Broiler breeding abnormity identification method and system based on drinking water consumption data
By installing monitoring equipment on the drinking water pipes of broiler farms, collecting and processing data in real time, and using neural network models to identify abnormal situations, the problem that the existing technology cannot fully cover and accurately judge the health status of broiler chickens, and the abnormality of broiler breeding is achieved Timely identification and monitoring have improved the scientificity and efficiency of breeding management.
Patent Information
- Application Number
- CN202411051827.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-01
- Publication Date
- 2024-11-12
AI Technical Summary
The prior art cannot fully cover abnormal situations in broiler breeding, and cannot accurately judge the health status of pets, resulting in incomplete or inaccurate monitoring results.
Install drinking water monitoring equipment on the drinking water pipelines of broiler farms, collect data in real time and perform pre-processing, extract drinking water consumption characteristics, use neural network models for training and abnormal identification, identify abnormal situations and issue alarms.
Timely identification and accurate monitoring of broiler breeding abnormalities has been achieved, the potential impact of abnormalities on breeding has been reduced, the scientificity and efficiency of monitoring has been improved, and the farm managers have helped timely measures be taken.
Smart Images

Figure CN118940184A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of smart breeding technology, and in particular to a broiler breeding abnormality identification method and system based on drinking water consumption data. Background Art
[0002] As an important part of modern agriculture, broiler farming may encounter various abnormalities during the farming process, such as disease outbreaks and environmental changes, which directly affect the health status and farming efficiency of broilers. Therefore, it is of great significance to identify abnormalities in broiler farming based on real-time data monitoring. Drinking water consumption is one of the important monitoring indicators in broiler farming, because the amount of drinking water directly reflects the health status and growth status of broilers.
[0003] The invention patent with application number 202310390132.9 discloses a method for monitoring the health of pet drinking water. This technology performs water quality testing, filtration and purification, water volume reminders, and filter element replacement reminders; performs initial monitoring and recording of pet drinking water, determines the healthy drinking water interval and multiple common drinking water times; performs normalized monitoring and recording of pet drinking water, compares it with the healthy drinking water interval, and determines whether there is any drinking water abnormality; when there is a drinking water abnormality, controls the drinking water flow at multiple common drinking water times; determines whether the drinking water can return to normal, and when it cannot return to normal, issues a pet health abnormality alarm. It is possible to monitor and record pet drinking water, control the drinking water flow when there is a drinking water abnormality, and determine whether it can return to normal, and when it cannot return to normal, issues a pet health abnormality alarm, so as to determine the health status of the pet through pet drinking machine monitoring, so as to detect the pet's health abnormality as early as possible and issue an abnormality alarm.
[0004] However, the monitoring and control capabilities of the pet drinking fountain are limited by the stability and accuracy of the equipment, and cannot cover all possible abnormal conditions or accurately determine the health status of the pet. This will lead to incomplete or inaccurate monitoring results of abnormal conditions of animals. Summary of the invention
[0005] In order to overcome the deficiencies of the prior art, the object of the present invention is to provide a method and system for identifying broiler farming abnormalities based on drinking water consumption data.
[0006] To achieve the above object, the present invention provides the following solutions:
[0007] A broiler farming abnormality identification method based on drinking water consumption data, comprising:
[0008] Install drinking water monitoring equipment on the drinking water pipes of the target breeding farm, and use the drinking water monitoring equipment to collect drinking water volume data of each breeding zone in real time and record the corresponding timestamp;
[0009] Preprocessing the collected water intake data and the corresponding timestamp to obtain preprocessed data;
[0010] extracting drinking water consumption features associated with abnormal conditions from the preprocessed data;
[0011] Use the preset historical data to train the initial neural network to obtain a trained anomaly recognition model;
[0012] The drinking water consumption feature is input into the abnormality recognition model to obtain an abnormality recognition result.
[0013] Preferably, the drinking water consumption characteristics include: changing trend of drinking water volume, average daily drinking water volume, drinking water peak and valley values, and drinking water volume volatility.
[0014] Preferably, the preprocessing process includes: removing noise data, time synchronization and data smoothing.
[0015] Preferably, extracting drinking water consumption features related to abnormal situations from the preprocessed data includes:
[0016] The pre-processed data are grouped by day, and the average daily water intake of each group of pre-processed data is calculated;
[0017] Performing linear regression on the preprocessed data to fit a drinking water trend line;
[0018] Extracting the regression coefficient of the drinking water trend line as the change trend;
[0019] Calculate the maximum and minimum values of the preprocessed data within a preset time period, and determine the maximum and minimum values as the drinking water peak and valley values;
[0020] The standard deviation of the numerical values of the preprocessed data is calculated to obtain the fluctuation of the water intake.
[0021] Preferably, the initial neural network is trained using preset historical data to obtain a trained anomaly recognition model, including:
[0022] Obtain the preset water intake change trend data set, daily average water intake data set and water intake peak and valley value data set;
[0023] Constructing an initial neural network; the initial neural network includes an input layer, an LSTM layer, a fully connected layer, and an output layer connected in sequence;
[0024] Obtain the drinking water consumption change trend data set, the daily average drinking water consumption data set, the drinking water consumption peak and valley value data set, and the drinking water consumption fluctuation data set in the historical database;
[0025] The water intake change trend data set, the daily average water intake data set, the water intake peak and valley value data set, and the water intake volatility data set are input into the initial neural network for training to obtain a trained abnormality recognition model.
[0026] Preferably, the abnormality identification result includes: abnormality type, abnormality time, abnormality partition and abnormality degree.
[0027] Preferably, the abnormal types include: excessive water intake, low water intake, drastic fluctuations in water intake, and abnormal drinking behavior caused by environmental or health problems.
[0028] A broiler farming abnormality identification system based on drinking water consumption data, comprising:
[0029] A data monitoring module is used to install a drinking water monitoring device on the drinking water pipe of the target breeding farm, and use the drinking water monitoring device to collect the drinking water volume data of each breeding zone in real time and record the corresponding timestamp;
[0030] A preprocessing module, used for preprocessing the collected water intake data and the corresponding timestamp to obtain preprocessed data;
[0031] A feature extraction module, for extracting drinking water consumption features related to abnormal situations from the preprocessed data;
[0032] A training module is used to train the initial neural network using preset historical data to obtain a trained anomaly recognition model;
[0033] The recognition module is used to input the drinking water consumption characteristics into the abnormality recognition model to obtain an abnormality recognition result.
[0034] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:
[0035] The present invention provides a method and system for identifying abnormalities in broiler farming based on drinking water consumption data, the method comprising: installing a drinking water monitoring device on the drinking water pipeline of the target farm, and using the drinking water monitoring device to collect the drinking water volume data of each farming partition in real time, and recording the corresponding timestamp; preprocessing the collected drinking water volume data and the corresponding timestamp to obtain preprocessed data; extracting drinking water consumption features related to abnormal situations from the preprocessed data; using preset historical data to train the initial neural network to obtain a trained abnormal recognition model; inputting the drinking water consumption features into the abnormal recognition model to obtain abnormal recognition results. The present invention quickly identifies abnormal situations in farming by real-time monitoring of drinking water consumption data, so as to take timely countermeasures. The present invention uses deep learning models such as neural networks for training and abnormality recognition, which can better mine the potential characteristics and laws of data and realize intelligent abnormality detection. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0037] Figure 1 A flow chart of a method provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0038] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0039] The purpose of the present invention is to provide a broiler farming anomaly identification method and system based on drinking water consumption data, which uses deep learning models such as neural networks for training and anomaly identification, can better mine the potential characteristics and laws of the data and realize intelligent anomaly detection.
[0040] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.
[0041] Figure 1 A flow chart of a method provided by an embodiment of the present invention, such as Figure 1 As shown, the present invention provides a broiler farming abnormality identification method based on drinking water consumption data, comprising:
[0042] Step 100: Install drinking water monitoring equipment on the drinking water pipe of the target breeding farm, and use the drinking water monitoring equipment to collect drinking water volume data of each breeding zone in real time, and record the corresponding timestamp;
[0043] Step 200: preprocessing the collected water intake data and the corresponding timestamp to obtain preprocessed data;
[0044] Step 300: extracting drinking water consumption features related to abnormal situations from the preprocessed data;
[0045] Step 400: Use preset historical data to train the initial neural network to obtain a trained anomaly recognition model;
[0046] Step 500: Input the drinking water consumption feature into the abnormality recognition model to obtain an abnormality recognition result.
[0047] In actual application, this embodiment deploys the trained model to the central data processing system to calculate the degree of deviation of the current water intake data from the normal mode in real time. When the deviation exceeds the preset threshold, it is judged as an abnormality. After identifying the abnormality, the system notifies the farm manager by SMS, email or other means. Based on the system prompts, the manager can further check and take corresponding measures. This embodiment can record all identified abnormal events, analyze them in detail, summarize the causes of the abnormalities and treatment methods, so as to continuously optimize the breeding management strategy. Through continuous data accumulation and model optimization, the accuracy and efficiency of abnormal identification are continuously improved to ensure that the system runs for a long time, stably and accurately.
[0048] Preferably, the drinking water consumption characteristics include: changing trend of drinking water volume, average daily drinking water volume, drinking water peak and valley values, and drinking water volume volatility.
[0049] Specifically, the average daily drinking water volume in this embodiment is defined as the average total daily drinking water volume of broilers in each breeding zone. The extraction method is: grouping the pre-processed data by day, calculating the total drinking water volume of each group of data, and calculating the average drinking water volume of each group of data.
[0050] This example uses Python for programming, and the sample code is as follows:
[0051] import pandas as pd
[0052] #Assume that the data is stored in a DataFrame with two columns: timestamp and water_consumption
[0053] data=pd.read_csv('water_data.csv')
[0054] data['timestamp']=pd.to_datetime(data['timestamp'])
[0055] #Group by day and calculate the average daily water intake
[0056] daily_avg=data.resample('D',on='timestamp').water_consumption.mean()
[0057] Furthermore, the variation trend of water intake in this embodiment is defined as the variation trend of water intake over time, which can be expressed by linear regression and other methods. The extraction method is: linear regression is performed on the time series data, a trend line is fitted, and then the regression coefficient is extracted as the trend feature.
[0058] This embodiment uses Python for programming, and the sample code is as follows:
[0059] from sklearn.linear_model import LinearRegression
[0060] import numpy as np
[0061] #Convert timestamp to numeric value
[0062] data['timestamp_numeric']=data['timestamp'].apply(lambdax:x.timestamp())
[0063] #Prepare linear regression model
[0064] X=data['timestamp_numeric'].values.reshape(-1,1)
[0065] y=data['water_consumption'].values
[0066] model=LinearRegression()
[0067] model.fit(X,y)
[0068] #Extract the regression coefficient (slope) as a trend feature
[0069] trend_slope = model.coef_[0]
[0070] Furthermore, the drinking water peak and valley values of this embodiment are defined as the peak and valley values of the drinking water amount, indicating the extreme conditions of the drinking water amount. The extraction method is: calculating the maximum and minimum values within a period of time.
[0071] Furthermore, the volatility of the amount of water consumed in this embodiment is defined as the degree of fluctuation of the amount of water consumed, which can be represented by statistics such as standard deviation. The extraction method is: calculating the standard deviation of the amount of water consumed within a period of time.
[0072] In addition, the drinking water consumption characteristics of broilers in this embodiment may also include: periodic characteristics of drinking water, which are defined as periodic changes in drinking water in different time periods, such as daily cycles, weekly cycles, etc. The extraction method is: using Fourier transform or wavelet transform to extract periodic characteristics.
[0073] Through the above method, this embodiment can extract a variety of features related to abnormal situations from the pre-processed water intake data, such as average daily water intake, water intake change trend, water intake peak and valley values, volatility and periodicity characteristics. These features can be used as inputs and further used for training and prediction of anomaly detection models. Different features can provide different perspectives to help more comprehensively understand the changing patterns and abnormal situations of water intake data.
[0074] Preferably, the preprocessing process includes: removing noise data, time synchronization and data smoothing.
[0075] Specifically, the noise data in this embodiment refers to abnormal data points caused by sensor failure, environmental interference or other abnormal factors. These data points may mislead the training and prediction of the model, so they need to be removed in the data preprocessing stage. The removal process is as follows:
[0076] 1. Identify noise: Use statistical methods (such as Z-score, IQR) or rule-based methods (such as the physical limit of water intake) to identify obvious noise data points. Z-score is specifically calculated for each data point. If the Z-score exceeds a certain threshold (such as 3), it is considered noise. IQR (interquartile range) is the IQR range of the calculated data. Any point that exceeds 1.5 times the IQR range is considered noise.
[0077] 2. Remove noise: Remove the identified noise data points from the data set or replace them with reasonable values (such as mean, neighboring values).
[0078] Furthermore, time synchronization in this embodiment refers to aligning data with different timestamps to ensure that the data is analyzed within the same time interval. This embodiment takes into account the importance of time synchronization between sensors, that is, time synchronization is particularly important for data fusion and analysis of multiple sensors. The process is as follows:
[0079] 1. Time standardization: convert the timestamps of all data into a unified time format (such as UTC).
[0080] 2. Time alignment: resample or interpolate the data according to preset time intervals (such as every minute or every hour) to ensure that all data points are in the same time interval.
[0081] 3. Resampling: Use data frame tools (such as Pandas) to perform time resampling to aggregate the data into specified time intervals.
[0082] 4. Interpolation: interpolate missing data points. Commonly used methods include linear interpolation, spline interpolation, etc. to fill in the missing values in the time interval.
[0083] Going a step further, data smoothing is the process of eliminating short-term fluctuations and noise to make the data smoother and more continuous, thereby better reflecting long-term trends and patterns. The process is as follows:
[0084] 1. Moving average: calculate the moving average of each data point. The commonly used window size can be set according to the characteristics of the data and analysis requirements (such as 3 points, 5 points, 7 points, etc.).
[0085] 2. Exponential smoothing uses the exponentially weighted average method to smooth the data, with more recent data points having higher weights.
[0086] 3. Filter, use signal processing methods such as low-pass filter to remove high-frequency noise and retain low-frequency signals.
[0087] In summary, this embodiment can significantly improve the quality and reliability of data by removing noise data, time synchronization and data smoothing, ensuring the accuracy and effectiveness of subsequent feature extraction and model training. These steps provide clean, continuous and aligned input data for the anomaly recognition model, thereby improving the anomaly detection performance of the model.
[0088] Preferably, extracting drinking water consumption features related to abnormal situations from the preprocessed data includes:
[0089] The pre-processed data are grouped by day, and the average daily water intake of each group of pre-processed data is calculated;
[0090] Performing linear regression on the preprocessed data to fit a drinking water trend line;
[0091] Extracting the regression coefficient of the drinking water trend line as the change trend;
[0092] Calculate the maximum and minimum values of the preprocessed data within a preset time period, and determine the maximum and minimum values as the drinking water peak and valley values;
[0093] The standard deviation of the numerical values of the preprocessed data is calculated to obtain the fluctuation of the water intake.
[0094] Preferably, the initial neural network is trained using preset historical data to obtain a trained anomaly recognition model, including:
[0095] Obtain the preset water intake change trend data set, daily average water intake data set and water intake peak and valley value data set;
[0096] Constructing an initial neural network; the initial neural network includes an input layer, an LSTM layer, a fully connected layer, and an output layer connected in sequence;
[0097] Obtain the drinking water consumption change trend data set, the daily average drinking water consumption data set, the drinking water consumption peak and valley value data set, and the drinking water consumption fluctuation data set in the historical database;
[0098] The water intake change trend data set, the daily average water intake data set, the water intake peak and valley value data set, and the water intake volatility data set are input into the initial neural network for training to obtain a trained abnormality recognition model.
[0099] Specifically, the model structure design of this embodiment is as follows:
[0100] 1. Input layer: input features include: time series water intake data.
[0101] 2. LSTM layer, two LSTM layers are used to capture the long-term dependencies and features of time series.
[0102] 3. Fully connected layer, the LSTM output passes through one or more fully connected layers (also called Dense layers) to continue high-level feature extraction and fusion.
[0103] 4. Output layer: The output layer can be a unit for binary classification (normal vs. abnormal), or multiple units for multi-classification problems. In this embodiment, multiple units of multi-classification problems are used to display abnormality recognition results.
[0104] Preferably, the abnormality identification result includes: abnormality type, abnormality time, abnormality partition and abnormality degree.
[0105] Preferably, the abnormal types include: excessive water intake, low water intake, drastic fluctuations in water intake, and abnormal drinking behavior caused by environmental or health problems.
[0106] Specifically, the abnormal types include excessive water intake, low water intake, drastic fluctuations in water intake, drinking equipment failure, and abnormal drinking behaviors caused by environmental or health problems. The abnormal time is the timestamp of the abnormality. For example: start time, duration. The abnormal partition refers to the farming partition where the specific abnormality occurs. For example: partition number or name. The degree of abnormality refers to the severity of the abnormality or the degree of deviation from the normal range. For example, it can be expressed by standard deviation, ratio or percentage: slight abnormality, moderate abnormality and severe abnormality.
[0107] Furthermore, this embodiment can also display the cause of the abnormality, that is, analyze the possible causes of the abnormality by combining other sensor data (feed, temperature and humidity sensors) or model analysis results, such as: disease outbreak, feed quality change, water leakage, abnormal temperature or humidity, human error.
[0108] In addition, this embodiment can make correction suggestions based on the identification results. For example: check and repair drinking water equipment, check and adjust stocking density or feeding strategy, monitor and adjust environmental conditions, health check and treatment
[0109] Specifically, this embodiment is designed with a visualization function, that is, a chart of abnormal data is displayed to help users understand and locate abnormalities more intuitively. Specific visualization charts include water intake time series charts, abnormality marks, and partition comparison charts.
[0110] Corresponding to the above method, this embodiment further provides a broiler farming abnormality identification system based on drinking water consumption data, including:
[0111] A data monitoring module is used to install a drinking water monitoring device on the drinking water pipe of the target breeding farm, and use the drinking water monitoring device to collect the drinking water volume data of each breeding zone in real time and record the corresponding timestamp;
[0112] A preprocessing module, used for preprocessing the collected water intake data and the corresponding timestamp to obtain preprocessed data;
[0113] A feature extraction module, for extracting drinking water consumption features related to abnormal situations from the preprocessed data;
[0114] A training module is used to train the initial neural network using preset historical data to obtain a trained anomaly recognition model;
[0115] The recognition module is used to input the drinking water consumption characteristics into the abnormality recognition model to obtain an abnormality recognition result.
[0116] The beneficial effects of the present invention are as follows:
[0117] (1) The present invention utilizes drinking water monitoring equipment to collect drinking water data of each breeding zone in real time, which can detect abnormal situations in time and reduce the potential impact of abnormalities on broiler breeding.
[0118] (2) The data collected by the monitoring equipment of the present invention is highly accurate and can accurately record the drinking water situation of each breeding area, helping to accurately identify abnormalities.
[0119] (3) The present invention utilizes deep learning models such as neural networks for training and anomaly identification, which can better mine the potential features and patterns of data and realize intelligent anomaly detection.
[0120] (4) By extracting drinking water consumption characteristics related to abnormal situations, the abnormality recognition model of the present invention can quickly identify potential abnormalities and issue early warnings, helping farm managers to take timely measures.
[0121] (5) The present invention can detect abnormal conditions in a timely manner and deal with them, thereby avoiding losses and improving production efficiency, and helping to optimize the breeding environment and the health of the chickens.
[0122] (6) The present invention is based on historical data training and model application, which can help farm managers make data-driven decisions and improve the scientificity and accuracy of production management.
[0123] (7) The present invention can realize automatic monitoring, processing and reminder of data through an automated abnormality identification system, reduce manual intervention and improve efficiency. And by accurately identifying abnormal situations and quickly adjusting the breeding environment, the quality of breeding products can be improved and sales competitiveness can be increased.
[0124] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.
[0125] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of the present invention. At the same time, for those skilled in the art, according to the idea of the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.
Claims
1. A method for identifying abnormalities in broiler farming based on drinking water consumption data, characterized in that: include: Install drinking water monitoring equipment on the drinking water pipes of the target breeding farm, and use the drinking water monitoring equipment to collect drinking water volume data of each breeding zone in real time and record the corresponding timestamp; Preprocessing the collected water intake data and the corresponding timestamp to obtain preprocessed data; extracting drinking water consumption features associated with abnormal conditions from the preprocessed data; Use the preset historical data to train the initial neural network to obtain a trained anomaly recognition model; The drinking water consumption feature is input into the abnormality recognition model to obtain an abnormality recognition result.
2. The broiler farming abnormality identification method based on drinking water consumption data according to claim 1 is characterized in that: The drinking water consumption characteristics include: changing trend of drinking water volume, average daily drinking water volume, drinking water peak and valley values, and drinking water volume volatility.
3. The broiler farming abnormality identification method based on drinking water consumption data according to claim 1 is characterized in that: The preprocessing process includes: removing noise data, time synchronization and data smoothing.
4. The broiler farming abnormality identification method based on drinking water consumption data according to claim 2 is characterized in that: Extracting drinking water consumption features related to abnormal situations from the preprocessed data includes: The pre-processed data are grouped by day, and the average daily water intake of each group of pre-processed data is calculated; Performing linear regression on the preprocessed data to fit a drinking water trend line; Extracting the regression coefficient of the drinking water trend line as the change trend; Calculate the maximum and minimum values of the preprocessed data within a preset time period, and determine the maximum and minimum values as the drinking water peak and valley values; The standard deviation of the numerical values of the preprocessed data is calculated to obtain the fluctuation of the water intake.
5. The broiler farming abnormality identification method based on drinking water consumption data according to claim 2 is characterized in that: Use the preset historical data to train the initial neural network to obtain a trained anomaly recognition model, including: Obtain the preset water intake change trend data set, daily average water intake data set and water intake peak and valley value data set; Constructing an initial neural network; the initial neural network includes an input layer, an LSTM layer, a fully connected layer, and an output layer connected in sequence; Obtain the drinking water consumption change trend data set, the daily average drinking water consumption data set, the drinking water consumption peak and valley value data set, and the drinking water consumption fluctuation data set in the historical database; The water intake change trend data set, the daily average water intake data set, the water intake peak and valley value data set, and the water intake volatility data set are input into the initial neural network for training to obtain a trained abnormality recognition model.
6. The broiler farming abnormality identification method based on drinking water consumption data according to claim 5 is characterized in that: The abnormality identification result includes: abnormality type, abnormality time, abnormality partition and abnormality degree.
7. The broiler farming abnormality identification method based on drinking water consumption data according to claim 6 is characterized in that: The abnormal types include: drinking too much water, drinking too little water, drastic fluctuations in water drinking, and abnormal drinking behavior caused by environmental or health problems.
8. A broiler farming abnormality identification system based on drinking water consumption data, characterized in that: include: A data monitoring module is used to install a drinking water monitoring device on the drinking water pipe of the target breeding farm, and use the drinking water monitoring device to collect the drinking water volume data of each breeding zone in real time and record the corresponding timestamp; A preprocessing module, used to preprocess the collected water intake data and the corresponding timestamp to obtain preprocessed data; A feature extraction module, used to extract drinking water consumption features related to abnormal situations from the preprocessed data; A training module is used to train the initial neural network using preset historical data to obtain a trained anomaly recognition model; The recognition module is used to input the drinking water consumption characteristics into the abnormality recognition model to obtain an abnormality recognition result.
Citation Information
Patent Citations
Water drinking health monitoring method for pets
CN116391644A
Multi-channel broiler automatic feeding, food and water consumption amount and weight detection device
CN109169405A
Livestock and poultry water intake real-time monitoring method
CN110954186A
Water treatment time sequence data anomaly detection method based on LSTM
CN112765896A
Abnormal monitoring and early warning method for industrial water consumption of enterprises in specific area
CN114611764A
Cited By
Quantitative monitoring method and system for water intake of live pigs
CN120092721A
