Intelligent monitoring method and system in caged chicken house

Abstract

The present application discloses an intelligent monitoring method and system in a caged chicken house, which relates to the field of intelligent poultry farming technology, and aims to solve the problem that the traditional caged chicken inspection method has low efficiency in monitoring the health of chickens. The intelligent monitoring method in a caged chicken house includes: inputting the environmental parameter value into a first multidimensional data processing device for data analysis and prediction to obtain a first prediction result; inputting the behavior data into a second multidimensional data processing device for data analysis and prediction to obtain a second prediction result; calculating the first prediction result and the second prediction result according to a preset algorithm to obtain a third prediction result; sending the third prediction result to a central control device so that the central control device displays the third prediction result to the user; obtaining the control instruction fed back by the user, and based on the control instruction, intelligently managing the physical environment management device in the chicken house. The above steps improve the management efficiency of the chicken house by the breeder.

Description

Technical Field

[0001] The present application relates to the technical field of intelligent poultry farming, and in particular to an intelligent monitoring method and system in a caged chicken house. Background Art

[0002] In the livestock industry, especially in caged chicken farms, it is crucial to inspect and manage the health and environmental conditions of chickens. Traditional chicken house inspection methods usually rely on manual inspections and empirical judgment, which has several obvious defects that limit its efficiency and accuracy.

[0003] First, the traditional manual inspection method has the problems of high human resource costs and time costs. Large chicken farms usually raise hundreds of thousands of chickens and need to regularly inspect thousands of chicken cages, which places high demands on manual inspection personnel and easily leads to low inspection efficiency and missed inspections.

[0004] Secondly, manual inspections often make it difficult to monitor and accurately assess the health of chickens in real time. Chicken health problems may worsen if they are not discovered in the early stages, thus affecting the production performance and health of the entire flock.

[0005] In addition, traditional methods have limitations in data collection and analysis. Although some chicken farms may use manual records or simple spreadsheets to record data, these methods usually cannot provide real-time data analysis and trend identification, limiting the accuracy and timeliness of overall health management of chicken farms.

[0006] In summary, the traditional cage chicken inspection method has many limitations, which make it difficult for chicken farm managers to achieve efficient, accurate and timely monitoring and management of the health of the chickens. Therefore, it is necessary to introduce a new technical solution to solve these problems and improve the management level and production efficiency of cage chicken farms. The technical solution proposed in this application is designed to meet these challenges. By combining advanced technologies such as the Internet of Things, image recognition and data analysis, real-time monitoring and precise management of the chicken farm environment and the health of the chickens can be achieved, thereby improving the operational efficiency and economic benefits of the chicken farm. Summary of the invention

[0007] The present application provides an intelligent monitoring method and system for caged chicken houses, aiming to solve the problem of low efficiency of traditional caged chicken inspection methods in monitoring the health of chicken flocks.

[0008] In order to solve the above technical problems, the embodiment of the present application provides: a method for intelligent monitoring in a caged chicken house, the method comprising the following steps:

[0009] The environmental parameter values ​​are input into a first multidimensional data processing device for data analysis and prediction to obtain a first prediction result; the behavioral data are input into a second multidimensional data processing device for data analysis and prediction to obtain a second prediction result; wherein the environmental parameter values ​​include humidity values, temperature values, target gas concentration values, lighting values ​​and feed consumption data, which are obtained by monitoring the environment in the caged chicken house through an environmental sensor; the behavioral data include behavioral video data and behavioral image data, which are obtained by monitoring the behavior of all chickens in the caged chicken house through a behavioral data acquisition device;

[0010] After calculating the first prediction result and the second prediction result according to a preset algorithm, a third prediction result is obtained;

[0011] sending the third prediction result to the central control device so that the central control device displays the third prediction result to the user;

[0012] The control instructions fed back by the user are obtained, and based on the control instructions, the physical environment management device in the chicken house is intelligently managed.

[0013] As some optional implementation methods of the present application, the step of inputting the environmental parameter value into a first multidimensional data processing device for data analysis and prediction to obtain a first prediction result; and inputting the behavior data into a second multidimensional data processing device for data analysis and prediction to obtain a second prediction result includes:

[0014] Inputting the environmental parameter value into a time series prediction model for data analysis and prediction to obtain a first prediction result;

[0015] The behavior data is input into a behavior recognition model and a target detection model for data analysis and prediction to obtain a second prediction result.

[0016] As some optional implementations of the present application, the first prediction result includes temperature fluctuation data, humidity fluctuation data, gas concentration change data, lighting change data and feed consumption data at a specific time in the future;

[0017] The second prediction result includes activity pattern data, appetite change data and abnormal behavior data of the target chicken.

[0018] As some optional implementations of the present application, the step of inputting the environmental parameter value into the time series prediction model for data analysis and prediction to obtain a first prediction result includes:

[0019] Perform seasonal decomposition on the environmental parameter to obtain a first environmental parameter and a second environmental parameter; wherein the first environmental parameter has seasonal changes or periodic changes; and the second environmental parameter has no seasonal changes or periodic changes;

[0020] Inputting the first environmental parameter into the Prophet time series prediction model for data analysis and prediction to obtain a first environmental prediction parameter value;

[0021] The second environmental parameter is input into the ARIMA time series prediction model for data analysis and prediction to obtain the second environmental prediction parameter value.

[0022] As some optional implementations of the present application, the step of inputting the behavior data into the behavior recognition model and the target detection model for data analysis and prediction to obtain a second prediction result includes:

[0023] After preprocessing the behavior data, the behavior data is input into a behavior recognition model for data analysis to screen and obtain data of abnormal chickens; the abnormal chickens have abnormal behaviors;

[0024] Inputting the data of the abnormal chicken into the target detection model, locating the abnormal chicken, and obtaining the real-time position of the abnormal chicken; and based on the real-time position of the abnormal chicken, monitoring its behavior in real time to obtain the real-time behavior data of the abnormal chicken;

[0025] The real-time behavior data of the abnormal chickens is input into the long short-term memory network model for analysis and prediction to obtain a second prediction result.

[0026] As some optional implementations of the present application, the step of obtaining a third prediction result after calculating the first prediction result and the second prediction result according to a preset algorithm includes:

[0027] After preprocessing, the first prediction result and the second prediction result are input into a trained decision tree model, so that the decision tree model outputs prediction information;

[0028] The prediction information includes production prediction information, environmental prediction information, feed consumption prediction information, chicken flock behavior prediction information and chicken flock health status prediction information.

[0029] As some optional implementations of the present application, after the step of outputting prediction information by the decision tree model, the following steps are further included:

[0030] Inputting the forecast information into a decision support system to obtain management suggestion information output by the decision support system; the decision support system includes judgment conditions corresponding to each forecast information;

[0031] When the processing level of each of the prediction information is less than or equal to level 1, the management suggestion information does not require manual intervention;

[0032] When the processing level of each of the prediction information is greater than level one and less than or equal to level two, the management suggestion information is for automatic control equipment to make adjustments;

[0033] When the processing level of each of the prediction information is greater than level 2 and less than or equal to level 3, the management suggestion information requires manual intervention.

[0034] As some optional implementations of the present application, the decision tree model is trained by the following steps:

[0035] Obtain historical environmental parameter data and historical chicken behavior data in the chicken house;

[0036] Preprocessing the historical environmental parameter data and the historical chicken behavior data in the chicken house to obtain standardized training data;

[0037] An initial decision tree model is constructed by a decision tree algorithm, and the standardized training data is input into the initial decision tree model for training to obtain a trained decision tree model.

[0038] As some optional implementations of the present application, the decision support system is trained by the following steps:

[0039] Collecting chicken house historical data; the chicken house historical data includes chicken house environment historical parameters, chicken flock behavior historical data, feed consumption historical data, chicken flock health status historical data and production historical data;

[0040] After preprocessing the chicken house historical data, extract characteristic information of the chicken house historical data to obtain the target chicken house historical data; the characteristic information includes the average value and change rate of chicken house environmental parameters, feed consumption trend, frequency of chicken group activity patterns, chicken group health status change trend and production average value;

[0041] The target chicken house historical data is divided into a training set and a validation set. After the initial machine learning model is trained based on the training set, the hyperparameters of the initial machine learning model are optimized based on the validation set to obtain a target decision support system.

[0042] On the other hand, the present application also provides an intelligent monitoring system for caged chicken houses, comprising:

[0043] A first multi-dimensional data processing device is used to input the environmental parameter values ​​into the first multi-dimensional data processing device for data analysis and prediction to obtain a first prediction result; the environmental parameter values ​​include humidity values, temperature values, target gas concentration values, lighting values ​​and feed consumption data, which are obtained by monitoring the environment in the cage chicken house through an environmental sensor;

[0044] A second multi-dimensional data processing device is used to input the behavior data into the second multi-dimensional data processing device for data analysis and prediction to obtain a second prediction result; the behavior data includes behavior video data and behavior image data, which are obtained by monitoring the behavior of all chickens in the cage chicken house through the behavior data acquisition device;

[0045] A prediction device, configured to calculate the first prediction result and the second prediction result according to a preset algorithm to obtain a third prediction result; and send the third prediction result to a user platform so that the user platform displays the third prediction result to a user;

[0046] A central control device, used to obtain control instructions fed back by the user platform, and intelligently manage the physical environment management device in the chicken house based on the control instructions;

[0047] The intelligent monitoring system for caged chicken houses also includes a user platform for receiving the third prediction result sent by the prediction device, and generating a control instruction based on the third prediction result; and feeding back the control instruction to the central control device, so that the central control device can intelligently manage the physical environment management device in the chicken house based on the control instruction.

[0048] Compared with the prior art, the intelligent monitoring method in the caged chicken house provided by the present application provides more accurate prediction results by analyzing and predicting the comprehensive data, and at the same time, through the real-time monitoring and feedback mechanism, intelligent management is carried out, and the optimization efficiency is improved. Specifically, the method described in the present application uses two multidimensional data processing devices to process environmental parameters and behavioral data respectively, and performs data analysis and prediction. This comprehensive analysis capability enables the system to fully understand the situation in the chicken house from multiple dimensions, including environmental parameters such as humidity, temperature, gas concentration, lighting and feed consumption, as well as the behavior patterns and activities of chickens. And the first prediction result and the second prediction result obtained by the multidimensional data processing device are calculated by a preset algorithm to obtain a third prediction result. This process utilizes data analysis and prediction technology, which can provide more accurate and reliable predictions. For example, it can predict the impact of environmental changes in the chicken house on the behavior of chickens, or predict the environmental status in the chicken house in the future. By sending the third prediction result to the central control device, the user can understand the status and prediction information of the chicken house in real time. This real-time monitoring and feedback mechanism helps to adjust the management strategy in time and improve the efficiency and production quality of chicken farming. Finally, according to the control instructions fed back by the user, the system can intelligently manage the physical environment management device in the chicken house. This means that the system can automatically adjust parameters such as temperature, humidity, and lighting to optimize the living environment of the chickens, thereby improving breeding efficiency and health levels. It can be seen that the intelligent monitoring method described in this application can not only provide comprehensive data analysis and prediction capabilities, but also significantly improve the management efficiency and production quality of the chicken farm through real-time monitoring and feedback mechanisms and intelligent management. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the specific implementation of the present application or the technical solution in the prior art, the following is a brief introduction to the drawings required for the specific implementation or the prior art description. In all drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, each element or part is not necessarily drawn according to the actual scale.

[0050] Figure 1 A schematic diagram of the computer device structure of the hardware operating environment involved in this application;

[0051] Figure 2 A schematic diagram of the process of the intelligent monitoring method in the caged chicken house involved in this application;

[0052] Figure 3 This is a schematic diagram of the framework of the intelligent monitoring system in the cage chicken house involved in this application.

[0053] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0055] It should be noted that all directional indications in the embodiments of the present application (such as up, down, left, right, front, back, etc.) are only used to explain the relative position relationship, movement status, etc. between the components in a certain specific posture. If the specific posture changes, the directional indication will also change accordingly.

[0056] In this application, unless otherwise clearly specified and limited, the terms "connection", "fixation", etc. should be understood in a broad sense. For example, "fixation" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements, unless otherwise clearly defined. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0057] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present application, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the meaning of "and / or" appearing in the full text includes three parallel schemes. Taking "A and / or B" as an example, it includes scheme A, or scheme B, or a scheme that satisfies both A and B. In addition, the technical solutions between the various embodiments can be combined with each other, but it must be based on the ability of ordinary technicians in the field to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0058] Reference Figure 1 , Figure 1 This is a schematic diagram of the computer device structure of the hardware operating environment involved in this embodiment, such as Figure 1As shown, the computer device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (WIreless-FIdelity, WI-FI) interface). The memory 1005 may be a high-speed random access memory (Random Access Memory, RAM) memory, or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk memory. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0059] Those skilled in the art will understand that Figure 1 The structure shown in the figure does not constitute a limitation on the computer device, and may include more or less components than shown in the figure, or combine some components, or arrange the components differently.

[0060] like Figure 1 As shown, the memory 1005 as a storage medium may include an operating system, a network communication module, a user interface module, and an electronic program, and may also include a data storage module.

[0061] exist Figure 1 In the computer device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the computer device of this embodiment can be set in the computer device, and the computer device calls the caged chicken house intelligent monitoring system stored in the memory 1005 through the processor 1001, and executes the caged chicken house intelligent monitoring method provided in this embodiment.

[0062] It should be noted that the above-mentioned computer device can be an external hardware device that can run independently, or it can be a hardware device that comes with the intelligent monitoring system in the cage chicken house.

[0063] Reference Figure 2 Based on the aforementioned hardware environment, this embodiment further provides an intelligent monitoring method in a caged chicken house, the method comprising the following steps:

[0064] Step S10, inputting the environmental parameter values ​​into a first multidimensional data processing device for data analysis and prediction to obtain a first prediction result; inputting the behavior data into a second multidimensional data processing device for data analysis and prediction to obtain a second prediction result; wherein the environmental parameter values ​​include humidity values, temperature values, target gas concentration values, lighting values ​​and feed consumption data, which are obtained by monitoring the environment in the cage chicken house through environmental sensors; the behavior data include behavior video data and behavior image data, which are obtained by monitoring the behavior of all chickens in the cage chicken house through a behavior data acquisition device.

[0065] It should be noted that the step of inputting the environmental parameter value into a first multidimensional data processing device for data analysis and prediction to obtain a first prediction result; and inputting the behavior data into a second multidimensional data processing device for data analysis and prediction to obtain a second prediction result in step S10 includes:

[0066] Step S11: Input the environmental parameter value into the time series prediction model for data analysis and prediction to obtain a first prediction result. The first prediction result includes temperature fluctuation data, humidity fluctuation data, gas concentration change data, lighting change data and feed consumption data at a specific time in the future.

[0067] Specifically, step S11 may include the following steps:

[0068] Step S111, seasonally decomposing the environmental parameter to obtain a first environmental parameter and a second environmental parameter; wherein the first environmental parameter has seasonal changes or periodic changes; and the second environmental parameter has no seasonal changes or periodic changes;

[0069] Step S112: input the first environmental parameter into the Prophet time series prediction model for data analysis and prediction to obtain a first environmental prediction parameter value;

[0070] Step S113: input the second environmental parameter into the ARIMA time series prediction model for data analysis and prediction to obtain a second environmental prediction parameter value.

[0071] The advantages of the scheme described in step S11 are mainly concentrated on the accuracy and applicability of environmental parameter prediction. Specifically:

[0072] 1) By seasonally decomposing the environmental parameters, the first environmental parameter is separated into seasonal changes or cyclical changes. This decomposition can more accurately capture the cyclical trends in the data, such as seasonal changes or cyclical changes, and thus help to more accurately predict future data trends.

[0073] 2) The Prophet model is suitable for processing time series data with strong seasonality and periodicity, and can automatically detect and fit these patterns to provide robust forecasting results. Inputting the first environmental parameter into the Prophet model and using its optimized seasonal analysis capabilities for data analysis and forecasting can improve the accuracy and reliability of the forecast. For the second environmental parameter, it does not have obvious seasonal or periodic changes, so it is more appropriate to choose the ARIMA model for forecasting. The ARIMA model is suitable for stationary or differentially stationary time series data, and can effectively capture the long-term trends and cyclical changes of the data, thereby generating reliable forecasting results. It can be seen that by selecting different forecasting models, this application can better adapt to the characteristics and data patterns of different environmental parameters, thereby improving the pertinence and accuracy of the forecast. The forecast results of the Prophet and ARIMA models are integrated to comprehensively consider the impact of seasonal and cyclical changes on environmental parameters, thereby providing more comprehensive and reliable forecast parameter values.

[0074] In summary, the above scheme utilizes seasonal decomposition and different time series prediction models, which can provide accurate and reliable data analysis and prediction capabilities in environmental parameter prediction, helping decision makers to perform accurate environmental management and regulation in practical applications.

[0075] In a further technical solution, the seasonal decomposition can help decompose time series data into trend, seasonal, cyclical and random components. By extracting the seasonal component separately, we can more accurately analyze and predict the long-term trend in the time series. This decomposition can eliminate the impact of seasonal factors on trend analysis, thereby improving the accuracy of prediction.

[0076] For example, in chicken house management, seasonal changes have a significant impact on the growth, egg production and health of chickens. For example, factors such as temperature, humidity and sunshine will change in different seasons, which directly affects the physiological state and behavioral habits of chickens. Through seasonal decomposition, we can more clearly understand how these environmental factors change with the seasons, so as to adjust management strategies in a targeted manner.

[0077] Seasonal decomposition helps predict the changing trends of environmental parameters in the chicken house. For example, by analyzing the seasonal patterns in historical data, it is possible to predict the changing trends of temperature, humidity or other environmental parameters in the chicken house in a future season. This is very important for reasonable feeding plans, feed supply adjustments and health management.

[0078] The adaptability and physiological response of chickens to the environment will change with seasonal changes. By seasonally decomposing environmental data, the period and amplitude of seasonal changes can be identified, and the environmental control in the chicken house can be optimized. For example, in cold seasons, insulation measures can be increased or lighting time can be adjusted to provide a suitable growth environment; in hot seasons, ventilation and cooling measures need to be adjusted.

[0079] In addition, understanding seasonal changes can help farmers manage resources effectively and reduce costs. Based on the prediction of environmental parameters in different seasons, the use of feed and water resources can be reasonably planned to avoid waste of resources and excessive environmental pressure.

[0080] Therefore, seasonal decomposition of chicken house environmental data can improve feeding efficiency, optimize resource utilization, and improve the growth and health of chickens.

[0081] The seasonal decomposition can be obtained by any of the following methods:

[0082] 1) Moving average method: The moving average method is one of the simplest and most direct seasonal decomposition methods. It smoothes the time series data by calculating the average value of the data points within a rolling window to identify and extract seasonal patterns. The moving average method is particularly suitable for seasonal patterns with relatively stable cycles.

[0083] 2) Decompose time series data into trend, seasonal and random components. Common decomposition methods include additive model and multiplicative model. The additive model assumes that the fluctuation of seasonal components at different time points is fixed, while the multiplicative model assumes that the seasonal components grow as the trend grows.

[0084] 3) Seasonal Index Method: The seasonal index method calculates the average of the data points within each seasonal cycle and compares it to the overall average to determine the magnitude of seasonal changes. This method can help identify seasonal patterns and the size of seasonal effects.

[0085] Through these implementation methods, the seasonal components in time series data can be effectively analyzed and utilized, thereby improving the accuracy and practicality of data analysis and prediction.

[0086] Step S12: input the behavior data into the behavior recognition model and the target detection model for data analysis and prediction to obtain a second prediction result. The second prediction result includes the activity pattern data, appetite change data and abnormal behavior data of the target chicken.

[0087] The step of inputting the behavior data into the behavior recognition model and the target detection model for data analysis and prediction to obtain a second prediction result in step S12 includes:

[0088] Step S121, after pre-processing the behavior data, input it into a behavior recognition model for data analysis to screen and obtain data of abnormal chickens; the abnormal chickens have abnormal behaviors.

[0089] The above preprocessing steps can eliminate noise and anomalies in the data, ensure the high quality of data input into the behavior recognition model, and improve the accuracy of subsequent analysis and prediction. The behavior recognition model can analyze a large amount of behavior data, quickly and accurately filter out the data of chickens with abnormal behavior, and improve the efficiency of anomaly detection. Through model analysis, abnormal behaviors that are significantly different from normal behavior patterns can be accurately identified, helping to detect and deal with abnormal situations early.

[0090] For example, a large amount of behavioral data on chickens can be collected, including activity patterns, appetite, sleep time, etc.; behavioral recognition models can be used to analyze these data to identify abnormal behaviors, such as illness, injury, or abnormal aggressive behavior, to help farmers investigate and handle abnormal situations in a timely manner to ensure the health and comfort of the chickens.

[0091] Step S122, inputting the data of the abnormal chicken into the target detection model, locating the abnormal chicken, and obtaining the real-time position of the abnormal chicken; and based on the real-time position of the abnormal chicken, monitoring its behavior in real time to obtain the real-time behavior data of the abnormal chicken.

[0092] The target detection model can quickly and accurately locate the position of abnormal chickens, and realize timely monitoring and response to chicken behavior. Based on the target detection model, the location information of each abnormal chicken can be accurately obtained, which not only helps real-time management, but also facilitates subsequent operations. After obtaining the real-time location, the behavioral changes of abnormal chickens can be continuously monitored and tracked, and their behavioral data can be updated in real time to provide real-time input for subsequent analysis.

[0093] For example, if the monitoring system finds that a chicken exhibits abnormal behavior, such as unusual attacks on fellow chickens or unusual activity patterns, the target detection model captures the location of the chicken through the camera and immediately notifies the breeder, who can quickly rush to the scene to confirm and make necessary interventions to prevent injuries or disease transmission among the chickens.

[0094] Step S123: input the real-time behavior data of the abnormal chicken into the long short-term memory network model for analysis and prediction to obtain a second prediction result.

[0095] The long short-term memory (LSTM) network model is suitable for processing time series data and can capture the behavior patterns and changing trends of abnormal chickens. By analyzing real-time behavior data with the LSTM model, it is possible to predict the behavior that abnormal chickens may exhibit in the future and help take necessary measures in advance. Based on the prediction results of the model, the feeding strategy and management measures can be dynamically adjusted to minimize the adverse effects of abnormal behavior.

[0096] For example, if a chicken shows frequent abnormal aggressive behavior, its future behavior trend and possible impacts need to be predicted. The LSTM model analyzes the chicken's real-time behavior data and predicts its possible behavior patterns and potential risks. Based on this, the breeder can adjust the environment or management methods in the chicken house to reduce conflicts and injuries between chickens, and improve breeding efficiency and chicken welfare.

[0097] Through the above steps, the efficiency of chicken house management can be improved, such as:

[0098] Farmers use intelligent monitoring systems to monitor the behavior of chickens in the chicken house. By analyzing the data, the system found that the activity pattern of a certain chicken was obviously different, such as frequent solitary activities or not eating for a long time. The system identified these abnormal behaviors and recorded them. During the monitoring, a chicken was found to have abnormal behavior of continuously attacking its companions. The system uses cameras and target detection algorithms to quickly lock the location of the chicken, so that farmers can arrive at the scene in time, observe and take necessary treatment measures, such as isolation or treatment. During the monitoring process, the system found that a chicken showed signs of illness, such as abnormal walking or loss of appetite. The LSTM model analyzes its real-time behavior data and predicts that the chicken may develop a more serious disease. According to the prediction results, the farmers adjust the feeding conditions in time, strengthen monitoring or take medical measures to ensure the health of the chickens.

[0099] Through the above steps, the system can identify and locate abnormal behaviors at an early stage and warn farmers in advance; real-time monitoring and positioning ensure rapid response and handling of abnormal chickens; through predictive models, feeding management strategies are optimized and the risk of chicken losses and disease transmission is reduced.

[0100] Step S20: Calculate the first prediction result and the second prediction result according to a preset algorithm to obtain a third prediction result.

[0101] The step of obtaining a third prediction result after calculating the first prediction result and the second prediction result according to a preset algorithm in step S20 includes:

[0102] Step S21: pre-process the first prediction result and the second prediction result, and input them into a trained decision tree model, so that the decision tree model outputs prediction information;

[0103] Step S22, the prediction information includes production prediction information, environmental prediction information, feed consumption prediction information, flock behavior prediction information and flock health status prediction information. The production prediction information includes egg production rate or hatchability.

[0104] In the above steps S21-S22, step S21 describes how to preprocess the first prediction result and the second prediction result and pass them as input to the trained decision tree model. The preprocessing process usually includes steps such as data cleaning, feature selection, normalization or standardization to ensure the quality and consistency of the input data. The decision tree model receives the preprocessed data and makes decisions and predicts the output according to its internal tree structure. Its advantages are:

[0105] Using a decision tree model, you can make accurate predictions based on a pre-trained data set. The decision tree analyzes the characteristics of the input data, makes branching judgments step by step, and finally gives the corresponding prediction information. This method can provide high accuracy and reliability when dealing with complex prediction tasks.

[0106] Decision tree models are easy to understand and interpret because they visually display each decision path and branching conditions. This enables users to understand how the model works and make adjustments and optimizations as needed.

[0107] Step S22 lists a variety of prediction information, such as production prediction, environmental prediction, feed consumption prediction, flock behavior prediction and flock health status prediction. The decision tree model can adapt to and predict these diverse information according to the different characteristics of the input data, making it widely applicable in fields such as agriculture and breeding.

[0108] The decision tree model can be applied to multiple prediction targets in the technical solution described in this application, such as:

[0109] Production performance prediction: Predicting the egg production rate, hatchability or weight growth rate of broilers. This information is very important for farmers to manage chicken houses, optimize feed rations and adjust production plans.

[0110] Environmental control optimization: predicting the internal environmental parameters of the chicken house, such as temperature, humidity, carbon dioxide concentration, etc. These predictions can help farmers adjust the ventilation, heating or cooling systems in time to provide the best growth environment.

[0111] Feed Consumption Forecasting: Predict feed consumption for the flock within the house. This helps farmers plan feed purchases and storage efficiently, avoid waste and ensure the birds get enough nutrition.

[0112] Chicken behavior prediction: Predict chicken behavior patterns, such as chicken activity patterns, flocking behavior, etc. This is very helpful for optimizing chicken house layout, improving chicken welfare and reducing stress reactions.

[0113] Disease outbreak prediction: Predict the possible outbreak or spread of disease in the chicken house. By monitoring the health status of the chickens and environmental parameters, the decision tree model can help detect potential health problems early and take preventive measures.

[0114] Chicken house operating cost forecast: Forecast the cost of chicken house operation, including expenses on feed, labor, energy, etc. This helps farmers to make budgets and financial plans.

[0115] These prediction goals can be achieved by collecting a large amount of historical data (such as internal environment data of the chicken house, chicken health data, feed usage data, etc.) to train the decision tree model, and use the predictive ability of the model to optimize the management and operation efficiency of the chicken house.

[0116] If the prediction target is the egg production rate of the chickens, the first prediction result of step S20 may include predictions of factors such as feed type, temperature and humidity of the breeding environment, and the second prediction result may include predictions of the behavior pattern and health status of the chickens.

[0117] Step S21: preprocessing the first prediction result and the second prediction result, which may include converting these prediction results into standardized numerical forms to facilitate the input of the decision tree model. For example, normalizing the predicted temperature and humidity data and converting the behavior prediction into classification features.

[0118] Step S22: The decision tree model will operate here based on the preprocessed data and output production prediction information, that is, the expected egg production rate of the flock. This information can help farmers make better decisions in feeding management, such as adjusting feed ratios, optimizing environmental control, or adjusting health monitoring strategies to maximize output and ensure the health of the flock and the egg production rate of the flock.

[0119] It should be noted that the decision tree model is trained through the following steps: obtaining historical environmental parameter data and historical chicken behavior data in the chicken house; preprocessing the historical environmental parameter data and historical chicken behavior data in the chicken house to obtain standardized training data; constructing an initial decision tree model through a decision tree algorithm, and inputting the standardized training data into the initial decision tree model for training to obtain a trained decision tree model.

[0120] In a further technical solution, after the step of outputting prediction information by the decision tree model, the following steps are further included:

[0121] Inputting the forecast information into a decision support system to obtain management suggestion information output by the decision support system; the decision support system includes judgment conditions corresponding to each forecast information;

[0122] When the processing level of each of the prediction information is less than or equal to level 1, the management suggestion information does not require manual intervention;

[0123] When the processing level of each of the prediction information is greater than level one and less than or equal to level two, the management suggestion information is for automatic control equipment to make adjustments;

[0124] When the processing level of each of the prediction information is greater than level 2 and less than or equal to level 3, the management suggestion information requires manual intervention.

[0125] In the above steps, the decision tree model is used to predict various data in the chicken house, such as predicting the temperature change in the next hour, the change of ammonia concentration, etc. The prediction information may include parameters such as expected temperature, humidity, and ammonia concentration.

[0126] The prediction information output by the decision tree model is input into the decision support system, which analyzes the prediction information and preset judgment conditions and generates corresponding management suggestion information.

[0127] Generate management advice information based on the processing level of the forecast information (which can be understood here as the accuracy and urgency of the forecast):

[0128] If the processing level of the forecast information is less than or equal to level 1, it indicates that the forecast result is highly credible and the problem is not urgent. It can be recommended that no human intervention is required, for example, the current temperature and humidity are within the ideal range.

[0129] If the processing level of the forecast information is greater than level one and less than or equal to level two, it indicates that the forecast result is relatively accurate and there are certain problems or changing trends. It is recommended that the automatic control equipment be adjusted, such as slightly higher temperature or slightly increased ammonia concentration.

[0130] If the processing level of the forecast information is greater than level 2 and less than or equal to level 3, it indicates that the forecast result is not accurate enough or the problem is already significant, and manual intervention is recommended, such as a rapid rise in temperature or an excessive ammonia concentration.

[0131] It can be seen that the above decision tree model can quickly generate forecast information based on real-time data, helping chicken house managers to take timely actions. The decision support system can automatically generate management suggestions based on forecast information, reduce the need for manual intervention, and improve management efficiency. According to the processing level classification, managers can quickly understand the urgency of the forecast results, respond and adjust in time, and improve emergency response capabilities.

[0132] For example, if the decision tree model predicts that the ammonia concentration in the chicken house is about to exceed the standard (the treatment level is greater than level 2), the decision support system can immediately issue management advice information requiring human intervention, reminding managers to check the ventilation system of the chicken house or start the air purification equipment.

[0133] For example, if the forecast shows that the temperature and humidity in the chicken house remain within the ideal range (treatment level is less than or equal to level 1), the management recommendation information can be that no manual intervention is required and the current setting state can be maintained. Such application scenarios demonstrate the actual value and application advantages of decision tree models and decision support systems in chicken house management.

[0134] The decision support system is trained by the following steps:

[0135] Collecting chicken house historical data; the chicken house historical data includes chicken house environment historical parameters, chicken flock behavior historical data, feed consumption historical data, chicken flock health status historical data and production historical data;

[0136] After preprocessing the chicken house historical data (such as cleaning data, standardizing data format, etc.), extracting characteristic information of the chicken house historical data to obtain target chicken house historical data; the characteristic information includes average values ​​and change rates of chicken house environmental parameters, feed consumption trends, frequency of chicken flock activity patterns, chicken flock health status change trends, and production average values;

[0137] The target chicken house historical data is divided into a training set and a validation set. After the initial machine learning model is trained based on the training set, the hyperparameters of the initial machine learning model (such as a regression model, decision tree, random forest or neural network, etc.) are optimized based on the validation set to obtain a target decision support system.

[0138] Step S30: Send the third prediction result to the central control device, so that the central control device displays the third prediction result to the user.

[0139] In the above technical solution, the third prediction result predicted by the decision tree model, for example, predicts that the temperature and humidity in the chicken house will exceed the safe range, or the ammonia concentration has reached a level that requires human intervention; these prediction results are sent to the central control device of the chicken house through the network or other communication methods. After receiving the prediction results, the central control device displays the relevant information to the manager or staff of the chicken house according to the set display interface or user interface. The displayed information may include specific predicted values ​​(such as temperature, humidity, ammonia concentration), prediction level (treatment level) and recommended countermeasures (requiring human intervention or automatic adjustment).

[0140] It can be seen that by sending the third prediction result to the central control device and displaying it to the user, the manager can understand the status and prediction of the chicken house in real time and respond in time. At the same time, the prediction results displayed by the central control device provide remote decision-making support for managers, so that even if they are not on site, they can understand the operating status of the chicken house and the measures to be taken in time. In some preferred schemes, if the prediction results show that manual intervention is required (the processing level is greater than level 2), the central control device can automatically issue an alarm or prompt to remind the manager to take necessary measures, such as adjusting the ventilation system or adding heat dissipation equipment. In addition, displaying the third prediction result not only helps managers to understand the operation of the chicken house in real time, but also optimizes the chicken house management strategy and facility design through long-term data analysis and trend prediction.

[0141] For example, if the decision tree model predicts that the temperature in the chicken house is about to exceed the safe range, the central control device will display this information to the manager after receiving it. The display content may include the current temperature, the predicted rising trend, and the recommended adjustment measures (such as turning on the cooling equipment). This real-time display and feedback of the prediction results help managers effectively manage and optimize the environmental conditions of the chicken house in different time periods to ensure the health and production efficiency of the chickens.

[0142] Step S40: Obtain the control instruction fed back by the user, and intelligently manage the physical environment management device in the chicken house based on the control instruction.

[0143] As can be seen, the chicken house management system can receive feedback from users, which may be manually entered instructions, set preference parameters, or response requirements for specific conditions. User feedback can be entered and submitted through a mobile application, a web interface, or a dedicated control panel.

[0144] Once the system obtains the user's control instructions, the intelligent chicken house management device will perform corresponding operations according to these instructions and adjust the physical environment management devices in the chicken house. These management devices may include temperature control devices (such as heaters, ventilation systems), humidity control devices (such as humidifiers, dehumidifiers), light adjustment devices or other monitoring and adjustment devices.

[0145] Users can adjust the parameters of the physical environment based on actual needs and changes in the chicken house. For example, users can set the temperature range and increase the temperature setting of the heater when cold weather arrives. The intelligent management system can respond to user instructions immediately and adjust the environment according to real-time data and user settings to ensure the comfort and health of the chickens. Intelligent management based on user feedback can help optimize energy use, such as reasonably controlling the operating time and intensity of heating or ventilation equipment based on forecasts or user instructions without affecting the comfort of the chickens. Users can manage the chicken house through remote control without on-site operation. This is especially important when the farmer or chicken house manager needs to be away for a long time.

[0146] For example, if the user sets a temperature adjustment command in the remote monitoring system, when the system receives this command, the intelligent management device will automatically adjust the temperature setting of the heating equipment to ensure that the temperature in the chicken house remains within the ideal range. This intelligent management not only improves management efficiency, but also improves the environmental control ability of the chicken house, which is beneficial to the growth and health of the chickens.

[0147] In summary, the technical solution described in this application has the following advantages:

[0148] Intelligent management and real-time monitoring: The technical solution uses a decision tree model to predict and analyze environmental parameters in the chicken house, such as temperature, humidity, and ammonia concentration. This real-time monitoring can help managers identify potential problems in a timely manner and take necessary measures to ensure the health and production efficiency of the chickens.

[0149] Remote control and ease of operation: The technical solution allows users to obtain prediction results and perform environmental management through remote control devices. No matter where the user is, they can adjust the physical environment in the chicken house through mobile devices or computers, which improves the convenience and flexibility of operation.

[0150] Personalized and data-driven decision-making: Users can set personalized control instructions and management strategies based on their own needs and predicted results. This data-driven decision support system not only improves management efficiency, but also optimizes resource utilization and energy consumption, promoting sustainable development of the farm.

[0151] Safety and early warning function: The technical solution has an early warning function. Once the system detects potential problems or abnormal conditions (such as excessive temperature or excessive ammonia concentration), it will immediately alert the user and display detailed information. This timely warning can help managers take countermeasures early to avoid possible losses or health risks.

[0152] Energy saving and environmental protection: Through intelligent management, technical solutions can optimize energy efficiency and reduce unnecessary energy consumption, thereby reducing the cost of chicken house operation and reducing the impact on the environment, in line with the concept of sustainable development.

[0153] On the other hand, Figure 3 As shown, in order to solve the above technical problems, the embodiment of the present application also provides: an intelligent monitoring system in a caged chicken house, comprising:

[0154] A first multi-dimensional data processing device is used to input the environmental parameter values ​​into the first multi-dimensional data processing device for data analysis and prediction to obtain a first prediction result; the environmental parameter values ​​include humidity values, temperature values, target gas concentration values, lighting values ​​and feed consumption data, which are obtained by monitoring the environment in the cage chicken house through an environmental sensor;

[0155] A second multi-dimensional data processing device is used to input the behavior data into the second multi-dimensional data processing device for data analysis and prediction to obtain a second prediction result; the behavior data includes behavior video data and behavior image data, which are obtained by monitoring the behavior of all chickens in the cage chicken house through the behavior data acquisition device;

[0156] A prediction device, configured to calculate the first prediction result and the second prediction result according to a preset algorithm to obtain a third prediction result; and send the third prediction result to a user platform so that the user platform displays the third prediction result to a user;

[0157] A central control device, used to obtain control instructions fed back by the user platform, and intelligently manage the physical environment management device in the chicken house based on the control instructions;

[0158] The intelligent monitoring system in the caged chicken house also includes a user platform, which is used to receive the third prediction result sent by the prediction device, and generate a control instruction based on the third prediction result; and feed back the control instruction to the central control device, so that the central control device can intelligently manage the physical environment management device in the chicken house based on the control instruction.

[0159] The system monitors and analyzes environmental parameters such as humidity, temperature, gas concentration, lighting, feed consumption, etc. through the first multi-dimensional data processing device. The real-time monitoring and prediction capabilities of these parameters help managers accurately grasp the environmental conditions inside the chicken house, adjust and optimize management strategies in a timely manner, and improve the growth and health of the chickens.

[0160] The second multi-dimensional data processing device is responsible for analyzing and predicting the chickens’ behavioral data, including video and image data. By monitoring the behavioral data, the activity patterns, health status, and interactions of the chickens can be understood in real time. This detailed monitoring helps to detect any abnormal behaviors or health problems early, prevent the occurrence of diseases in advance, and thus reduce losses.

[0161] The prediction device integrates the first prediction result and the second prediction result, and generates a third prediction result by applying a preset algorithm. These prediction results not only include a comprehensive analysis of environmental parameters and behavioral data, but also take into account possible trends and changes in the future. After the third prediction result is sent to the user platform, managers can make more accurate management decisions based on this information, optimize breeding conditions, and improve production efficiency.

[0162] The user platform receives and displays the third prediction result, generates control instructions and feeds them back to the central control device. This two-way information flow mechanism allows managers to remotely monitor and control the physical environment in the chicken house at any time, respond to emergencies or user needs, and improve management flexibility and responsiveness.

[0163] That is, it integrates intelligent environmental management and health monitoring systems to effectively optimize energy consumption and resource utilization efficiency. Through accurate data analysis and prediction, the system can help farmers use resources more effectively, reduce waste, reduce operating costs, and improve the sustainable development capabilities of the farm.

[0164] Based on the same inventive concept as the aforementioned embodiment, this embodiment provides a computer-readable storage medium, on which a computer program is stored, and a processor executes the computer program to implement the aforementioned method.

[0165] In some embodiments, the computer readable storage medium may be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, flash memory, magnetic surface memory, optical disk, or CD-ROM; or various devices including one or any combination of the above memories. The computer may be various computing devices including intelligent terminals and servers.

[0166] The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. An intelligent monitoring method in a caged chicken house, characterized in that: The method comprises the following steps: Input the environmental parameter values ​​into the time series prediction model for data analysis and prediction to obtain a first prediction result; wherein the environmental parameter values ​​include humidity values, temperature values, target gas concentration values, lighting values ​​and feed consumption data, which are obtained by monitoring the environment in the cage chicken house through environmental sensors; the specific steps of obtaining the first prediction result are: seasonally decomposing the environmental parameters to obtain the first environmental parameter and the second environmental parameter; wherein the first environmental parameter has seasonal or periodic changes; the second environmental parameter does not have seasonal or periodic changes; the first environmental parameter is input into the Prophet time series prediction model for data analysis and prediction to obtain the first environmental prediction parameter value; the second environmental parameter is input into the ARIMA time series prediction model for data analysis and prediction to obtain the second environmental prediction parameter value; The behavior data is input into the behavior recognition model and the target detection model for data analysis and prediction to obtain a second prediction result; wherein the behavior data includes behavior video data and behavior image data, which are obtained by monitoring the behavior of all chickens in the caged chicken house through a behavior data acquisition device; the first prediction result includes temperature fluctuation data, humidity fluctuation data, gas concentration change data, lighting change data and feed consumption data at a specific time in the future; the second prediction result includes activity pattern data, appetite change data and abnormal behavior data of the target chicken; the specific steps of obtaining the second prediction result are: after pre-processing the behavior data, input it into the behavior recognition model for data analysis to screen and obtain data of abnormal chickens; the abnormal chickens have abnormal behaviors; the data of the abnormal chickens is input into the target detection model to locate the abnormal chickens and obtain the real-time position of the abnormal chickens; and based on the real-time position of the abnormal chickens, their behaviors are monitored in real time to obtain the real-time behavior data of the abnormal chickens; the real-time behavior data of the abnormal chickens is input into the long short-term memory network model for analysis and prediction to obtain the second prediction result; After calculating the first prediction result and the second prediction result according to a preset algorithm, a third prediction result is obtained; the specific steps of obtaining the third prediction result are: after preprocessing the first prediction result and the second prediction result, input them into a trained decision tree model, so that the decision tree model outputs third prediction information; the third prediction information includes production prediction information, environmental prediction information, feed consumption prediction information, chicken group behavior prediction information and chicken group health status prediction information; sending the third prediction result to the central control device so that the central control device displays the third prediction result to the user; The control instructions fed back by the user are obtained, and based on the control instructions, the physical environment management device in the chicken house is intelligently managed.

2. The intelligent monitoring method in a caged chicken house according to claim 1, characterized in that: After the step of outputting prediction information by the decision tree model, the method further includes: Inputting the forecast information into a decision support system to obtain management suggestion information output by the decision support system; the decision support system includes judgment conditions corresponding to each forecast information; When the processing level of each of the prediction information is less than or equal to level 1, the management suggestion information does not require manual intervention; When the processing level of each of the prediction information is greater than level one and less than or equal to level two, the management suggestion information is for automatic control equipment to make adjustments; When the processing level of each of the prediction information is greater than level 2 and less than or equal to level 3, the management suggestion information requires manual intervention.

3. The intelligent monitoring method in a caged chicken house according to claim 1, characterized in that: The decision tree model is trained by the following steps: Obtain historical environmental parameter data and historical chicken behavior data in the chicken house; Preprocessing the historical environmental parameter data and the historical chicken behavior data in the chicken house to obtain standardized training data; An initial decision tree model is constructed by a decision tree algorithm, and the standardized training data is input into the initial decision tree model for training to obtain a trained decision tree model.

4. The intelligent monitoring method in a caged chicken house according to claim 2, characterized in that: The decision support system is trained by the following steps: Collecting chicken house historical data; the chicken house historical data includes chicken house environment historical parameters, chicken flock behavior historical data, feed consumption historical data, chicken flock health status historical data and production historical data; After preprocessing the chicken house historical data, extract characteristic information of the chicken house historical data to obtain the target chicken house historical data; the characteristic information includes the average value and change rate of chicken house environmental parameters, feed consumption trend, frequency of chicken group activity patterns, chicken group health status change trend and production average value; The target chicken house historical data is divided into a training set and a validation set. After the initial machine learning model is trained based on the training set, the hyperparameters of the initial machine learning model are optimized based on the validation set to obtain a target decision support system.

5. An intelligent monitoring system for caged chicken houses, characterized in that: include: A first multidimensional data processing device is used to input environmental parameter values ​​into a time series prediction model for data analysis and prediction to obtain a first prediction result; wherein the environmental parameter values ​​include humidity values, temperature values, target gas concentration values, lighting values ​​and feed consumption data, which are obtained by monitoring the environment in the cage chicken house through environmental sensors; the specific steps of obtaining the first prediction result are: seasonally decomposing the environmental parameters to obtain a first environmental parameter and a second environmental parameter; wherein the first environmental parameter has seasonal or periodic changes; the second environmental parameter has no seasonal or periodic changes; the first environmental parameter is input into a Prophet time series prediction model for data analysis and prediction to obtain a first environmental prediction parameter value; the second environmental parameter is input into an ARIMA time series prediction model for data analysis and prediction to obtain a second environmental prediction parameter value; The second multidimensional data processing device is used to input the behavior data into the behavior recognition model and the target detection model for data analysis and prediction to obtain a second prediction result; wherein the behavior data includes behavior video data and behavior image data, which are obtained by monitoring the behavior of all chickens in the caged chicken house through the behavior data acquisition device; the first prediction result includes temperature fluctuation data, humidity fluctuation data, gas concentration change data, lighting change data and feed consumption data at a specific time in the future; the second prediction result includes activity pattern data, appetite change data and abnormal behavior data of the target chicken; the specific steps of obtaining the second prediction result are: after pre-processing the behavior data, input it into the behavior recognition model for data analysis to screen and obtain data of abnormal chickens; the abnormal chickens have abnormal behaviors; input the data of the abnormal chickens into the target detection model, locate the abnormal chickens, and obtain the real-time position of the abnormal chickens; and based on the real-time position of the abnormal chickens, monitor their behaviors in real time to obtain the real-time behavior data of the abnormal chickens; input the real-time behavior data of the abnormal chickens into the long short-term memory network model for analysis and prediction to obtain the second prediction result; A prediction device, used for calculating the first prediction result and the second prediction result according to a preset algorithm to obtain a third prediction result; the specific steps of obtaining the third prediction result are: pre-processing the first prediction result and the second prediction result, and inputting them into a trained decision tree model, so that the decision tree model outputs third prediction information; the third prediction information includes production prediction information, environmental prediction information, feed consumption prediction information, chicken group behavior prediction information and chicken group health status prediction information; A central control device, used to obtain control instructions fed back by the user platform, and based on the control instructions, intelligently manage the physical environment management device in the chicken house; The intelligent monitoring system in the caged chicken house also includes a user platform, which is used to receive the third prediction result sent by the prediction device, and generate a control instruction based on the third prediction result; and feed back the control instruction to the central control device, so that the central control device can intelligently manage the physical environment management device in the chicken house based on the control instruction.

Images