A modern intelligent deer breeding method and system

Abstract

The present invention relates to the technical field of feeding methods, and specifically to a modern intelligent deer breeding method and system thereof, including monitoring the temperature, humidity, and gas concentration in a deer house, monitoring the physiological index data and behavior data of a deer herd, transmitting the data to a cloud database, intelligently supplying feed to the deer herd based on the physiological index data and growth and development needs of the deer herd, establishing a deer herd behavior model, identifying the normal behavior and abnormal behavior of the deer herd, analyzing the changing trend of the physiological index data of the deer herd, combining intelligent drug delivery technology to improve the health level of the deer herd, combining environmental perception data and meteorological information to automatically adjust the environment inside and outside the deer house to improve the adaptability of the deer herd. The present invention introduces sensor technology and intelligent algorithms to realize real-time monitoring of the health status of the deer herd, uses physiological index data and intelligent technology to intelligently supply feed to the deer herd, avoids feed waste, guarantees the nutritional needs of the deer herd, and improves feed utilization and economic benefits.

Description

Technical Field

[0001] The present invention relates to the technical field of breeding methods, and in particular to a modern intelligent deer breeding method and system thereof. Background Art

[0002] In modern deer farming, it is crucial to effectively monitor and manage the health of deer. In traditional deer farming, monitoring the health of deer and managing feed supply usually rely on manual observation and experience. Farmers need to regularly inspect the deer house, observe the behavior and physical condition of the deer, and manually supply feed. This traditional deer farming method has some problems. First, due to the limitations of manual observation, farmers often cannot detect health problems of deer in time, leading to the spread of diseases and delays in treatment. Secondly, manual feed supply and environmental adjustment are prone to inaccuracy and waste, resulting in feed waste and increased costs, and also affecting the growth efficiency and health of deer. Therefore, the existing technology has problems in monitoring and management, such as untimely monitoring, inaccurate feed supply, and inaccurate environmental adjustment.

[0003] In order to solve these problems, the present invention introduces advanced sensor technology, data analysis algorithms and intelligent management technology, realizes real-time monitoring, early warning and management of the health status of deer herds, as well as intelligent feed supply and environmental regulation, and improves the management level and production efficiency of the deer farming industry. Summary of the invention

[0004] Based on the above objectives, the present invention provides a modern intelligent deer breeding method and system.

[0005] A modern intelligent deer breeding method comprises the following steps:

[0006] S1, environmental perception: by deploying a multi-sensor network, the temperature, humidity, gas concentration and harmful gas content in the deer house are monitored in real time;

[0007] S2, health monitoring and early warning: using biosensors and image recognition technology to continuously monitor the physiological indicators and behavioral data of deer, including body temperature, heart rate, movement and appetite;

[0008] S3, data collection and storage: Transmit environmental perception data, physiological indicator data, and behavioral data to the cloud database through the Internet of Things to provide a data basis for subsequent data analysis and decision-making;

[0009] S4, feed supply: based on the physiological index data and growth and development needs of the deer herd, realize intelligent feed supply to the deer herd, meet the nutritional needs of the deer herd, and reduce feed waste;

[0010] S5, Behavior analysis and guidance: By analyzing the behavior data of deer, a deer behavior model is established to identify the normal and abnormal behaviors of deer, and behavior guidance is carried out through sound and light to improve the quality of life and behavioral regularity of deer;

[0011] S6, Intelligent disease management: Use data mining technology to analyze the changing trends of physiological indicators of deer herds, realize disease diagnosis and management, and combine intelligent drug delivery technology to provide treatment and preventive measures to improve the health level of deer herds;

[0012] S7, ecological breeding environment optimization: combining environmental perception data and meteorological information to automatically adjust the environment inside and outside the deer house, optimize the breeding environment, and improve the adaptability and growth efficiency of the deer herd.

[0013] Furthermore, the environmental perception includes the following steps:

[0014] S11, temperature sensor: a thermocouple sensor is used to monitor the temperature changes in the deer house;

[0015] S12, humidity sensor: a resistive humidity sensor is used to monitor the humidity level in the deer house;

[0016] S13, gas sensor: using infrared absorption sensor to monitor the gas composition in the deer house, including the concentration of oxygen and carbon dioxide;

[0017] S14, Harmful Gas Sensor: An electrochemical gas sensor is used to monitor the harmful gas content in the deer house, including ammonia and hydrogen sulfide.

[0018] Furthermore, the health monitoring and early warning includes the following steps:

[0019] S21, biosensors monitor physiological indicators of deer herds, including:

[0020] S211, implanting a body temperature sensor into the body of a deer or attaching it to the surface of the deer to continuously monitor the changes in the body temperature of the deer;

[0021] S212, attaching a heart rate sensor to the surface of the deer body to monitor the heart rate of the deer in real time;

[0022] S213, uses motion sensors to monitor the movement and activity levels of deer;

[0023] S214, using weighing sensors to monitor the eating behavior and food intake of deer;

[0024] S22, image recognition technology monitors behavioral data, including:

[0025] S221, camera installation: cameras are placed inside or around the deer house to cover the deer activity area and to capture the deer behavior in real time;

[0026] S222, Image processing and analysis: Use computer vision technology to process and analyze images captured by the camera to identify the behavioral characteristics of deer;

[0027] S223, Behavior classification and tracking: Classify and track the behavior of deer based on image recognition algorithms;

[0028] S224, abnormal behavior detection: abnormal behavior is detected by comparing the behavior patterns of deer.

[0029] Furthermore, the image recognition algorithm in S223 adopts a convolutional recurrent neural network algorithm, including:

[0030] S2231, convolution layer calculation formula: In the convolution layer, the features of the image are extracted through the convolution operation. Assuming that the input image is X, the convolution kernel is W, the bias is b, and the output feature map is Y, the calculation formula of the convolution operation is: Y i，j =f(∑ m ∑ n X i+m,j+n ·W m，n + b);

[0031] Where f is the activation function, X i，j Represents the pixel value of the input image, W m,n represents the weight of the convolution kernel, b represents the bias term, and Y i，j Represents the pixel value of the output feature map after convolution;

[0032] S2232, loop layer calculation formula: In the loop layer, the long short-term memory network is used to model the sequence data. Assume that the sequence data is X = {x1, x2, ..., x T}, the hidden state is H = {h1, h2, ..., h T}, then the calculation formula of the loop layer is as follows: t =f RNN (x t ,h t-1 );

[0033] Among them, f RNN represents the computational function of the loop layer, h t represents the hidden state at time step t;

[0034] S2233, Forward propagation of convolutional recurrent neural network: In the convolutional recurrent neural network, the forward propagation process is to first extract image features through the convolution layer, and then input the feature sequence into the recurrent layer for sequence modeling. The specific steps are as follows:

[0035] S22331, use the convolution layer to extract features from the input image to obtain a feature map sequence;

[0036] S22332, input the feature map sequence as sequence data into the circulation layer, perform sequence modeling, and obtain the final output sequence.

[0037] Further, the feed supply comprises the following steps:

[0038] S41, Nutritional requirement model: Based on the physiological index data of deer herds and the growth and development requirements of deer herds, which include protein, cellulose, water, energy, vitamins and minerals, a nutritional requirement model of deer herds is established;

[0039] S42, feed ingredient preparation: Based on the nutritional demand model, intelligent algorithms are used to intelligently prepare the feed ingredients for the deer herd. According to the physiological index data and growth and development stage of the deer herd, the feed ingredients are dynamically adjusted to ensure that the nutritional content of the feed meets the needs of the deer herd;

[0040] S421, intelligent algorithm: The multivariate linear regression algorithm is used to establish the relationship between physiological index data and required nutrients. Its mathematical expression is:

[0041] There are n sample data, each of which contains m features (independent variables), expressed as:

[0042] x ij , j = 1, 2, ..., m, and a target variable (dependent variable), denoted as y i , i=1,2,...,n;

[0043] The basic form of the regression model is: i =β0+β1x i1 +β2x i2 +…+β m x im +ε i ;

[0044] Where: y i represents the target variable (feed ingredient requirement) of the ith sample, x ij represents the jth feature (physiological index data) of the i-th sample, β0, β1, ..., β m is the regression coefficient, which indicates the influence of the independent variable on the dependent variable, ε i is the error term, which represents the random error that cannot be explained by the model, and is calculated by minimizing the observed value y i With the model prediction value The residual sum of squares between the two is used to estimate the regression coefficients β0, β1, ..., βm , that is, to solve by minimizing the loss function:

[0045] in It represents the predicted value of the target variable of the i-th sample by the model, which is calculated as follows:

[0046]

[0047] The goal of the multi-distance linear regression algorithm is to find the regression coefficients β0, β1, ..., β that minimize the loss function. m , thereby establishing a linear relationship model between physiological index data and feed ingredient requirements;

[0048] S43, quantitative supply: intelligent feeders are used to control the amount of feed supplied according to the number, weight and nutritional needs of the deer;

[0049] S44, real-time monitoring and feedback: The feed consumption of the deer herd is monitored in real time through weighing sensors, and the weighing sensor data is fed back to the intelligent feeder to adjust the feed supply according to the real-time situation.

[0050] Furthermore, the behavior analysis guidance includes:

[0051] S51, establish behavioral model: Based on behavioral data analysis, unsupervised learning algorithm is used to establish a deer behavior model to distinguish normal behavior from abnormal behavior;

[0052] S52, abnormal behavior detection and alarm: monitor and analyze the behavior data of deer herds according to the behavior model. When abnormal behavior is detected, the alarm is automatically triggered to notify relevant personnel or managers to take corresponding measures

[0053] S53, sound and light guidance: Analyze the behavior patterns of deer and determine the sound and light signals corresponding to the behavior patterns. When abnormal behavior of deer is detected, guide the deer by issuing sound signals or adjusting the light intensity to restore normal behavior.

[0054] Furthermore, the unsupervised learning algorithm in S51 is a K-means clustering algorithm. The K-means clustering algorithm is used to divide the behavior data of the deer herd into different clusters, each cluster representing a behavior pattern. The distinction between normal behavior and abnormal behavior includes:

[0055] S511, cluster center analysis: for each cluster, calculate the center point of the cluster, which represents the behavior pattern of the cluster;

[0056] S512, anomaly detection: for each cluster, the average distance (or other appropriate distance measurement) from the data points in the cluster to the cluster center is calculated. If the distance from the data point to the cluster center is much larger than the average distance, it indicates that the behavior corresponding to the data point is an abnormal behavior;

[0057] S513, threshold setting: setting a threshold according to actual conditions to determine whether the behavior is abnormal;

[0058] S514, manual confirmation and adjustment: Behaviors initially judged to be abnormal are manually confirmed and adjusted to ensure the accuracy of the judgment.

[0059] Furthermore, the calculation formula of the K-means clustering algorithm includes:

[0060] S515, in K-means clustering, the Euclidean distance is used to calculate the distance between the data point and the cluster center; the Euclidean distance formula is:

[0061] Among them, x and y are the feature vectors of two data points, x i and i They represent the i-th component in the eigenvector, and n is the dimension of the eigenvector;

[0062] S516, outlier detection: For each cluster, the average distance from the data points in the cluster to the cluster center is calculated, and the average distance plus or minus the number of standard deviations is used as the threshold of the outlier. The specific calculation formula includes:

[0063] S5161, average distance: For each cluster C j , calculate the distance from all data points in the cluster to the cluster center μ j The distance between the clusters is calculated and the average value is obtained, that is, the average distance within the cluster mean(d(x i , μ j )), where d(x i , μ j ) represents the data point x i To cluster center μ j distance;

[0064] S5162, standard deviation: Calculate the standard deviation σ of the distance from all data points in the cluster to the cluster center j , measures the degree of dispersion between data points and cluster centers;

[0065] S5163, outlier threshold: Set the outlier threshold to mean(d(x i , μ j ))±k×σ j , where k is a constant representing the multiple of the standard deviation and is used to adjust the sensitivity of outliers;

[0066] S5164, for each data point x i Calculate its distance to the center of its cluster μ j The distance d(x i , μ j ), and judge whether the data point is an outlier according to the outlier threshold. The calculation formula of outlier detection is expressed as:

[0067] Outlier threshold = mean(d(x i , μ j ))±k×σ j ;

[0068] Among them, mean(d(x i , μ j )) represents the average distance within the cluster, σ j It represents the standard deviation of the distance within the cluster, and k is a multiple of the standard deviation, which is used to adjust the sensitivity of outliers.

[0069] Furthermore, the intelligent disease management includes:

[0070] S61, establish health model: based on physiological index data, use evolutionary strategy algorithm to establish deer health model, use historical data to train health model, so that health model learns the characteristics of health status from physiological index data, and identifies healthy and sick deer;

[0071] S611, the evolution strategy includes the following main steps:

[0072] S612, initialization: randomly generate a set of initial solutions as individuals in the population;

[0073] S613, mutation: Perform random mutation operations on individuals in the population to generate new candidate solutions. The mutation operation includes random perturbations or changes to the parameters of the individuals.

[0074] S614, selection: evaluate the quality of candidate solutions according to a certain fitness function, and select individuals with high fitness as the parents of the next generation population;

[0075] S615, Recombination: Recombining the selected parent individuals through crossover or other operations to produce offspring of the next generation population;

[0076] S616, Update: Replace the original parent generation with the newly generated offspring to form the next generation population;

[0077] S617, termination condition: repeat the above steps until the preset optimization target is met;

[0078] S618, the specific steps of the evolution strategy are as follows:

[0079] Assume that the optimization problem is to minimize the objective function f(x), where x is the parameter vector to be optimized

[0080] The optimization process of evolution strategy is described as:

[0081] S6181, objective function: f(x);

[0082] S6182, population size: λ, the number of individuals in each generation;

[0083] S6183, mutation operation: x′=x+σ·∈, where σ is the mutation step length and ∈ is the random perturbation vector;

[0084] S6184, fitness function: fitness(x), measures the fitness or quality of each individual;

[0085] S6185, selection operation: select parent individuals according to the value of fitness function;

[0086] S6186, recombination operation: the selected parent individuals are recombined to produce offspring;

[0087] S6187, update operation: replace the original parent generation with the newly generated offspring to form the next generation population;

[0088] S6188, termination condition: satisfying the preset optimization goal;

[0089] S62, Disease Diagnosis and Management: Based on the established health model, newly collected physiological indicator data are predicted and diagnosed to determine the current health status of the deer herd;

[0090] S63, intelligent drug delivery technology: using intelligent drug feeders to deliver drugs to deer that need treatment or prevention based on disease diagnosis results and health model predictions.

[0091] A modern intelligent deer breeding system, used to implement the above-mentioned modern intelligent deer breeding method, includes the following modules:

[0092] Environmental perception module: deploy a multi-sensor network, including temperature sensors, humidity sensors, gas sensors, and harmful gas sensors, to monitor the environmental parameters inside and outside the deer house in real time;

[0093] Health monitoring module: using biosensors and image recognition technology to continuously monitor the physiological indicators and behavioral data of deer, and achieve real-time monitoring and early warning of the health status of deer;

[0094] Data collection and storage module: transmits environmental perception data, physiological index data and behavioral data to the cloud database through the Internet of Things;

[0095] Intelligent algorithm module: Use data mining technology and machine learning algorithms to analyze monitoring data in real time, identify the health status and behavioral characteristics of deer, and provide corresponding warnings and suggestions;

[0096] Intelligent feed supply module: based on the physiological index data and growth and development needs of deer, realize intelligent feed supply for deer;

[0097] Ecological breeding environment optimization module: combines environmental perception data and meteorological information to automatically adjust the environment inside and outside the deer house.

[0098] The present invention introduces sensor technology and intelligent algorithms to achieve real-time monitoring and early warning of the health status of deer. By monitoring the physiological index data, behavioral data and environmental parameters of the deer, the system can promptly detect abnormal behaviors and health problems of the deer, and issue early warning notifications to help breeders take timely intervention measures, effectively prevent the occurrence and spread of diseases, and improve the health level of the deer.

[0099] The present invention uses physiological index data and intelligent technology to realize intelligent feed supply for deer. According to the growth and development needs and nutritional status of the deer, the system can accurately adjust the feed supply, avoid feed waste, ensure the nutritional needs of the deer, and improve feed utilization and economic benefits.

[0100] The present invention combines environmental sensing technology and intelligent technology to achieve automatic adjustment of the environment inside and outside the deer house. According to environmental sensing data and meteorological information, the temperature, humidity, ventilation and other parameters inside and outside the deer house are automatically adjusted, which improves the quality of life and growth efficiency of the deer herd, reduces the workload of breeders, and improves the management level of the deer breeding industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0101] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0102] Figure 1 A schematic diagram of the process flow of the intelligent deer breeding method according to an embodiment of the present invention;

[0103] Figure 2 The figure is a flow chart of the intelligent deer breeding system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0104] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0105] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0106] like Figure 1 As shown, a modern intelligent deer breeding method comprises the following steps:

[0107] S1, environmental perception: by deploying a multi-sensor network, the temperature, humidity, gas concentration and harmful gas content in the deer house are monitored in real time;

[0108] S2, health monitoring and early warning: using biosensors and image recognition technology to continuously monitor the physiological indicators and behavioral data of deer, including body temperature, heart rate, movement and appetite;

[0109] S3, data collection and storage: Transmit environmental perception data, physiological indicator data, and behavioral data to the cloud database through the Internet of Things to provide a data basis for subsequent data analysis and decision-making;

[0110] S4, feed supply: based on the physiological index data and growth and development needs of the deer herd, realize intelligent feed supply to the deer herd, meet the nutritional needs of the deer herd, and reduce feed waste;

[0111] S5, Behavior analysis and guidance: By analyzing the behavior data of deer, a deer behavior model is established to identify the normal and abnormal behaviors of deer, and behavior guidance is carried out through sound and light to improve the quality of life and behavioral regularity of deer;

[0112] S6, Intelligent disease management: Use data mining technology to analyze the changing trends of physiological indicators of deer herds, realize disease diagnosis and management, and combine intelligent drug delivery technology to provide treatment and preventive measures to improve the health level of deer herds;

[0113] S7, ecological breeding environment optimization: combining environmental perception data and meteorological information to automatically adjust the environment inside and outside the deer house, optimize the breeding environment, and improve the adaptability and growth efficiency of the deer herd.

[0114] Environmental perception includes the following steps:

[0115] S11, temperature sensor: a thermocouple sensor is used to monitor the temperature changes in the deer house;

[0116] S12, humidity sensor: a resistive humidity sensor is used to monitor the humidity level in the deer house;

[0117] S13, gas sensor: using infrared absorption sensor to monitor the gas composition in the deer house, including the concentration of oxygen and carbon dioxide;

[0118] S14, Harmful Gas Sensor: An electrochemical gas sensor is used to monitor the harmful gas content in the deer house, including ammonia and hydrogen sulfide.

[0119] Health monitoring and early warning include the following steps:

[0120] S21, biosensors monitor physiological indicators of deer herds, including:

[0121] S211, implanting a body temperature sensor into the body of a deer or attaching it to the surface of the deer body to continuously monitor the changes in the body temperature of the deer;

[0122] S212, attaching a heart rate sensor to the surface of the deer body to monitor the heart rate of the deer in real time;

[0123] S213, using motion sensors, such as accelerometers or gyroscopes, to monitor the movement and activity level of deer;

[0124] S214, using weighing sensors to monitor the eating behavior and food intake of deer;

[0125] S22, image recognition technology monitors behavioral data, including:

[0126] S221, camera installation: cameras are placed inside or around the deer house to cover the deer activity area and to capture the deer behavior in real time;

[0127] S222, Image processing and analysis: Use computer vision technology to process and analyze images captured by the camera to identify the behavioral characteristics of deer;

[0128] S223, behavior classification and tracking: Based on image recognition algorithms, the behavior of deer is classified and tracked, such as standing, walking, eating, resting, etc.;

[0129] S224, abnormal behavior detection: detecting abnormal behaviors, such as diseases or abnormal conditions in the deer herd, by comparing the behavior patterns of the deer herd.

[0130] The image recognition algorithm in S223 uses a convolutional recurrent neural network algorithm, including:

[0131] S2231, convolution layer calculation formula: In the convolution layer, the features of the image are extracted through the convolution operation. Assuming that the input image is X, the convolution kernel is W, the bias is b, and the output feature map is Y, the calculation formula of the convolution operation is: Y i，j =f(∑ m ∑ n X i+m,j+n ·W m，n + b);

[0132] Where f is the activation function, X i，j Represents the pixel value of the input image, W m,n represents the weight of the convolution kernel, b represents the bias term, and Y i，j Represents the pixel value of the output feature map after convolution;

[0133] S2232, loop layer calculation formula: In the loop layer, the long short-term memory network is used to model the sequence data. Assume that the sequence data is X = {x1, x2, ..., x T}, the hidden state is H = {h1, h2, ..., h T}, then the calculation formula of the loop layer is as follows: t =f RNN (x t ,h t-1 );

[0134] Among them, f RNN represents the computational function of the loop layer, h t represents the hidden state at time step t;

[0135] S2233, Forward propagation of convolutional recurrent neural network: In the convolutional recurrent neural network, the forward propagation process is to first extract image features through the convolution layer, and then input the feature sequence into the recurrent layer for sequence modeling. The specific steps are as follows:

[0136] S22331, use the convolution layer to extract features from the input image to obtain a feature map sequence;

[0137] S22332, input the feature graph sequence as sequence data into the recurrent layer, perform sequence modeling, and obtain the final output sequence;

[0138] The convolutional recurrent neural network combines the advantages of convolutional operations and recurrent operations, performs well in processing serialized image data, and can effectively capture the spatiotemporal information of deer behavior images.

[0139] Feed supply includes the following steps:

[0140] S41, Nutritional requirement model: Based on the physiological index data of deer, such as weight, body temperature, digestive system condition, etc., combined with the growth and development needs of deer, which include protein, cellulose, water, energy, vitamins and minerals, a nutritional requirement model of deer is established;

[0141] S42, feed ingredient preparation: Based on the nutritional demand model, intelligent algorithms are used to intelligently prepare the feed ingredients for the deer herd. According to the physiological index data and growth and development stage of the deer herd, the feed ingredients are dynamically adjusted to ensure that the nutritional content of the feed meets the needs of the deer herd;

[0142] S421, intelligent algorithm: The multivariate linear regression algorithm is used to establish the relationship between physiological index data and required nutrients. Its mathematical expression is:

[0143] There are n sample data, each of which contains m features (independent variables), expressed as:

[0144] x ij , j = 1, 2, ..., m, and a target variable (dependent variable), denoted as y i , i=1,2,...,n;

[0145] The basic form of the regression model is: i =β0+β1x i1 +β2x i2 +…+β m x im +ε i ;

[0146] Where: y i represents the target variable (feed ingredient requirement) of the ith sample, x ij represents the jth feature (physiological index data) of the i-th sample, β0, β1, ..., β m is the regression coefficient, which indicates the influence of the independent variable on the dependent variable, ε i is the error term, which represents the random error that cannot be explained by the model, and is calculated by minimizing the observed value y i With the model prediction value The residual sum of squares between the two is used to estimate the regression coefficients β0, β1, ..., β m , that is, to solve by minimizing the loss function:

[0147] in It represents the predicted value of the target variable of the i-th sample by the model, which is calculated as follows:

[0148]

[0149] The goal of the multi-distance linear regression algorithm is to find the regression coefficients β0, β1, ..., β that minimize the loss function. m , thereby establishing a linear relationship model between physiological index data and feed ingredient requirements;

[0150] S43, quantitative supply: intelligent feeders are used to control the amount of feed supplied according to the number, weight and nutritional needs of the deer;

[0151] S44, real-time monitoring and feedback: The feed consumption of the deer herd is monitored in real time through weighing sensors, and the weighing sensor data is fed back to the intelligent feeder to adjust the feed supply according to the real-time situation.

[0152] Behavior analysis guidance includes:

[0153] S51, establish behavioral model: Based on behavioral data analysis, unsupervised learning algorithm is used to establish a deer behavior model to distinguish normal behavior from abnormal behavior;

[0154] S52, abnormal behavior detection and alarm: monitor and analyze the behavior data of deer herds according to the behavior model. When abnormal behavior is detected, the alarm is automatically triggered to notify relevant personnel or managers to take corresponding measures

[0155] S53, sound and light guidance: Analyze the behavior patterns of the deer herd, determine the sound and light signals corresponding to the behavior patterns, and when abnormal behavior of the deer herd is detected, guide the deer herd by issuing sound signals or adjusting the light intensity to restore normal behavior;

[0156] For example, sound signals are used to simulate the warning sounds or feeding sounds of deer to attract them back to their normal activity areas. At the same time, by adjusting the brightness and color of the light, the atmosphere of the environment is changed to affect the behavioral regularity of the deer.

[0157] The unsupervised learning algorithm in S51 is the K-means clustering algorithm. Through the K-means clustering algorithm, the behavior data of the deer herd is divided into different clusters. Each cluster represents a behavior pattern. The distinction between normal behavior and abnormal behavior includes:

[0158] S511, cluster center analysis: for each cluster, the center point of the cluster is calculated to represent the behavior pattern of the cluster, and by comparing the distance between the current behavior data of the deer herd and each cluster center, it is determined which behavior pattern the current behavior is closer to;

[0159] S512, anomaly detection: for each cluster, the average distance (or other appropriate distance measurement) from the data points in the cluster to the cluster center is calculated. If the distance from the data point to the cluster center is much larger than the average distance, it indicates that the behavior corresponding to the data point is an abnormal behavior;

[0160] S513, threshold setting: setting a threshold according to actual conditions to determine whether the behavior is abnormal, for example, determining the threshold according to the distribution of data points within the cluster and the statistical characteristics of the distance;

[0161] S514, manual confirmation and adjustment: For behaviors initially judged as abnormal, manual confirmation and adjustment are performed to ensure the accuracy of the judgment;

[0162] Using unsupervised learning algorithms to establish behavioral models and distinguish between normal and abnormal behaviors can help monitor the behavior of deer in real time and detect abnormal behaviors in a timely manner, so that appropriate measures can be taken to deal with them and improve the quality of life and behavioral regularity of the deer.

[0163] The calculation formula of the K-means clustering algorithm includes:

[0164] S515, in K-means clustering, the Euclidean distance is used to calculate the distance between the data point and the cluster center; the Euclidean distance formula is:

[0165] Among them, x and y are the feature vectors of two data points, x i and i They represent the i-th component in the eigenvector, and n is the dimension of the eigenvector;

[0166] S516, outlier detection: For each cluster, the average distance from the data points in the cluster to the cluster center is calculated, and the average distance plus or minus the number of standard deviations is used as the threshold of the outlier. The specific calculation formula includes:

[0167] S5161, average distance: For each cluster C j , calculate the distance from all data points in the cluster to the cluster center μ j The distance between the clusters is calculated and the average value is obtained, that is, the average distance within the cluster mean(d(x i , μ j )), where d(x i , μ j ) represents the data point x i To cluster center μ j distance;

[0168] S5162, standard deviation: Calculate the standard deviation σ of the distance from all data points in the cluster to the cluster center j , measures the degree of dispersion between data points and cluster centers;

[0169] S5163, outlier threshold: Set the outlier threshold to mean(d(x i , μ j ))±k×σ j , where k is a constant representing the multiple of the standard deviation and is used to adjust the sensitivity of outliers;

[0170] S5164, for each data point x i Calculate its distance to the center of its cluster μ j The distance d(x i , μ j ), and judge whether the data point is an outlier according to the outlier threshold. The calculation formula of outlier detection is expressed as:

[0171] Outlier threshold = mean(d(x i , μ j ))±k×σ j ;

[0172] Among them, mean(d(x i , μ j )) represents the average distance within the cluster, σ j It represents the standard deviation of the distance within the cluster, and k is a multiple of the standard deviation, which is used to adjust the sensitivity of outliers.

[0173] Intelligent disease management includes:

[0174] S61, establish health model: based on physiological index data, use evolutionary strategy algorithm to establish deer health model, use historical data to train health model, so that health model learns the characteristics of health status from physiological index data, and identifies healthy and sick deer;

[0175] S611, the evolution strategy includes the following main steps:

[0176] S612, initialization: randomly generate a set of initial solutions as individuals in the population;

[0177] S613, mutation: Perform random mutation operations on individuals in the population to generate new candidate solutions. The mutation operation includes random perturbations or changes to the parameters of the individuals.

[0178] S614, selection: evaluate the quality of candidate solutions according to a certain fitness function, and select individuals with high fitness as the parents of the next generation population;

[0179] S615, Recombination: Recombining the selected parent individuals through crossover or other operations to produce offspring of the next generation population;

[0180] S616, Update: Replace the original parent generation with the newly generated offspring to form the next generation population;

[0181] S617, termination condition: repeat the above steps until the preset optimization target is met;

[0182] S618, the specific steps of the evolution strategy are as follows:

[0183] Assume that the optimization problem is to minimize the objective function f(x), where x is the parameter vector to be optimized

[0184] The optimization process of evolution strategy is described as:

[0185] S6181, objective function: f(x);

[0186] S6182, population size: λ, the number of individuals in each generation;

[0187] S6183, mutation operation: x′=x+σ·∈, where σ is the mutation step length and ∈ is the random perturbation vector;

[0188] S6184, fitness function: fitness(x), measures the fitness or quality of each individual;

[0189] S6185, selection operation: select parent individuals according to the value of fitness function;

[0190] S6186, recombination operation: the selected parent individuals are recombined to produce offspring;

[0191] S6187, update operation: replace the original parent generation with the newly generated offspring to form the next generation population;

[0192] S6188, termination condition: satisfying the preset optimization goal;

[0193] S62, Disease Diagnosis and Management: Based on the established health model, newly collected physiological indicator data are predicted and diagnosed to determine the current health status of the deer herd;

[0194] If diseases or abnormal conditions are found in the herd, timely treatment and management measures are taken, which may include adjusting feed ingredients, providing specific medical care, adjusting environmental conditions, etc.

[0195] S63, intelligent drug delivery technology: using intelligent drug feeders to deliver drugs to deer that need treatment or prevention based on disease diagnosis results and health model predictions;

[0196] Combining the evolutionary strategy algorithm and the intelligent medicine feeder, the health status of the deer herd can be monitored, diagnosed and managed, and timely treatment and preventive measures can be given, thereby improving the health level of the deer herd and the breeding efficiency.

[0197] like Figure 2 As shown, a modern intelligent deer breeding system is used to implement the above-mentioned modern intelligent deer breeding method, including the following modules:

[0198] Environmental perception module: deploy a multi-sensor network, including temperature sensors, humidity sensors, gas sensors, and harmful gas sensors, to monitor the environmental parameters inside and outside the deer house in real time;

[0199] Health monitoring module: using biosensors and image recognition technology to continuously monitor the physiological index data and behavioral data of deer, and realize real-time monitoring and early warning of the health status of deer;

[0200] Data collection and storage module: transmits environmental perception data, physiological index data and behavioral data to the cloud database through the Internet of Things;

[0201] Intelligent algorithm module: Use data mining technology and machine learning algorithms to analyze monitoring data in real time, identify the health status and behavioral characteristics of deer, and provide corresponding warnings and suggestions;

[0202] Intelligent feed supply module: Based on the physiological index data and growth and development needs of deer, it realizes intelligent feed supply to deer, reduces feed waste and improves feed utilization;

[0203] Ecological breeding environment optimization module: Combine environmental perception data and meteorological information to automatically adjust the environment inside and outside the deer house, optimize the breeding environment, and improve the quality of life and growth efficiency of the deer herd

[0204] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present invention is limited to these examples. Under the concept of the present invention, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the present invention as described above, which are not provided in detail for the sake of simplicity.

[0205] The present invention is intended to cover all such substitutions, modifications and variations that fall within the broad scope of the claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A modern and intelligent method for raising deer, characterized in that: The following steps are involved: S1, environmental perception: by deploying a multi-sensor network, the temperature, humidity and gas concentration in the deer house are monitored in real time; S2, health monitoring and early warning: using biosensors and image recognition technology to continuously monitor the physiological indicators and behavioral data of deer, including body temperature, heart rate, movement and appetite; S3, data collection and storage: transmitting environmental perception data, physiological indicator data, and behavioral data to the cloud database through the Internet of Things; S4, feed supply: based on the physiological index data and growth and development needs of the deer herd, realize intelligent feed supply to the deer herd, meet the nutritional needs of the deer herd, and reduce feed waste; S5, Behavior analysis and guidance: By analyzing the behavior data of deer, a deer behavior model is established to identify the normal and abnormal behaviors of deer, and behavior guidance is carried out through sound and light to improve the quality of life and behavioral regularity of deer; S6, Intelligent disease management: Use data mining technology to analyze the changing trends of physiological indicators of deer herds, realize the diagnosis and management of diseases, combine with intelligent drug delivery technology, provide treatment and preventive measures, and improve the health level of deer herds. The intelligent disease management specifically includes: S61, establish health model: based on physiological index data, use evolution strategy algorithm to establish health model of deer herd, use historical data to train health model, make health model learn health status characteristics from physiological index data, identify healthy and sick deer herds; S611, the evolution strategy algorithm includes the following steps: S612, initialization: randomly generate a set of initial solutions as individuals in the population; S613, mutation: Perform random mutation operations on individuals in the population to generate new candidate solutions. The mutation operation includes random perturbations or changes to the parameters of the individuals. S614, selection: evaluate the quality of candidate solutions according to the fitness function, and select individuals with high fitness as the parents of the next generation population; S615, Recombination: Recombining the selected parent individuals through crossover to produce offspring of the next generation population; S616, Update: Replace the original parent generation with the newly generated offspring to form the next generation population; S617, termination condition: repeat the above steps until the preset optimization target is met; S618, the specific steps of the evolution strategy are as follows: Assume that the optimization problem is to minimize the objective function ,in is the parameter vector to be optimized; The optimization process of evolution strategy is described as: S6181, objective function: ; S6182, population size: , that is, the number of individuals in each generation; S6183, mutation operation: ,in It is a variable asynchronous length. is the random perturbation vector; S6184, fitness function: fitness , which measures the fitness or quality of each individual; S6185, selection operation: select parent individuals according to the value of fitness function; S6186, recombination operation: the selected parent individuals are recombined to produce offspring; S6187, update operation: replace the original parent generation with the newly generated offspring to form the next generation population; S6188, termination condition: satisfying the preset optimization goal; S62, Disease Diagnosis and Management: Based on the established health model, newly collected physiological indicator data are predicted and diagnosed to determine the current health status of the deer herd; S63, intelligent drug delivery technology: using intelligent drug feeders to deliver drugs to deer that need treatment or prevention based on disease diagnosis results and health model predictions; S7, ecological breeding environment optimization: combining environmental perception data and meteorological information to automatically adjust the environment inside and outside the deer house, optimize the breeding environment, and improve the adaptability and growth efficiency of the deer herd.

2. A modern intelligent deer breeding method according to claim 1, characterized in that: The environmental perception comprises the following steps: S11, temperature sensor: a thermocouple sensor is used to monitor the temperature changes in the deer house; S12, humidity sensor: a resistive humidity sensor is used to monitor the humidity level in the deer house; S13, gas sensor: using infrared absorption sensor to monitor the gas composition in the deer house, including the concentration of oxygen and carbon dioxide; S14, Harmful Gas Sensor: An electrochemical gas sensor is used to monitor the concentration of harmful gases in the deer house, including ammonia and hydrogen sulfide.

3. A modern intelligent deer breeding method according to claim 1, characterized in that: The health monitoring and early warning comprises the following steps: S21, biosensors monitor physiological indicators of deer herds, including: S211, implanting a body temperature sensor into the body of a deer or attaching it to the surface of the deer body to continuously monitor the changes in the body temperature of the deer; S212, attaching a heart rate sensor to the surface of the deer body to monitor the heart rate of the deer in real time; S213, uses motion sensors to monitor the movement and activity levels of deer; S214, using weighing sensors to monitor the food intake of deer; S22, image recognition technology monitors behavioral data, including: S221, camera installation: cameras are placed inside or around the deer house to cover the deer activity area and to capture the deer behavior in real time; S222, Image processing and analysis: Use computer vision technology to process and analyze images captured by the camera to identify the behavioral characteristics of deer; S223, Behavior classification and tracking: Classify and track the behavior of deer based on image recognition algorithms; S224, abnormal behavior detection: abnormal behavior is detected by comparing the behavior patterns of deer.

4. A modern intelligent deer breeding method according to claim 3, characterized in that: The image recognition algorithm in S223 adopts a convolutional recurrent neural network algorithm, including: S2231, convolution layer calculation formula: In the convolution layer, the features of the image are extracted through the convolution operation. Assuming that the input image is X, the convolution kernel is W, the bias is b, and the output feature map is Y, the calculation formula of the convolution operation is: ; Where f is the activation function, represents the pixel value of the input image, represents the weight of the convolution kernel, b represents the bias term, Represents the pixel value of the output feature map after convolution; S2232, loop layer calculation formula: In the loop layer, the long short-term memory network is used to model the sequence data. Assume that the sequence data is , the hidden state is , then the calculation formula of the circulation layer is as follows: ; in, represents the computational function of the loop layer, Indicates that at time step The hidden state of S2233, forward propagation of convolutional recurrent neural network: In the convolutional recurrent neural network, the forward propagation process is to first extract image features through the convolution layer, and input the feature sequence into the recurrent layer for sequence modeling. The specific steps are as follows: S22331, use the convolution layer to extract features from the input image to obtain a feature map sequence; S22332, input the feature graph sequence as sequence data into the circulation layer, perform sequence modeling, and obtain an output sequence.

5. A modern intelligent deer breeding method according to claim 2, characterized in that: The feed supply comprises the following steps: S41, Nutritional requirement model: Based on the physiological index data of deer herds and the growth and development requirements of deer herds, which include protein, cellulose, water, energy, vitamins and minerals, a nutritional requirement model of deer herds is established; S42, feed ingredient preparation: Based on the nutritional demand model, intelligent algorithms are used to intelligently prepare the feed ingredients for the deer herd. According to the physiological index data and growth and development stage of the deer herd, the feed ingredients are dynamically adjusted to ensure that the nutritional content of the feed meets the needs of the deer herd; S421, Intelligent Algorithm: Multiple linear regression algorithm is used to establish the relationship between physiological index data and required nutrients. Its mathematical expression is: There are n sample data, each of which contains m features, expressed as: , and a target variable, expressed as ; The basic form of the regression model is: ; in: Indicates The target variable for each sample is Indicates The sample Features, is the regression coefficient, which indicates the influence of the independent variable on the dependent variable. is the error term, which represents the random error that cannot be explained by the model, and is calculated by minimizing the observed value With the model prediction value The residual sum of squares between the two is used to estimate the regression coefficient , that is, to solve by minimizing the loss function: ; in Represents the model for The predicted value of the target variable for samples is calculated as: ； S43, quantitative supply: intelligent feeders are used to control the amount of feed supplied according to the number, weight and nutritional needs of the deer; S44, real-time monitoring and feedback: The feed consumption of the deer herd is monitored in real time through weighing sensors, and the weighing sensor data is fed back to the intelligent feeder to adjust the feed supply according to the real-time situation.

6. A modern and intelligent deer breeding method according to claim 3, characterized in that: The behavior analysis guidance includes: S51, Establishing a deer behavior model: Based on behavioral data analysis, a deer behavior model is established in combination with an unsupervised learning algorithm to distinguish between normal and abnormal behaviors; S52, abnormal behavior detection and alarm: monitor and analyze the behavior data of the deer herd according to the deer herd behavior model. When abnormal behavior is detected, an alarm is automatically triggered to notify relevant personnel or managers to take corresponding measures; S53, sound and light guidance: Analyze the behavior patterns of deer and determine the sound and light signals corresponding to the behavior patterns. When abnormal behavior of deer is detected, guide the deer by issuing sound signals or adjusting the light intensity to restore normal behavior.

7. A modern and intelligent deer breeding method according to claim 6, characterized in that: The unsupervised learning algorithm in S51 is a K-means clustering algorithm. The K-means clustering algorithm is used to divide the behavior data of the deer herd into different clusters. Each cluster represents a behavior pattern. The distinction between normal behavior and abnormal behavior includes: S511, cluster center analysis: for each cluster, calculate the center point of the cluster, which represents the behavior pattern of the cluster; S512, anomaly detection: for each cluster, the average distance between the data points in the cluster and the cluster center is calculated. If the distance between the data point and the cluster center is much larger than the average distance, it indicates that the behavior corresponding to the data point is an abnormal behavior. S513, threshold setting: setting a threshold according to actual conditions to determine whether the behavior is abnormal; S514, manual confirmation and adjustment: Behaviors initially judged to be abnormal are manually confirmed and adjusted to ensure the accuracy of the judgment.

8. A modern and intelligent deer breeding method according to claim 7, characterized in that: The calculation formula of the K-means clustering algorithm includes: S515, in K-means clustering, the Euclidean distance is used to calculate the distance between the data point and the cluster center; the Euclidean distance formula is: ; Among them, x and y are the feature vectors of two data points, and They represent the first components, n is the dimension of the eigenvector; S516, outlier detection: For each cluster, the average distance from the data points in the cluster to the cluster center is calculated, and the average distance plus or minus the number of standard deviations is used as the threshold of the outlier. The specific calculation formula includes: S5161, average distance: for each cluster , calculate the distance from all data points in the cluster to the cluster center The distance is calculated and the average value is obtained, which is the average distance within the cluster. ,in Represents data points To cluster center distance; S5162, Standard Deviation: Calculates the standard deviation of the distances from all data points in a cluster to the cluster center. , measures the degree of dispersion between data points and cluster centers; S5163, Outlier threshold: Set the outlier threshold to , where k is a constant representing the multiple of the standard deviation and is used to adjust the sensitivity of outliers; S5164, for each data point Calculate the center of the cluster to which it belongs Distance , according to the outlier threshold, whether the data point is an outlier is determined. The calculation formula for outlier detection is expressed as: Outlier Threshold ; in, represents the average distance within a cluster, It represents the standard deviation of the distance within the cluster, and k is a multiple of the standard deviation, which is used to adjust the sensitivity of outliers.

9. A modern intelligent deer breeding system, used to implement a modern intelligent deer breeding method as claimed in any one of claims 1 to 8, characterized in that: Includes the following modules: Environmental perception module: deploy a multi-sensor network, including temperature sensors, humidity sensors, and gas sensors, to monitor the environmental parameters inside and outside the deer house in real time; Health monitoring module: using biosensors and image recognition technology to continuously monitor the physiological indicators and behavioral data of deer, and achieve real-time monitoring and early warning of the health status of deer; Data collection and storage module: transmits environmental perception data, physiological index data and behavioral data to the cloud database through the Internet of Things; Intelligent algorithm module: Use data mining technology and machine learning algorithms to analyze monitoring data in real time and identify the health status and behavioral characteristics of deer; Intelligent feed supply module: based on the physiological index data and growth and development needs of deer, realize intelligent feed supply for deer; Ecological breeding environment optimization module: combines environmental perception data and meteorological information to automatically adjust the environment inside and outside the deer house.

Images