adaptability analysis module includes an environmental data acquisition submodule, a correlation analysis submodule, and a fitness calculation submodule, wherein: Environmental data collection submodule: Based on the health status monitoring results, collect data related to the environment, record environmental changes through real-time monitoring equipment, and generate environmental data; Correlation analysis submodule: Based on the health status monitoring results and environmental data, the correlation between environmental factors and health status is analyzed item by item, and the correlation results of environmental factors are generated by comparing the relationship between each environmental change and health fluctuation; Fitness calculation submodule: Based on the environmental factor association results, calculate the environmental fitness, quantify the impact of environmental changes on animal health, and generate environmental adaptability analysis results.

8. The pastoral area in-situ stress-free full life cycle production performance measurement system according to claim 1 is characterized in that: The risk warning and response module includes a data acquisition submodule, a warning generation submodule, and a warning processing submodule, wherein: Data collection submodule: Based on the health status monitoring results and environmental adaptability analysis results, collect climate change data, feeding adjustment data and historical fluctuation data, monitor the environment and production data in real time through sensors, record climate, humidity, temperature and feeding adjustment information item by item, and generate collected environment and production data; Warning generation submodule: Based on the collected environmental and production data, by comparing the current health status with historical fluctuation data, combined with climate change and environmental data, it is determined whether the animal's health status data, environmental data and production data exceed the set risk threshold, and compared with the set risk standards to generate risk warning results; Early warning processing submodule: Based on the risk early warning results, check whether there is any deviation from the standard threshold, perform early warning processing operations, mark the deviated data and take processing measures to generate health and production risk early warning results.

9. The pastoral area in-situ stress-free full life cycle production performance measurement system according to claim 1, characterized in that: The breeding strategy adjustment module includes a data analysis submodule, a strategy judgment submodule and an adjustment execution submodule, wherein: Data analysis submodule: Based on the health status monitoring results and health and production risk warning results, combined with current production performance data, analyzes health and production fluctuation data, and generates health and production fluctuation analysis results by comparing production efficiency and health status fluctuations item by item; Strategy judgment submodule: Based on the health and production fluctuation analysis results, it determines whether the feeding management plan needs to be adjusted. By comparing the current feeding conditions and production status, it evaluates whether the feed ratio and feeding method should be adjusted, and generates the strategy adjustment demand results; Adjustment execution submodule: Based on the strategy adjustment demand results, execute the feeding strategy adjustment operation, update the feeding management plan in real time by modifying the feed ratio and adjusting the animal feeding environment, and generate the feeding strategy adjustment results.

Images