Beef cattle individual information management system

Abstract

The invention relates to the technical field of beef cattle information management, in particular to a beef cattle individual information management system. According to the technical scheme, the system comprises a data acquisition layer, a data transmission layer, a data processing and storage layer and an application layer, wherein the data acquisition layer is used for acquiring biological characteristics, physiological indexes, behavior information and breeding environment information of beef cattle; the data transmission layer is responsible for transmitting the data acquired by the data acquisition layer to the data processing and storage layer; the data processing and storage layer is used for cleaning, fusing, analyzing and storing the data transmitted by the data transmission layer; and the application layer is based on the processing result of the data processing and storage layer. Through accurate and comprehensive information acquisition, strong data processing capability, abundant intelligent application functions and guarantee of data safety and stability, fine management and scientific decision of beef cattle breeding are realized, breeding benefits and market competitiveness are improved, and development of the beef cattle breeding industry is promoted.

Description

Technical Field

[0001] The present invention relates to the technical field of beef cattle information management, and particularly to a beef cattle individual information management system. Background Art

[0002] The beef cattle breeding management technology has experienced the development from traditional to modern. Traditional breeding management mainly relies on manual experience. Information collection is obtained by visual observation, handheld instrument measurement and simple tools, processing depends on manual records and a small amount of data analysis, and decision-making relies on past experience. With the progress of information technology, modern management systems have emerged, which automatically collect some physiological and environmental indicators by sensors, transmit data using communication technology, and have certain storage and simple analysis functions.

[0003] The existing technologies have many deficiencies in aspects of information collection, processing and transmission, and application functions. For example, the collection range is limited, the accuracy and real-time performance are poor, it is difficult to comprehensively and accurately obtain the biological characteristics, behavior rules and health conditions of beef cattle; the processing ability is weak, lacking in-depth integration and mining, unable to provide a scientific decision-making basis, and the transmission security and stability are insufficient; the application functions are single, lacking intelligent early warning and personalized services, unable to meet the needs of refined breeding, resulting in the ineffective improvement of breeding efficiency and low market competitiveness.

[0004] In view of the above deficiencies, the present application proposes a beef cattle individual information management system to solve the above problems existing in the prior art. The information collection of this system is comprehensive and accurate, covering biological characteristics, physiology, behavior and environmental information, and advanced equipment is used to ensure accurate and real-time data; it has a powerful data processing and fusion ability, and scientific algorithms are used for cleaning, fusion and analysis to provide a basis for decision-making; the application functions are rich and intelligent, including breeding management, early warning and decision support, which can achieve refined management and scientific decision-making; at the same time, it ensures the security and stability of data transmission, adopts encryption and backup mechanisms to ensure the reliable operation of the system; it realizes the needs of refined breeding, effectively improves breeding efficiency and market competitiveness. Summary of the Invention

[0005] The purpose of the present invention is to propose a beef cattle individual information management system that can achieve refined management and scientific decision-making in view of the problems existing in the background art.

[0006] The present application provides a beef cattle individual information management system, which includes the following four levels:

[0007] Data collection layer: used for collecting the biological characteristics, physiological indicators, behavior information and breeding environment information of beef cattle;

[0008] Data transmission layer: responsible for transmitting the data collected by the data collection layer to the data processing and storage layer;

[0009] Data processing and storage layer: Cleans, fuses, analyzes, and stores the data transmitted by the data transmission layer;

[0010] Application layer: Based on the processing results of the data processing and storage layer, provides functions of breeding management, intelligent warning, and decision-making support for beef cattle breeding.

[0011] Optionally, the data acquisition layer includes a biometric acquisition module, including:

[0012] Gene information acquisition unit: Uses a gene sequencer to perform whole-genome sequencing on beef cattle blood or tissue samples after DNA extraction and library construction; the sequencing read length is 2×250bp, and the gene data obtained by sequencing is stored in FASTQ format;

[0013] Appearance feature acquisition unit: Consists of multiple high-definition 3D cameras, which are distributed and installed at different positions in the cowshed; automatically takes images and videos of beef cattle at preset time intervals, records the relevant information of the shooting time and position, and simultaneously performs preprocessing operations of denoising and enhancement on the collected images and videos.

[0014] Optionally, the data acquisition layer includes a physiological index acquisition module, including:

[0015] Body temperature sensor unit: Uses a wearable digital temperature sensor with a sampling period of 1 minute; the digital temperature sensor sends the measured data to the data aggregation node through the built-in ZigBee wireless communication module;

[0016] Heart rate sensor unit: Selects a heart rate sensor and wears it on the beef cattle; the measurement range is 25-220 beats per minute, and the sampling frequency is 1 time per second; the heart rate sensor transmits the collected heart rate data to the data acquisition terminal through Bluetooth, and the data is preliminarily processed at the data acquisition terminal to remove the outliers;

[0017] Respiratory rate sensor unit: Consists of a humidity sensor combined with a respiratory induction sensor; the measurement range is 10-60 times per minute, and the sampling period is 5 seconds; transmits the collected humidity data to the data processing module to analyze and calculate the respiratory rate of the beef cattle.

[0018] Optionally, the data acquisition layer includes a behavior information acquisition module, including:

[0019] Movement trajectory tracking unit: Installs a GPS positioning tracker on the beef cattle collar; collects the position information of the beef cattle at a preset time interval of 5 minutes, uploads the collected position information data to the cloud server through the 4G network, analyzes the collected movement trajectory data, and calculates the parameters of the activity range and movement speed of the beef cattle;

[0020] Feeding and Drinking Monitoring Unit: It includes an intelligent feeding trough unit and an intelligent water trough unit. The intelligent feeding trough is equipped with a weighing sensor and an RFID identification module inside, which can monitor the weight change of the feed in the trough in real time. At the same time, it can identify the electronic ear tag of the beef cattle through the RFID identification module, and record the feeding time and feed intake of the beef cattle. The intelligent water trough is equipped with a flow sensor and a liquid level sensor, which are used to measure the drinking water volume of the beef cattle and monitor the water level in the trough in real time. When the water level is lower than the set value, the water filling device will be automatically triggered. The feeding and drinking data are processed to analyze the feeding and drinking patterns of the beef cattle.

[0021] Optionally, the data acquisition layer includes an environmental information acquisition module, including:

[0022] Temperature and Humidity Sensor Unit: Select a temperature and humidity sensor and install it at different positions in the cattle shed. The sampling period is 10 minutes. The temperature and humidity sensor transmits the collected temperature and humidity data to the data aggregation node through LoRa wireless communication technology, and then the aggregation node uploads the data to the cloud server.

[0023] Air Quality Sensor Unit: It includes an ammonia sensor and a hydrogen sulfide sensor, which are used to monitor the concentrations of ammonia and hydrogen sulfide in the cattle shed in real time. When the ammonia concentration exceeds 20 ppm and the hydrogen sulfide concentration exceeds 5 ppm, the system will automatically send out a warning message.

[0024] Optionally, the data transmission layer uses RaspberryPi4B as the data aggregation node, which is responsible for collecting the data collected by each sensor, and performing preliminary processing and storage on these data. It supports ZigBee, LoRa, Bluetooth and 4G / 5G wireless communication protocols, and will select a suitable communication protocol according to the type and transmission distance of the sensor. During the data transmission process, the AES encryption algorithm is used to ensure the security of data transmission, and at the same time, a backup mechanism is established to prevent data loss.

[0025] Optionally, the data cleaning module in the data processing and storage layer processes the data using the following methods:

[0026] Outlier Detection: The Z-score method based on statistical analysis is used to detect outliers. The calculation formula is:

[0027]

[0028] Among them, Z i represents the value of the data point, x i represents the i-th data point, μ is the mean value of the data of this index, σ is the standard deviation. When |Z i |>3, it is determined that this data point is an outlier.

[0029] Missing Value Processing: The linear interpolation method is used to process missing values. For the missing data x j, calculate based on the adjacent data x before and after j-1 and x j+1 The formula is:

[0030]

[0031] where x j represents the missing data of the j-th data point, j + 1 represents the previous data point of the j-th data point, and j - 1 represents the next data point of the j-th data point;

[0032] Data standardization: Use the Min-Max standardization method to standardize the cleaned data. The calculation formula is:

[0033]

[0034] where x is the original data, min(x) and max(x) are the minimum and maximum values of this group of data respectively, and x' is the standardized data.

[0035] Optionally, the data fusion module in the data processing and storage layer uses the Kalman filter algorithm to fuse multi-source data, specifically as follows:

[0036] The system state equation is: X k+1 = A k X k + B k U k + W k ;

[0037] where X k+1 represents the system state vector at time k + 1, A k is the state transition matrix, X k is the system state vector at time k, B k is the control input matrix, U k is the control input vector at time k, and W k is the process noise vector;

[0038] The observation equation is: Z k = H k X k + V k ;

[0039] where Z k is the observation vector at time k, H k is the observation matrix, and V k is the observation noise vector;

[0040] The Kalman filter steps include:

[0041] Prediction:

[0042] This formula is used to estimate the state at time k-1 to predict the state estimate at time k for predicting the covariance matrix P at time k k|k-1 , where P k-1|k-1 is the covariance matrix at time k-1, and Q k is the process noise covariance matrix;

[0043] Update: Calculate the Kalman gain K k , which is used to balance the weights of the predicted value and the observed value; Update the state estimate at time k according to the Kalman gain Update the covariance matrix P at time k k|k , where I is the identity matrix and R k is the observation noise covariance matrix.

[0044] Optionally, the data analysis module in the data processing and storage layer includes:

[0045] Health assessment unit: Adopt the support vector machine algorithm for health assessment. For the linearly separable case, the optimization objective is:

[0046] The constraint condition is s.t. y i (w T x i +b)≥1, i = 1, 2,..., n; where w is the weight vector, b is the bias, and x i is the input sample, y i is the sample label (y i ∈{-1, +1}), and n is the number of samples; Use historical data to train the SVM model and select the optimal model parameters by using the cross-validation method;

[0047] Growth prediction unit: Use the long short-term memory network algorithm to predict the growth of beef cattle. The core formula of the LSTM unit includes:

[0048] Input gate: i t =σ(W ii x t +W hi h t-1 +b i ), where i t is the output of the input gate at time t, σ is the sigmoid function, W ii is the weight matrix input to the input gate, x t is the input at time t, W hi is the weight matrix from the previous hidden state to the input gate, and ht-1 is the hidden state at time t-1, b i is the bias vector of the input gate, and the input gate determines the amount of information input that needs to be updated into the cell state at the current time;

[0049] Forget gate: f t = σ(W if x t + W hf h t-1 + b f ), f t is the output of the forget gate at time t, W if is the weight matrix input to the forget gate, W hf is the weight matrix from the hidden state of the previous time to the forget gate, b f is the bias vector of the forget gate, and the forget gate determines the amount of information in the cell state of the previous time that needs to be forgotten;

[0050] Cell state update: is the candidate cell state at time t, tanh is the hyperbolic tangent function, W ic is the weight matrix input to the candidate cell state, W hc is the weight matrix from the hidden state of the previous time to the candidate cell state, b c is the bias vector of the candidate cell state; C t is the cell state at time t, ⊙ is element-wise multiplication, and this formula updates the cell state according to the outputs of the forget gate and the input gate;

[0051] Output gate: o t = σ(W io x t + W ho h t-1 + b o ), o t is the output of the output gate at time t, W io is the weight matrix input to the output gate, W ho is the weight matrix from the hidden state of the previous time to the output gate, b o is the bias vector of the output gate;

[0052] Hidden state: h t = o t ⊙ tanh(C t ), h t is the hidden state at time t, and the output gate determines the amount of information in the cell state that needs to be output to the hidden state; The LSTM model is trained using the historical growth data of beef cattle, and the future growth prediction results can be output by inputting the current growth data.

[0053] Optionally, the data storage module in the data processing and storage layer includes a distributed file system (Hadoop HDFS) and a relational database (MySQL); among them, the distributed file system is used to store original sensor data, images, and large amounts of video-related data files; the relational database is used to store the processed and analyzed data, which includes the basic information of beef cattle, health assessment results, and growth prediction data.

[0054] Compared with the prior art, the present application includes at least one of the following beneficial technical effects:

[0055] 1. Precise and comprehensive information collection: It comprehensively collects the biological characteristics, physiological indicators, behavior information, and breeding environment information of beef cattle. Advanced equipment is used to ensure accurate and real-time data, providing a solid foundation for subsequent management. For example, a gene sequencer is used to obtain the whole-genome information for seed selection and breeding, and a high-precision camera captures the subtle changes in appearance.

[0056] 2. Powerful data processing ability: It has the ability to clean, fuse, analyze, and store data. Scientific algorithms are used to process outliers and fuse multi-source data, and professional algorithms are used to evaluate and predict the health and growth of beef cattle, providing a scientific basis for decision-making.

[0057] 3. Rich and intelligent application functions: It provides functions such as breeding management, intelligent warning, and decision support, realizing refined management of the whole process of beef cattle breeding. It can monitor abnormalities in real time and give warnings, and provide suggestions for slaughter and adjustment of breeding strategies in combination with the market and costs.

[0058] 4. Ensure data security and stability: Multiple wireless communication protocols are adopted, and encryption algorithms are used to ensure transmission security. A backup mechanism is established to prevent data loss, ensuring the reliable operation of the system. Brief Description of the Drawings

[0059] Figure 1 It is a principle block diagram of a beef cattle individual information management system. Detailed Embodiments

[0060] The technical solutions of the present invention will be further described below in conjunction with the drawings and specific embodiments.

[0061] Embodiment

[0062] As Figure 1 shown, a beef cattle individual information management system proposed by the present invention includes the following four levels:

[0063] Data acquisition layer: It is used to collect the biological characteristics, physiological indicators, behavior information, and breeding environment information of beef cattle;

[0064] Data transmission layer: It is responsible for transmitting the data collected by the data acquisition layer to the data processing and storage layer;

[0065] Data processing and storage layer: Cleans, fuses, analyzes, and stores the data transmitted by the data transmission layer;

[0066] Application layer: Based on the processing results of the data processing and storage layer, provides functions of breeding management, intelligent early warning, and decision-making support for beef cattle breeding.

[0067] Furthermore, the data acquisition layer includes a biometric acquisition module, a physiological index acquisition module, a behavior information acquisition module, and an environmental information acquisition module:

[0068] The biometric acquisition module includes:

[0069] Gene information acquisition unit: Uses a gene sequencer of model Illumina NovaSeq 6000. By collecting beef cattle blood or tissue samples, after DNA extraction and library construction, whole-genome sequencing is carried out; the sequencing read length is 2×250bp, the accuracy rate is above 99.9%, and the gene data obtained by sequencing is stored in FASTQ format (a text-based file format for storing biological sequences and corresponding base or amino acid quality).

[0070] Appearance feature acquisition unit: Consists of multiple high-definition 3D cameras of model Basler acA4096-30um 4K, which are distributed and installed at different positions in the cowshed; the camera frame rate is 30fps, the resolution is 4096×3000 pixels, and it has depth perception ability; automatically takes images and videos of beef cattle at preset time intervals, records relevant information such as the shooting time and location, and at the same time performs preprocessing operations of denoising and enhancement on the collected images and videos.

[0071] The physiological index acquisition module includes:

[0072] Body temperature sensor unit: Uses a wearable digital temperature sensor of model DS18B20, which is fixed on the ear of beef cattle through a silicone patch; its measurement range is -55°C - 125°C, the accuracy is ±0.1°C, and the sampling period is 1 minute; this sensor sends the measurement data to the data aggregation node (a small wireless base station or gateway, which is responsible for collecting data from multiple sensor nodes and sending the data to the Internet or other network devices) through the built-in ZigBee wireless communication module, and is calibrated regularly to ensure the accuracy of the measurement.

[0073] Heart rate sensor unit: A Polar H10 heart rate sensor is selected and worn on the beef cattle in a chest strap manner. The measurement range is 25 - 220 beats per minute, the accuracy is ±1 beat per minute, and the sampling frequency is 1 time per second. The sensor transmits the collected heart rate data to the data acquisition terminal via Bluetooth. The data is preliminarily processed at the data acquisition terminal to remove outliers, and the data is verified by taking the average of multiple measurements.

[0074] Respiratory rate sensor unit: It consists of a humidity sensor of model Honeywell HIH - 4000 combined with a respiratory induction sensor. The measurement range is 10 - 60 breaths per minute, the accuracy is ±1 breath per minute, and the sampling period is 5 seconds. The collected humidity data is transmitted to the data processing module to analyze and calculate the respiratory rate of the beef cattle.

[0075] It should be noted that the above - mentioned respiratory induction sensor is composed of an induction main body, a humidity conduction component, and a fixing component, which are specifically as follows:

[0076] Induction main body: It is made of a highly sensitive flexible material that can closely fit the chest or abdomen of the beef cattle to accurately capture the subtle changes generated during the respiration of the beef cattle. This flexible material not only ensures sensitive perception of the respiration of the beef cattle but also does not cause restraint or discomfort to the beef cattle.

[0077] Humidity conduction component: Connects the induction main body to the Honeywell HIH - 4000 humidity sensor and is responsible for transmitting the humidity changes caused by respiration sensed to the humidity sensor. The conduction component uses a moisture - conducting material with good water absorption and moisture conductivity, which can quickly and accurately transmit the humidity information to the sensor to ensure the timeliness and accuracy of data collection.

[0078] Fixing component: Includes adjustable straps or clips for firmly fixing the respiratory induction sensor on the beef cattle. The design of the fixing component fully considers the movement characteristics of the beef cattle, ensuring that the device does not fall off during the movement of the beef cattle and does not hinder the normal movement of the beef cattle.

[0079] When installing the respiratory induction sensor, first select the appropriate adjustment method of the fixing component according to the body size of the beef cattle, and fix the respiratory induction sensor on the chest or abdomen of the beef cattle through straps or clips to ensure that the induction main body closely fits the body of the beef cattle, avoiding looseness and gaps, so as not to affect the induction effect. Correctly connect the humidity conduction component to the humidity sensor to ensure a stable connection and prevent data transmission interruption.

[0080] In the data acquisition stage, when the beef cattle breathe, the humidity change generated by breathing is captured by the sensing body, and then transmitted to the humidity sensor through the humidity conduction component. The humidity sensor collects humidity data according to a sampling period of 5 seconds and transmits this data to the data processing module.

[0081] The behavior information acquisition module includes:

[0082] The movement trajectory tracking unit: A GPS positioning tracker with the model of Quectel L76-G is installed on the beef cattle collar; the position information of the beef cattle is collected at a preset time interval of 5 minutes, and this data is uploaded to the cloud server through the 4G network. The collected movement trajectory data is analyzed to calculate the parameters of the activity range and movement speed of the beef cattle.

[0083] The feeding and drinking monitoring unit: It includes an intelligent feeding trough unit and an intelligent water trough unit. The intelligent feeding trough is equipped with a weighing sensor with the model of HX711 (accuracy of ±0.1 kg) and an RFID identification module, which can monitor the weight change of the feed in the feeding trough in real time. At the same time, the electronic ear tag of the beef cattle is identified through the RFID identification module to record the feeding time and feed intake of the beef cattle; the intelligent water trough is equipped with a flow sensor with the model of YF-S201 (accuracy of ±0.01 L) and a liquid level sensor with the model of E3F-DS30C4, which can accurately measure the drinking water volume of the beef cattle and monitor the water level in the water trough in real time. When the water level is lower than the set value, the water adding device is automatically triggered; the feeding and drinking data is processed to analyze the feeding and drinking rules of the beef cattle.

[0084] The environmental information acquisition module includes:

[0085] The temperature and humidity sensor unit: A temperature and humidity sensor with the model of Sensirion SHT31 is selected and installed at different positions in the cowshed; the temperature measurement range is -40°C - 125°C, the accuracy is ±0.3°C, the humidity measurement range is 0% - 100% RH, the accuracy is ±3% RH, and the sampling period is 10 minutes; the sensor transmits the collected temperature and humidity data to the data aggregation node through LoRa wireless communication technology, and then the aggregation node uploads it to the cloud server, and these temperature and humidity data will be displayed in the form of a chart on the system interface in real time.

[0086] The air quality sensor unit: It includes an ammonia sensor with the model of MQ-137 (measurement range of 0 - 100 ppm, accuracy of ±1 ppm) and a hydrogen sulfide sensor with the model of MQ-136 (measurement range of 0 - 50 ppm, accuracy of ±0.5 ppm), which are used to monitor the concentrations of ammonia and hydrogen sulfide in the cowshed in real time; when the ammonia concentration exceeds 20 ppm and the hydrogen sulfide concentration exceeds 5 ppm, the system will automatically send out a warning message.

[0087] Secondly, the data transmission layer uses a Raspberry Pi 4B (single-board computer) as the data aggregation node, which is responsible for collecting the data collected by each sensor, and performing preliminary processing and storage on this data; this layer supports ZigBee, LoRa, Bluetooth, and 4G wireless communication protocols, and will select the appropriate communication protocol according to the type of sensor and the transmission distance; during the data transmission process, the AES encryption algorithm is used to ensure the security of data transmission, and at the same time, a backup mechanism is established to prevent data loss.

[0088] Furthermore, for the data processing and storage layer, the specific operations of cleaning, fusing, analyzing, and storing the data transmitted by the data transmission layer are as follows:

[0089] The first step is to process the data in the data cleaning module of the data processing and storage layer using the following method:

[0090] Outlier detection: The Z-score method based on statistical analysis (also known as the standard score, which is a method for measuring the distance of a data point from the mean of the data set to which it belongs) is used to detect outliers. The calculation formula is:

[0091]

[0092] Among them, Z i represents the value of the data point, x i represents the i-th data point, μ is the mean of the data of this index, σ is the standard deviation, and when |Z i | > 3, it is determined that this data point is an outlier;

[0093] Missing value processing: The linear interpolation method is used to process missing values. For the missing data x j , according to the adjacent data x j-1 and x j+1 before and after, the calculation is performed, and the formula is:

[0094]

[0095] Among them, x j represents the missing data of the j-th data point, j + 1 represents the data point before the j-th data point, and j - 1 represents the data point after the j-th data point;

[0096] Data standardization: The Min-Max standardization method is used to standardize the cleaned data. The calculation formula is:

[0097]

[0098] Among them, x is the original data, min(x) and max(x) are the minimum and maximum values of this group of data respectively, and x' is the standardized data;

[0099] The second step is to use the Kalman filter algorithm to fuse multi-source data in the data fusion module of the data processing and storage layer, as follows:

[0100] The system state equation is: k+1 =A k X k +B k U k +W k ;

[0101] Among them, X k+1 represents the system state vector at time k+1, A k is the state transfer matrix, X k is the system state vector at time k, B k is the control input matrix, U k is the control input vector at time k, W k is the process noise vector;

[0102] The observation equation is: Z k =H k X k +v k ;

[0103] Among them, Z k is the observation vector at time k, H k is the observation matrix, V k is the observation noise vector;

[0104] The Kalman filter steps include:

[0105] predict:

[0106] This formula is used to estimate the value based on the state at time k-1. Predict the state estimate at time k The covariance matrix P used to predict time k k|k-1 , where P k-1|k-1 is the covariance matrix at time k-1, Q k is the process noise covariance matrix;

[0107] renew: Calculate the Kalman gain K k , used to weigh the weights of predicted values and observed values; Update the state estimate at time k according to the Kalman gain Update the covariance matrix P at time k k|k , where I is the identity matrix, R k is the observation noise covariance matrix;

[0108] In addition, the data fusion module will fuse the data by using a weighted fusion method according to the characteristics and importance of different types of data;

[0109] In the third step, when analyzing and processing the data in the data processing and storage layer, the data analysis module includes:

[0110] Health assessment unit: The support vector machine (SVM) algorithm is used for health assessment. For the linearly separable case, the optimization objective is:

[0111] The constraint condition is s.t. y i (w T x i +b)≥1, i = 1, 2, …, n; where w is the weight vector, b is the bias, x i is the input sample, y i is the sample label (y i ∈{-1, +1}), n is the number of samples; The SVM model is trained using historical data, and the optimal model parameters are selected by using the cross-validation method;

[0112] Growth prediction unit: The long short-term memory network (LSTM) algorithm is used to predict the growth of beef cattle. The core formulas of the LSTM unit include:

[0113] Input gate: i t =σ(W ii x i +W hi h t-1 +b i ), where i t is the output of the input gate at time t, σ is the sigmoid function, W ii is the weight matrix input to the input gate, x t is the input at time t, W hi is the weight matrix from the previous hidden state to the input gate, h t-1 is the hidden state at time t-1, b i is the bias vector of the input gate, and the input gate determines the amount of information input that needs to be updated into the cell state;

[0114] Forget gate: f t =σ(W if x t +W hf h t-1 +b f ), f t is the output of the forget gate at time t, W if is the weight matrix input to the forget gate, W hf is the weight matrix from the previous hidden state to the forget gate, bf is the bias vector of the forget gate, and the forget gate determines the amount of information in the cell state at the previous moment that needs to be forgotten;

[0115] Cell state update: is the candidate cell state at time t, tanh is the hyperbolic tangent function, W ic is the weight matrix input to the candidate cell state, W hc is the weight matrix from the hidden state at the previous moment to the candidate cell state, b c is the bias vector of the candidate cell state; C t is the cell state at time t, ⊙ is element-wise multiplication, and this formula updates the cell state according to the outputs of the forget gate and the input gate;

[0116] Output gate: o t = σ(W io x t + W ho h t-1 + b o ), o t is the output of the output gate at time t, W io is the weight matrix input to the output gate, W ho is the weight matrix from the hidden state at the previous moment to the output gate, b o is the bias vector of the output gate;

[0117] Hidden state: h t = o t ⊙ tanh(C t ), h t is the hidden state at time t, and the output gate determines the amount of information in the cell state that needs to be output to the hidden state; The LSTM model is trained using the historical growth data of beef cattle, and the future growth prediction results can be output by inputting the current growth data.

[0118] Secondly, the data storage module in the data processing and storage layer includes a distributed file system and a relational database; among them, the distributed file system is used to store original sensor data, images, and large amounts of video-related files; the relational database is used to store the processed and analyzed data, and the processed and analyzed data includes the basic information of beef cattle, health assessment results, and growth prediction data.

[0119] It is worth mentioning that the processing results based on the data processing and storage layer in the application layer provide breeding management, intelligent early warning and decision support functions for beef cattle breeding, specifically: understanding the basic information of beef cattle, mastering the eating habits and food intake of beef cattle, evaluating the health of beef cattle, predicting the growth of beef cattle, and issuing early warnings for abnormal data. By providing assistance to human decision-making through data, corresponding adjustments can be made to beef cattle breeding in a timely manner.

[0120] In this embodiment, the data collection layer covers the full range of beef cattle biological characteristics, physiological indicators, behavioral information and breeding environment information. The gene information collection unit uses the advanced Illumina NovaSeq6000 gene sequencer, which can accurately obtain the whole genome information of beef cattle and provide a scientific basis for breeding selection; the appearance feature collection unit uses multiple high-precision 4K high-definition 3D cameras to accurately capture the subtle changes in the appearance of beef cattle. The physiological indicator, behavioral information and environmental information collection module also uses high-precision and high-stability sensors and equipment to ensure the accuracy and real-time nature of the collected data.

[0121] The data processing and storage layer has advanced data cleaning, fusion, analysis and storage capabilities. The data cleaning unit uses scientific algorithms to detect and process outliers and missing values to ensure data quality; the data fusion unit uses the Kalman filter algorithm and weighted fusion method to effectively fuse multi-source data and explore the intrinsic connections between data; the data analysis unit uses the support vector machine (SVM) algorithm and the long short-term memory network (LSTM) algorithm to accurately evaluate and predict the health and growth of beef cattle, providing a scientific basis for breeding decisions.

[0122] In the data transmission layer, this system uses RaspberryPi4B as a data aggregation node, supports multiple wireless communication protocols, and can select the optimal protocol according to the sensor type and transmission distance. At the same time, the AES encryption algorithm is used to ensure data transmission security, and a backup mechanism is established to prevent data loss, ensuring the stability of system operation and data reliability.

[0123] The application layer provides the functions of breeding management, intelligent early warning and decision support. The breeding management module includes file management, feed management and breeding management units, which can realize the refined management of the whole process of beef cattle breeding. The intelligent early warning module can monitor the health of beef cattle and the breeding environment in real time, and issue early warnings and provide control suggestions when abnormalities occur. The decision support module provides scientific suggestions for market release decisions and breeding strategy adjustments based on data analysis results, combined with market price trends and breeding costs, to improve breeding efficiency and market competitiveness.

[0124] The above specific embodiments are merely several alternative embodiments of the present invention. Based on the technical solution of the present invention and the relevant revelations of the above embodiments, those skilled in the art can make various alternative improvements and combinations to the above specific embodiments.

Claims

1. A beef cattle individual information management system, characterized in that include: The data collection layer collects the biological characteristics, physiological indicators, behavioral information and breeding environment information of beef cattle; The data transmission layer transmits the data collected by the data collection layer to the data processing and storage layer; The data processing and storage layer cleans, integrates, analyzes, and stores the data transmitted from the data transmission layer; The application layer provides breeding management, intelligent early warning and decision support functions for beef cattle breeding based on the processing results of the data processing and storage layer.

2. The beef cattle individual information management system according to claim 1, characterized in that, The data collection layer includes a biometric feature collection module, and the biometric feature collection module includes: The gene information collection unit uses a gene sequencer to collect blood or tissue samples from beef cattle, extract DNA, and construct a library before performing whole genome sequencing; The appearance feature acquisition unit is composed of multiple high-definition 3D cameras installed in different locations in the cowshed. It automatically captures images and videos of beef cattle at preset time intervals, and records relevant information about the shooting time and location. At the same time, it performs denoising and enhancement pre-processing operations on the captured images and videos.

3. The beef cattle individual information management system according to claim 1, characterized in that, The data collection layer includes a physiological index collection module, including: Body temperature sensor unit: a wearable digital temperature sensor is used, and the digital temperature sensor sends the measurement data to the data aggregation node through a built-in wireless communication module; A heart rate sensor unit is selected and worn on the beef cattle. The heart rate sensor transmits the collected heart rate data to the data collection terminal, and the data is preliminarily processed at the data collection terminal to remove abnormal values; The respiratory rate sensor unit is composed of a humidity sensor combined with a respiratory sensing sensor. The collected humidity data is transmitted to the data processing module to analyze and calculate the respiratory rate of the beef cattle.

4. The beef cattle individual information management system according to claim 1, characterized in that The data collection layer includes a behavior information collection module, including: The motion trajectory tracking unit uses a GPS positioning tracker installed on the collar of the beef cattle; the location information of the beef cattle is collected at preset time intervals, the collected location information data is uploaded to the cloud server, the collected motion trajectory data is analyzed, and the parameters of the activity range and movement speed of the beef cattle are calculated; The feeding and drinking monitoring unit includes an intelligent trough unit and an intelligent water trough unit. The intelligent trough is equipped with a built-in weighing sensor and an RFID identification module to monitor the weight changes of the feed in the trough in real time. At the same time, the electronic ear tag of the beef cattle is identified through the RFID identification module to record the feeding time and feeding amount of the beef cattle. The intelligent water trough is equipped with a flow sensor and a liquid level sensor to measure the drinking amount of the beef cattle and monitor the water level in the trough in real time. When the water level is lower than the set value, the water adding device is automatically triggered. The feeding and drinking data are processed to analyze the feeding and drinking patterns of the beef cattle.

5. The beef cattle individual information management system according to claim 1, wherein The data collection layer includes an environmental information collection module, including: The temperature and humidity sensor unit collects temperature and humidity data from different locations in the cowshed and transmits them to the data aggregation node, which then uploads them to the cloud server. The air quality sensor unit, including an ammonia sensor and a hydrogen sulfide sensor, is used to monitor the concentrations of ammonia and hydrogen sulfide in the cowshed in real time. When the ammonia concentration exceeds 20ppm and the hydrogen sulfide concentration exceeds 5ppm, the system automatically issues a warning message.

6. The beef cattle individual information management system according to claim 1, characterized in that, The data transmission layer uses RaspberryPi4B as the data aggregation node, which is responsible for collecting the data collected by each sensor, and preliminarily processing and storing these data; it supports ZigBee, LoRa, Bluetooth, 4G / 5G wireless communication protocols, and selects the appropriate communication protocol according to the type and transmission distance of the sensor; during the data transmission process, the AES encryption algorithm is used to ensure the security of data transmission.

7. The beef cattle individual information management system according to claim 1, characterized in that The data cleaning module of the data processing and storage layer processes the data using the following method: Outlier detection: The Z-score method based on statistical analysis is used to detect outliers, and the calculation formula is: Among them, Z i represents the data point value, x i represents the i-th data point, μ is the mean of the indicator data, and σ is the standard deviation. When |Z i | > 3, it is determined that this data point is an outlier; Missing value handling: The linear interpolation method is used to handle missing values. For the missing data x j , based on the adjacent data x j-1 and x j+1 to calculate. The formula is: where x j represents the missing data of the j-th data point, j + 1 represents the data point before the j-th data point, and j - 1 represents the data point after the j-th data point; Data standardization: The Min-Max standardization method is used to standardize the cleaned data, and the calculation formula is: Among them, x is the original data, min(x) and max(x) are the minimum and maximum values of this group of data respectively, and x’ is the standardized data.

8. The beef cattle individual information management system according to claim 1, characterized in that, The data fusion module of the data processing and storage layer uses the Kalman filter algorithm to fuse multi-source data, specifically as follows: The system state equation is: X k+1 = A k X k + B k U k + W k ; Among them, X k+1 represents the system state vector at time k+1, A k is the state transition matrix, X k is the system state vector at time k, B k is the control input matrix, U k is the control input vector at time k, W k is the process noise vector; The observation equation is: Z k = H k X k + V k ; where, Z k is the observation vector at time k, H k is the observation matrix, V k is the observation noise vector; The Kalman filter steps include: Prediction: This formula is used to estimate the state at time k-1 to predict the state estimate at time k for predicting the covariance matrix P at time k k|k-1 , where P k-1|k-1 is the covariance matrix at time k-1, and Q k is the process noise covariance matrix; Update: Calculate the Kalman gain K k , which is used to weigh the weights of the predicted value and the observed value; Update the state estimate at time k according to the Kalman gain P k|k =(I - K k H k )P k|k-1 , update the covariance matrix P at time k k|k , where I is the identity matrix and R k is the observation noise covariance matrix.

9. The beef cattle individual information management system according to claim 1, characterized in that The data analysis module of the data processing and storage layer includes: The health assessment unit uses the support vector machine algorithm for health assessment. For the linearly separable case, the optimization objective is: The constraint is s.t. y i (w T x i + b) ≥ 1, i = 1, 2, …, n; where w is the weight vector, b is the bias, and x i is the input sample, y i is the sample label, and n is the number of samples; The SVM model is trained using historical data, and the optimal model parameters are selected using the cross-validation method; The growth prediction unit uses the long short-term memory network algorithm to predict the growth of beef cattle. The core formulas of the LSTM unit include: Input gate: i t = σ(W ii x t + W hi h t-1 + b i ), where i t is the output of the input gate at time t, σ is the sigmoid function, W ii is the weight matrix input to the input gate, x t is the input at time t, W hi is the weight matrix from the hidden state of the previous time step to the input gate, h t-1 is the hidden state at time t - 1, b i is the bias vector of the input gate, and the input gate determines the amount of information input that needs to be updated into the cell state at the current time; Forgotten gate: f t = σ(W if x t + W hf h t-1 + b f ), f t is the output of the forgotten gate at time t, W if is the weight matrix input to the forgotten gate, W hf is the weight matrix from the hidden state of the previous time step to the forgotten gate, b f is the bias vector of the forgotten gate. The forgotten gate determines the amount of information from the cell state of the previous time step that needs to be forgotten; Cell state update: is the candidate cell state at time t, tanh is the hyperbolic tangent function, W ic is the weight matrix input to the candidate cell state, W hc is the weight matrix from the previous hidden state to the candidate cell state, b c is the bias vector of the candidate cell state; C t is the cell state at time t, ⊙ is element-wise multiplication, and this formula updates the cell state according to the outputs of the forget gate and the input gate; Output gate: o t = σ(W io x t + W ho h t-1 + b o ), o t is the output of the output gate at time t, W io is the weight matrix input to the output gate, W ho is the weight matrix from the previous hidden state to the output gate, b o is the bias vector of the output gate; Hidden state: h t = o t ⊙tanh(C t ), h t is the hidden state at time t; the LSTM model is trained using the historical growth data of beef cattle, and the future growth prediction result can be output by inputting the current growth data.

10. A beef cattle individual information management system according to claim 1, characterized in that, The data storage module of the data processing and storage layer includes a distributed file system and a relational database; among them, The distributed file system is used to store original sensor data, images, and large data volume files related to videos; The relational database is used to store the data after processing and analysis. The data after processing and analysis includes the basic information of beef cattle, health assessment results, and growth prediction data.

Images