Cow oestrus prediction method, server, medium and program product
Through deep learning algorithms and sensor and camera technology, a cow estrus prediction model was established, which solved the problems of inaccurate and inefficient prediction of cow estrus in the existing technology, and achieved accurate prediction of cow estrus and ranch management optimization.
Patent Information
- Application Number
- CN202411114065.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-14
- Publication Date
- 2024-11-22
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing dairy cow estrus prediction technology relies on manual observation and simple physiological indicators, which are difficult to guarantee accuracy and inefficient, difficult to achieve real-time, comprehensive and continuous monitoring, and cannot meet the needs of large-scale breeding.
By obtaining the historical estrus records of dairy cows, establishing estrus sample data, combining deep learning algorithms to build estrus prediction models, using sensors and cameras to obtain monitoring data, determining estrus-related feature data, inputting the model to predict estrus index, and providing information through the display end .
It has achieved accurate prediction of the estrus of dairy cows, improved reproduction efficiency and breeding income, enhanced the accuracy and reliability of the prediction model, timely discovered the pre-estrus status, and provided scientific ranch management measures.
Smart Images

Figure CN119005802A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the biological field, and in particular to a method, server, medium and program product for predicting estrus of dairy cows. Background Art
[0002] In modern dairy farming, accurate prediction of cow estrus is crucial to improving reproductive efficiency and farming benefits. Efficient estrus prediction can help farms arrange breeding time reasonably, reduce the empty period, and thus increase milk production and the reproductive quality of the herd.
[0003] Existing dairy cow estrus prediction technologies usually rely on manual observation and simple physiological index measurements. For example, farmers observe the behavioral changes of dairy cows with the naked eye, such as restlessness and mounting, or use a thermometer to regularly measure the body temperature of dairy cows to determine whether they are in estrus. Some farms also determine estrus by testing the hormone levels in dairy cows. Manual observation is also easily affected by the external environment in which dairy cows are located, such as seasonal changes and weather conditions. Some farms also assist in determining estrus by monitoring the calls of dairy cows or observing their eating conditions.
[0004] However, this method has obvious defects. Manual observation is easily affected by the experience and subjective judgment of the breeders, and its accuracy is difficult to guarantee. Moreover, it is difficult to fully reflect the estrus status of dairy cows by measuring only a small number of physiological indicators such as body temperature, which is easy to cause misjudgment. This method requires a lot of manpower and time, is inefficient, and is easy to miss estrus individuals in large-scale breeding scenarios. Therefore, since most existing methods are intermittent and one-sided, it is difficult to achieve real-time, comprehensive and continuous monitoring of the estrus status of dairy cows, and thus it is difficult to meet the needs of large-scale breeding. Summary of the invention
[0005] The present application provides a method, server, medium and program product for predicting estrus of dairy cows, which can accurately predict the estrus of dairy cows in the existing technology, which relies on manual experience and may cause inaccurate predictions, so that managers can make breeding preparations in advance, thereby improving breeding efficiency and breeding income to a certain extent.
[0006] In a first aspect, the present application provides a method for predicting estrus of dairy cows, which is applied to a server. The method comprises: after obtaining historical estrus records of multiple dairy cows in a target dairy cow group, determining estrus sample data of the target dairy cow group, the estrus sample data at least includes monitoring data of the target dairy cow group with the same physical condition and reproductive status at different parities and ages during estrus, the monitoring data at least includes one of the historical reproductive records of the cattle, the length of the estrus cycle, physiological indicators, movement parameters and meteorological data during the estrus period; establishing an estrus prediction model between the monitoring data and estrus information of the target dairy cow based on the estrus sample data combined with a deep learning algorithm; after obtaining the monitoring data of the target dairy cow through a variety of sensors and camera devices, determining feature data related to estrus of the dairy cow; inputting the feature data into the estrus prediction model to determine estrus prediction information of the target dairy cow; determining an estrus index based on the estrus prediction information, the estrus index refers to the probability that the dairy cow is in estrus according to the estrus prediction model; and sending the estrus prediction information and the estrus index to a display terminal for display.
[0007] By adopting the above technical solution, the estrus prediction model is established by combining estrus sample data with deep learning algorithms, which can accurately predict the estrus of dairy cows. Monitoring data is obtained through sensors and cameras, characteristic data related to estrus is determined, and input into the estrus prediction model to obtain estrus prediction information and estrus index. This information is sent to the display end, so that farmers can understand the estrus of dairy cows in time, make breeding preparations in advance, improve breeding efficiency, and increase breeding income.
[0008] In combination with some embodiments of the first aspect, in some embodiments, after the step of inputting the characteristic data into the estrus prediction model to determine the estrus prediction information of the target cow, it also includes: establishing an individual health record of the target cow based on the monitoring data; evaluating the health status of the target cow based on the individual health record; determining estrus influencing factors based on the health status and the estrus prediction information; and sending the estrus influencing factors to a display terminal for display.
[0009] By adopting the above technical solution, individual health records can be established based on monitoring data, which can fully understand the health status of the target dairy cows. The factors affecting estrus can be determined by combining the health status and estrus prediction information, and sent to the display end, which helps farmers to deeply understand the reasons for estrus of dairy cows and take targeted measures to improve the reproductive performance and health level of dairy cows.
[0010] In combination with some embodiments of the first aspect, in some embodiments, after the step of establishing an estrus prediction model between the monitoring data and estrus information of the target cow based on the estrus sample data combined with the deep learning algorithm, it also includes: determining the interaction frequency and interaction method between the target cows according to the camera device; judging whether the target cow is in a set behavior state before estrus based on the interaction frequency and the interaction method combined with the preset estrus interaction frequency and estrus interaction method; if it is in the set behavior state, increasing the collection frequency of monitoring data and sending an early warning message to the management end, and the early warning message is used to prompt the management personnel to make subsequent preparations; if it is not in the set behavior state, continuing to monitor according to the regular monitoring frequency.
[0011] By adopting the above technical solution, the interaction frequency and interaction mode between target cows can be determined according to the camera equipment, so as to timely find out whether the cows are in the set behavior state of pre-estrus. If they are in this state, the frequency of monitoring data collection is increased and early warning information is sent to the management end, so that the management can prepare in advance and take timely measures to improve the success rate of reproduction.
[0012] In combination with some embodiments of the first aspect, in some embodiments, if in the set behavior state, after the step of increasing the frequency of monitoring data collection and sending early warning information to the management end, it also includes: after receiving question information related to cow estrus sent by the management end, sending the question information to the expert end; after receiving the suggested solution sent by the expert end, sending the suggested solution to the management end for display.
[0013] By adopting the above technical solutions, professional guidance and suggestions can be provided to managers to help them better deal with the pre-estrus situation of dairy cows and improve the level of breeding management.
[0014] In combination with some embodiments of the first aspect, in some embodiments, after the step of sending the estrus prediction information and the estrus index to the display terminal for display, it also includes: obtaining the actual estrus status of the target cow through a preset channel; if the estrus prediction information does not match the actual estrus status, determining the difference factors that cause the mismatch; and optimizing the estrus prediction model according to the difference factors.
[0015] By adopting the above technical solution, the actual estrus status of the target dairy cows can be obtained through the preset channels and compared with the estrus prediction information, so as to timely find out the deficiencies of the prediction model. The estrus prediction model can be optimized according to the difference factors, and the accuracy and reliability of the prediction model can be continuously improved to provide more accurate services for dairy farming.
[0016] In combination with some embodiments of the first aspect, in some embodiments, after the step of sending the estrus prediction information and the estrus index to the display end for display, it also includes: obtaining the pasture environment of the target cow sent by the management end; evaluating the potential impact of the pasture environment on the estrus of the cow in combination with the pasture environmental factors and the estrus prediction information, and determining the evaluation results; after determining the corresponding pasture management measures from a preset solution library based on the evaluation results, the pasture management measures are sent to the display end for display.
[0017] By adopting the above technical solution, the pasture environment sent by the management end is obtained, and the potential impact of the pasture environment on the estrus of dairy cows is evaluated by combining pasture environmental factors and estrus prediction information, which can provide scientific pasture management measures for farmers. Sending pasture management measures to the display end for display helps to improve the pasture environment and improve the estrus rate and reproductive performance of dairy cows.
[0018] In combination with some embodiments of the first aspect, in some embodiments, after obtaining the historical estrus records of multiple cows in the target cow group, after the step of determining the estrus sample data of the target cow group, it also includes: after extracting representative audio features through the call samples, constructing a cow call feature data set, the call sample is one of the physiological indicators in the monitoring data, which means that the call samples of the target cow at different estrus stages are collected in advance by audio collection equipment; training an estrus prediction model for the calls of cows at different estrus stages based on a deep learning algorithm; after obtaining the target audio monitoring data of the target cow, extracting the target audio features; inputting the target audio features into the trained estrus prediction model to determine the estrus stage of the target cow.
[0019] By adopting the above technical solution, representative audio features are extracted from the call samples, a cow call feature data set is constructed, and an estrus prediction model is trained, so that estrus information can be obtained from the calls of cows. After obtaining the target audio monitoring data of the target cow, the target audio features are extracted and input into the estrus prediction model, which can accurately determine the estrus stage of the cow, provide farmers with more estrus monitoring methods, and improve the accuracy of estrus monitoring.
[0020] In a second aspect, the present application provides a server, comprising: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to cause the server to execute the method described in the first aspect and any possible implementation of the first aspect.
[0021] In a third aspect, the present application provides a computer-readable storage medium comprising instructions, which, when executed on a server, causes the server to execute the method described in the first aspect and any possible implementation of the first aspect.
[0022] In a fourth aspect, the present application provides a computer program product, which, when executed on a server, enables the server to execute the method described in the first aspect and any possible implementation manner of the first aspect.
[0023] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. The estrus prediction model is established by combining estrus sample data with deep learning algorithms, and the monitoring data is obtained through sensors and cameras and input into the model. Therefore, the technical problem of inaccurate estrus prediction of dairy cows in the existing technology is effectively solved, thereby achieving the technical effect of accurately predicting the estrus of dairy cows, making breeding preparations in advance, and improving breeding efficiency and breeding income.
[0024] 2. Due to the technical means of obtaining the actual estrus situation and comparing it with the predicted information, and optimizing the estrus prediction model according to the difference factors, the technical problem of insufficient accuracy of the estrus prediction model in the existing technology is effectively solved, thereby achieving the technical effect of continuously improving the accuracy and reliability of the prediction model and providing more accurate services for dairy cattle farming.
[0025] 3. Due to the adoption of technical means to obtain pasture environment and combine estrus prediction information to evaluate potential impacts, as well as to determine and send pasture management measures, the technical problem of the difficulty in evaluating and managing the impact of pasture environment on dairy cow estrus in the existing technology has been effectively solved, thereby achieving the technical effect of improving pasture environment and improving dairy cow estrus rate and reproductive performance. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] Figure 1 This is a flow chart of a method for predicting estrus of dairy cows in an embodiment of the present application; Figure 2 This is another flow chart of the method for predicting estrus of dairy cows in an embodiment of the present application; Figure 3 It is a schematic diagram of a physical device structure of a server in an embodiment of the present application. DETAILED DESCRIPTION
[0027] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to be used as limitations to the present application. As used in the specification and appended claims of the present application, the singular expressions "one", "a kind of", "said", "above", "the" and "this" are intended to also include plural expressions, unless there is a clear indication to the contrary in the context. It should also be understood that the term "and / or" used in the present application refers to and includes any or all possible combinations of one or more listed items.
[0028] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as suggesting or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, unless otherwise specified, "plurality" means two or more.
[0029] For ease of understanding, an application scenario of the relevant method is introduced below.
[0030] Mr. Cao runs a small-scale dairy farm in a certain place, raising 80 cows. Mr. Cao mainly relies on himself and several workers to judge whether the cows are in estrus through manual observation. The specific method is to manually measure the body temperature of the cows at multiple times in the morning and evening every day, and observe whether the cows' behavior has the characteristics of the estrus period. One evening, Mr. Cao found that the body temperature of cow No. 27 was a little higher than usual, and it was a little restless, and there were some other estrus-related characteristics. He judged that No. 27 might be in estrus. Mr. Cao performed artificial insemination on No. 27, but after a month of pregnancy examination, No. 27 still did not conceive. Later, Mr. Cao consulted a veterinarian, and the veterinarian examined the ovarian status of No. 27 by B-ultrasound. The results showed that No. 27 was not in estrus at the time, and Mr. Cao's judgment was wrong. This not only wasted an opportunity for artificial insemination, but also prolonged No. 27's empty period.
[0031] It can be seen that with the expansion of the scale of breeding, manual observation alone has become increasingly difficult, and the estrus period of individual dairy cows is often missed, and it is impossible to continuously monitor the estrus status of each dairy cow. This directly affects the conception rate of dairy cows and seriously restricts the breeding efficiency of the farm.
[0032] By adopting the method for predicting estrus of dairy cows in the embodiment of the present application, the estrus of dairy cows can be accurately predicted, and the breeders can timely understand the estrus of dairy cows, make breeding preparations in advance, and increase breeding income. Assume that Mr. Cao adopts the method for predicting estrus of dairy cows in the present application, and establishes an estrus prediction model by analyzing the historical estrus records of the dairy cow group. The monitoring system continuously detects the body temperature, behavior and other data of dairy cows, and inputs the model to calculate the estrus index. The system will accurately prompt Mr. Cao about the estrus of dairy cows and predict the best breeding time, so that he can breed the cow at the right time instead of missing the estrus period.
[0033] For ease of understanding, the following describes the process of the method provided by this implementation in combination with the above scenario. Figure 1 , is a flow chart of a method for predicting estrus of dairy cows in an embodiment of the present application.
[0034] S101. After obtaining historical estrus records of multiple dairy cows in a target dairy cow group, determine estrus sample data of the target dairy cow group, the estrus sample data at least including monitoring data of the target dairy cow group with the same physical condition and reproductive status at different parities and ages during estrus, the monitoring data at least including one of the historical reproductive records of the cattle, the length of the estrus cycle, physiological indicators, movement parameters and meteorological data during the estrus period; Each cow can be equipped with a pedometer or other monitoring devices that can monitor the cow's physiological characteristics on its neck or foot. The pedometer can detect and record the cow's steps, exercise distance, exercise intensity and other parameters in real time, and monitor the cow's daily activity. The monitoring devices can include body temperature sensors, heart rate sensors, audio acquisition equipment, etc. These devices form an Internet of Things monitoring system, which can fully understand the situation of each cow.
[0035] For this step, the server needs to connect to the farm's database to query the basic information of each cow in the target dairy cow group, including breed, age, weight, milk production, pregnancy times and other data. The server needs to retrieve the estrus and mating records of each cow in the past certain period of time, and obtain relevant data such as estrus date, estrus cycle length, and the number of matings required for each time. Then, the server will select cows with similar physical condition and reproductive status from the target group to form the target group of sample data, such as cows with similar age, weight, milk production, and pregnancy times. Perform statistical analysis on the historical estrus cycle of each cow in the sample group, and calculate the average length and variation range of the estrus cycle. At the same time, query the monitoring data of the sample group at different estrus periods, including body temperature, activity, exercise parameters, appetite changes, and meteorological conditions (temperature, humidity, wind speed, etc.) at that time. And classify and organize these monitoring data to form an estrus sample data set to provide data support for subsequent modeling. For example, the server selects three cows numbered 1024, 2046, and 3071 from the target group as estrus samples. The parity of these three cows is 1, their age is between 26 and 30 months, their weight is between 550 and 600 kg, and their milk production is between 22 and 25 kg / day. The server calculates the average length of the estrus cycle of these three cows in the past year to be 21 days, 23 days, and 20 days, respectively. At the same time, the server retrieves the time series of the body temperature, activity level, exercise steps, and local meteorological data of these three cows before and after estrus in the past year to construct an estrus sample data set.
[0036] In addition, the server can also obtain real-time data of the target cow from the farm's monitoring system. The monitoring system monitors and records the cow's physiological parameters and activity data in real time through devices such as pedometers and temperature sensors installed on the cow's neck or legs. These monitoring devices form an Internet of Things system that can fully understand the situation of each cow. The server can regularly collect data from the monitoring system and integrate it into the estrus sample data set to enrich the dimensions of the sample data.
[0037] In some embodiments, the server can more specifically predict the estrus stage by analyzing the calls of cows. First, the server obtains the audio samples of the target cows at different estrus stages from the audio acquisition device. Then, using audio analysis technology, representative audio features such as pitch, rhythm, spectral features, etc. are extracted from these call samples to construct a feature data set of cow calls. Based on this data set, the server trains a cow call estrus prediction model based on deep learning. The deep learning model can automatically learn the implicit features of calls at different estrus stages. The server inputs audio samples marked with estrus stages to train the model. The training goal is to enable the model to analyze the input cow call features and correctly predict its estrus stage. After the model is successfully trained, the server can use it to predict the estrus stage of the target cow. When new call monitoring data of the target cow is obtained, the server first extracts the features of its audio, and then inputs it into the trained model. The model analyzes its features and outputs the estrus stage prediction result.
[0038] In this way, by training the model with call samples, the server can determine the stage of the estrus cycle of a cow by analyzing the characteristics of its calls, just like a human expert, and thus determine the time of estrus, providing another means of prediction.
[0039] S102, establishing an estrus prediction model between the monitoring data and estrus information of the target dairy cow based on the estrus sample data combined with a deep learning algorithm; After obtaining the estrus sample data of the target dairy cow group, the server needs to establish an estrus prediction model based on these sample data. Specifically, the server will preprocess the estrus sample data, including removing abnormal data, smoothing, standardization, etc., so that the sample data becomes standardized, and the preprocessed data is more conducive to model training. The server will extract features from the sample data, which may include physiological characteristics such as body temperature, activity, number of steps, appetite changes, and information such as estrus cycle length and meteorological data of the target dairy cow. The server can use a specific method to select features that are more closely related to estrus, and perform feature fusion to obtain key feature vectors that can reflect the estrus state of dairy cows. The server then constructs the network structure of the deep learning model, which can be simply compared to a "multi-layer filter". The estrus sample data collected by the server will first pass through the input layer, just like passing through the first filter, filtering out some unimportant features. Then it will pass through the first hidden layer, which is equivalent to the second filter, to further extract more important features. As the number of layers increases, each hidden layer is filtering and strengthening key features, and gradually extracting advanced features from the data. The number of nodes in the hidden layer determines the density and screening strength of each filter. The more nodes there are, the denser the filter is, the stronger the screening effect is, and more complex features can be extracted. However, too many nodes may also lead to over-filtering, so the number of nodes needs to be determined through verification to find the optimal number. To give a simple analogy, suppose that estrus data is like a pile of small balls of different colors and shapes. The input layer may first filter out some small balls with very light colors (unobvious features). The first hidden layer filters out some small balls with irregular shapes (noise data). The second hidden layer only leaves the darkest and most rounded small balls (key features). After multiple layers of filtering, the most important features are extracted from the initially messy data. Structures such as recurrent neural networks and convolutional neural networks also use similar multi-layer feature extraction ideas. It’s just that the calculation method and connection relationship of each layer are different, which is suitable for extracting features of different types of data. The server can select a suitable network structure based on the structure of the estrus sample data.
[0040] The server inputs the estrus sample data into the set deep learning network structure as training data for cyclic training. The optimization algorithm is used to iteratively update the network parameters to gradually minimize the loss function. After each parameter update, the value of the current loss function is calculated. Repeat multiple rounds of training until the network's loss function no longer decreases significantly, and the training can be stopped. After the training is completed, the verification data is used to test the effect of the network. If the predetermined effect is achieved, such as the accuracy rate and other indicators meet the requirements, the training is successful and the estrus prediction model is obtained.
[0041] S103, after acquiring monitoring data of the target dairy cow through various sensors and camera devices, determining characteristic data related to estrus of the dairy cow; After the server has established the estrus prediction model, it is necessary to obtain various monitoring data of the target cow in real time and extract features related to estrus from it. The server will obtain various monitoring data from the monitoring system pre-set by the farm. The monitoring system includes a variety of sensor devices, such as leg pedometers, collar body temperature sensors, heart rate monitors, etc. installed on the cows. These sensor devices can upload the physiological parameters and behavioral data of the cows in real time through wireless networks. The monitoring system also includes camera equipment for capturing the image information of the cows, and can determine the cow's activity level, number of ruminations, standing or lying time, etc. through image analysis. The server needs to subscribe to various monitoring data of the target cow from the monitoring system to obtain the information source.
[0042] After acquiring the monitoring data, feature extraction and screening are performed. Based on expert experience, the server will determine in advance which data fields are related to the estrus cycle of dairy cows, such as body temperature, activity level, number of ruminations, etc. The server will extract these estrus-related feature data. Some related indicators will also be calculated, such as body temperature fluctuation values, activity level surge indicators, etc. Before entering the estrus prediction model, feature selection is also required to delete non-critical features to obtain a feature subset that has a greater effect on estrus prediction.
[0043] S104, inputting the characteristic data into the estrus prediction model to determine the estrus prediction information of the target dairy cow; After the server extracts the characteristic data related to estrus, it needs to input it into a pre-established estrus prediction model, run the model to perform estrus prediction, and obtain the result information of estrus prediction.
[0044] Specifically, the server will organize the feature data to generate the input vector of the model. The dimension of the input vector is consistent with the number of features during model training. For example, if the input layer of the model requires 4 features, the server will select 4 features from the feature data and arrange them in an order consistent with the model to form a 4-dimensional vector. The generated input vector is then sent to the input layer of the estrus prediction model, and the model will generate output results based on the trained parameters through hidden layer calculations.
[0045] The output results are usually post-processed first. The server will parse the final output value of the model and convert it into the result information of estrus prediction, such as "estrus prediction index is 92". The estrus prediction index indicates the probability that the model predicts that the target cow is in the peak estrus period based on the current feature data of the target cow. The higher the prediction index, the more reliable the model is in judging that the cow is in estrus. In addition to generating the prediction index, the server can also configure the model to output information such as the possible start and end dates of estrus and the time when the peak estrus period occurs. This information constitutes the estrus prediction result of the target cow. If the input is image features, intuitive results such as the estrus calendar with the estrus period marked can also be displayed.
[0046] For example, the server extracts the three features of the cow's body temperature fluctuation range, activity level change, and rumination frequency in the past three days. A three-dimensional vector [1.2, 105, 312] is formed as the model input, and the model outputs "estrus prediction index 96" after calculation. It also gives the prediction of the peak estrus period as "August 2 to August 5". The server summarizes these results as the estrus prediction details of the target cow.
[0047] S105, determining an estrus index according to the estrus prediction information, where the estrus index refers to the probability that the dairy cow is in estrus according to the estrus prediction model; After obtaining the estrus prediction information of the target cow, the server needs to extract the estrus prediction index separately. The estrus prediction index is the probability of the cow being in estrus directly predicted by the model, which reflects the confidence of the estrus prediction. Specifically, the server will parse the output of the model and extract the value of the prediction index. The prediction index can be a value in a set range, such as a value between 60-100, which indicates the percentage of the correct judgment. For example, the estrus prediction information output by the model contains: "Estrus prediction index: 76". The server will extract the value of 76 as the estrus index. The higher the estrus index, the greater the possibility that the cow is in estrus. For example, an estrus index of 76 means that the model judges that there is a 76% probability that the cow is in estrus.
[0048] In addition to directly extracting the prediction index output by the model, the server can also calculate the estrus index based on the estrus prediction information. For example, the model outputs the predicted estrus start date and end date, and the server can calculate the ratio of the number of days of estrus to the total number of days in an estrus cycle of the cow as the estrus index. It is important to extract an accurate estrus index because it intuitively reflects the results and credibility of estrus prediction. The higher the estrus index, the more reliable the prediction result, and the server needs to use this to determine subsequent treatment measures, such as promptly reminding farmers to make preparations before estrus.
[0049] S106: Send the estrus prediction information and the estrus index to a display terminal for display.
[0050] After obtaining the estrus prediction information and estrus index of the target cow, the server needs to send the results to the display end to display to the user, so that the user can understand the estrus status of the cow and make corresponding operations. Specifically, the server will transmit the estrus prediction information and estrus index to the display device of the monitoring center of the farm through the network, which can be a display terminal that supports network access, such as a computer or Pad. The display end needs to install a software program that matches the server to receive and display the prediction results. To make the results more intuitive and easy to understand, the server can summarize the results before sending them. For example, describe the estrus prediction situation in concise language, and list the estrus index, possible estrus date range and other information in a table. You can also draw a bar chart of estrus probability and a calendar with estrus status marked, so that users can quickly grasp key information.
[0051] After the display end receives the result data from the server, the software program will automatically complete the information analysis and interface rendering, and display the estrus prediction results with unified format and complete content on the screen of the display device in the form of text, charts, etc.
[0052] In some embodiments, the server obtains the actual estrus of the target cow through a preset channel. The preset channel can be a production management system connected to the farm, or the farmer can feedback the actual estrus date and other information in the software. The server will take out the estrus prediction information and compare it with the actual estrus date, estrus duration and other actual conditions reported by the user. If there is a difference between the two, such as the predicted estrus peak is inconsistent with the actual, it indicates that the prediction may be wrong. The server will analyze the reasons for the difference and determine whether it is caused by model errors or other factors such as inaccurate or incomplete data. For example, the number of sample data used for prediction is insufficient; or there is a deviation in the training algorithm; or there are omissions in the actual input monitoring data. The server will adopt different model optimization strategies based on the analysis results. For example, expand the sample data optimization model, improve the algorithm logic, or increase the types of monitoring data. After repeated verification and optimization iterations, the server can gradually improve the accuracy of estrus prediction, so that the difference between the predicted information and the actual situation is getting smaller and smaller, and the model prediction effect is better.
[0053] By adopting the dairy cow estrus prediction method in the embodiment of the present application, an estrus prediction model is established by combining estrus sample data with a deep learning algorithm, and monitoring data is obtained through sensors and cameras and input into the model, thereby accurately predicting the estrus of dairy cows, making breeding preparations in advance, and improving breeding efficiency and breeding income. This not only solves the problem of inaccurate estrus prediction of dairy cows in the prior art, but also improves the economic benefits of the breeding industry.
[0054] In some embodiments, the server can collect and integrate various physiological parameter monitoring data of the target cow, such as body temperature, heart rate, respiratory rate, etc., as well as information such as medical history and milk production, to establish the individual reproduction, health and milk production file of the target cow. The file will be updated in real time and can fully reflect the individual reproduction, health and milk production status of the target cow. For example, the file will display the temperature fluctuation curve of the cow in the last two months, the average milk production data per week, etc. The server can also use the evaluation algorithm to measure and judge the status of the target cow based on the file data. For example, evaluate the number of fevers and whether the decrease in milk production is normal. After obtaining the estrus prediction information, the server will combine the evaluation results of the file, analyze the relationship between the estrus state and the individual reproduction, health and milk production status, and determine the health factors that may affect estrus. For example, it is judged that the estrus period is disturbed due to fever, or the estrus signal is weak due to the decrease in milk production. Finally, the server will send the judged estrus health influencing factors to the monitoring terminal of the farm manager through the network for display, so that users can pay attention to these factors and adjust the feeding plan in a targeted manner.
[0055] By establishing profiles of target cows, the server can more comprehensively evaluate the individual cows' reproduction, health, and milk production status, and analyze their impact on estrus, thereby taking effective measures to help users improve the reproductive performance of cows.
[0056] In some embodiments, the server can obtain the environmental data of the pasture where the target cow is located submitted by the management end, including information such as temperature, humidity, light, air quality, etc., which can also be directly obtained through a camera or sensor set in a set place. The server will input these pasture environmental data and the estrus prediction information of the cow into the environmental impact assessment model. The model will analyze the positive or negative effects of different environmental factors on the estrus of the cow, and calculate the overall potential impact of the environment on the estrus. For example, suitable temperature and humidity are conducive to estrus, while insufficient light may inhibit estrus. The model can give each environmental factor a plus or minus score, and finally evaluate the environmental impact score. After obtaining the environmental impact assessment results, the server will search for solutions to the corresponding situation in the preset pasture management solution library. For example, if the assessment results show that light has a negative impact on estrus, the management measure of "increasing light duration" is matched. The server will send the pasture management measure solution found that is suitable for the environmental impact assessment result to the farm management end through the network for display. Users can adjust the pasture environment accordingly according to the management measure solution recommended by the server to create more optimized external conditions, which will help improve the estrus rate of dairy cows.
[0057] In some embodiments, when a cow enters the proestrus stage, it will exhibit specific interactive behaviors, such as frequently sniffing the genital area of other cows, attempting to mount other cows, etc. At this time, the server can increase monitoring of the target cows, and if the management personnel are unclear about the status of the estrus cows during the monitoring and handling process, they can obtain help through the pre-set expert consultation channel.
[0058] After combining the above scenarios, the following is a more detailed description of the process of the method provided by this implementation. Figure 2 , is another flow chart of the method for predicting estrus of dairy cows in an embodiment of the present application.
[0059] S201, determining the interaction frequency and interaction mode between the target cows according to the camera device; In the proestrus period, cows will have specific social interactions. In order to detect whether the target cow has entered the proestrus state, the server needs to analyze the interactions between cows.
[0060] Specifically, the server will connect to the camera equipment set up in the farm, which will record the social behavior between cows in real time. The server will use image recognition and behavior analysis algorithms to detect interactive behaviors such as physical contact and sniffing of reproductive organs between cows. Then, the server will count the total number of social interactions of the target cows within a certain period of time (for example, 1 hour) and calculate the frequency of interaction, such as 10 interactions per hour. In addition to the frequency of interaction, the server can also analyze the distribution of different interaction methods such as the number of sniffs, the number of attempts to mount, and the number of head rubbing. These data on interaction frequency and method can be obtained with the support of image recognition algorithms. If necessary, the server can also pass relevant videos and statistical data to experts to request further confirmation of the type of interaction. Finally, the server can build a database of interaction frequency and method characteristics for each target cow, providing a basis for judging the behavior status of the proestrus.
[0061] S202, judging whether the target cow is in a set behavior state of proestrus according to the interaction frequency and the interaction mode in combination with a preset estrus interaction frequency and estrus interaction mode; After obtaining the target cow's interaction frequency and method, the server needs to determine whether it has entered proestrus based on preset parameters. The server database will store the normal interaction frequency and method distribution data of proestrus cows, which can be obtained by analyzing a large number of samples. The server will compare the actual interaction frequency and method of the target cow with the baseline proestrus data in the database to determine whether the difference exceeds the preset threshold range. For example, if the target cow's sniffing interaction times reach 30 times within 1 hour, far exceeding the proestrus baseline of 10 times, then the server will determine that the cow is likely to have entered proestrus.
[0062] If the comparison results of interaction frequency and key interaction methods (such as sniffing genitals) exceed the threshold limit, the server will determine that the target cow is in the set behavioral state of proestrus.
[0063] If it is in the set behavior state, execute step S204; If it is not in the set behavior state, execute step S203.
[0064] S203. Continue to monitor according to the regular monitoring frequency; If it is determined that the target cow has not entered the proestrus state, the server will continue the regular monitoring frequency, regularly capture and analyze the interactive behavior of the target cow, and determine whether it has entered the proestrus state. For example, the server may analyze the interactive behavior of the target cow every 2 hours. If there is no obvious abnormality in the frequency and method of interaction, the server will wait for the next monitoring cycle and continue to monitor the group interactive behavior of the target cow. At the same time, the server is also monitoring other physiological parameters, such as body temperature, heart rate, activity level, etc. All monitoring results are comprehensively judged to determine the estrus state of the cow. Continuing the regular monitoring frequency can avoid unnecessary calculations and energy consumption caused by repeated analysis, and can also continuously observe changes in the state of the cow and discover problems at the first time.
[0065] S204, increasing the frequency of collecting monitoring data and sending warning information to the management end, the warning information is used to remind the management personnel to make subsequent preparations; If the target cow is judged to have entered the pre-estrus state, it is necessary to increase the monitoring of the cow in order to grasp the estrus dynamics in time. Generally speaking, the server will re-plan the working mode of the monitoring equipment and increase the frequency of monitoring data collection. For example, the interval of body temperature monitoring is increased from once an hour to once every 15 minutes; the activity volume statistics are increased from once a day to once every 2 hours. Increasing the monitoring frequency can obtain more and more real-time health indicator data for model prediction. At the same time, the server will send pre-estrus warning information to the terminal of the farm manager. The warning information will pop up in the form of a message reminder, informing the user that "cow number XX has entered the pre-estrus, please make subsequent preparations." The manager will prepare the necessary estrus breeding tools and supplies in advance based on this.
[0066] In some embodiments, the server can also prompt the possible peak period of estrus, allowing managers to arrange their work time. The server can also link some estrus feeding knowledge or processes in the early warning information for user reference.
[0067] S205, after receiving the question information related to cow estrus sent by the management end, send the question information to the expert end; If managers have questions when dealing with pre-estrus cows, they can get expert guidance through the online consultation function of the server. Specifically, the software on the management side will have a built-in online consultation window. When the user enters a question related to estrus and sends it, the server will receive the question information. For example, "What are the things to pay attention to when the cows in this field are in pre-estrus?" The server will first try to find the standard answer to the corresponding question in the knowledge base. If the knowledge base cannot reply, the server will forward the user's question to the connected expert side. The expert side here can be an expert team in the field of animal nutrition or reproduction, or a veterinary consulting platform. The expert side will receive the user's question information and prepare professional consulting answers.
[0068] S206: After receiving the suggestion from the expert, the suggestion is sent to the management end for display.
[0069] The expert side generates professional advice and answers in text, audio or video format based on the questions raised by the user. Then, the consultation results are sent back to the server through the application. The server will obtain various forms of answers fed back by the expert side. Then, these professional advice solutions are transmitted through the network and sent to the software interface of the user side for display. Users can read or listen to the text, audio and video information provided by experts to obtain professional knowledge and suggestions on the feeding and breeding of dairy cows before estrus. This can help users better monitor and handle the estrus of target dairy cows and carry out subsequent work scientifically. If the expert side feedbacks some generalized knowledge, the server can store the content in the knowledge base to form standard question-answer pairs for subsequent user inquiries. By connecting to expert resources, the server can provide timely online guidance services for farm managers to improve work efficiency and the level of management of dairy cows during estrus.
[0070] The following describes the server in the embodiment of the present invention from the perspective of hardware processing. Figure 3 , is a schematic diagram of a physical device structure of a server in an embodiment of the present application.
[0071] It should be noted that Figure 3 The structure of the server shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.
[0072] like Figure 3As shown, the server includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 302 or the program loaded from the storage part 308 to the random access memory (RAM) 303, such as executing the method described in the above embodiment. In RAM 303, various programs and data required for system operation are also stored. CPU 301, ROM 302 and RAM 303 are connected to each other through bus 304. Input / output (I / O) interface 305 is also connected to bus 304.
[0073] The following components are connected to the I / O interface 305: an input section 306 including an audio input device, a button switch, etc.; an output section 307 including a liquid crystal display (LCD) and an audio output device, an indicator light, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed so that a computer program read therefrom is installed into the storage section 308 as needed.
[0074] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 309, and / or installed from a removable medium 311. When the computer program is executed by the central processing unit (CPU) 301, various functions defined in the present invention are performed.
[0075] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, apparatus, or device.
[0076] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that marked in the accompanying drawings.
[0077] Specifically, the server of this embodiment includes a processor and a memory. The memory stores a computer program. When the computer program is executed by the processor, the cow estrus prediction method provided in the above embodiment is implemented.
[0078] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the server described in the above embodiment; or may exist independently without being assembled into the server. The above storage medium carries one or more computer programs, and when the above one or more computer programs are executed by a processor of the server, the server implements the cow estrus prediction method provided in the above embodiment.
[0079] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.
[0080] As used in the above embodiments, the term "when..." may be interpreted to mean "if..." or "after..." or "in response to determining..." or "in response to detecting...", depending on the context. Similarly, the phrases "upon determining..." or "if (the stated condition or event) is detected" may be interpreted to mean "if determining..." or "in response to determining..." or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)", depending on the context.
[0081] Those skilled in the art can understand that to implement all or part of the processes in the above-mentioned embodiments, the processes can be completed by computer programs to instruct related hardware, and the programs can be stored in computer-readable storage media. When the programs are executed, they can include the processes of the above-mentioned method embodiments. The aforementioned storage media include: ROM or random access memory RAM, magnetic disk or optical disk and other media that can store program codes.
Claims
1. A method for predicting estrus of dairy cows, applied to a server, characterized in that: The method comprises: After obtaining historical estrus records of multiple cows in a target dairy cow group, estrus sample data of the target dairy cow group is determined, wherein the estrus sample data at least includes monitoring data of the target dairy cow group with the same physical condition and reproductive status at different postpartum days or months of age during estrus, and the monitoring data at least includes one of the historical reproductive records of the cattle, the length of the estrus cycle, physiological indicators, movement parameters, and meteorological data during the estrus period; Establishing an estrus prediction model between the monitoring data and estrus information of the target dairy cow based on the estrus sample data combined with a deep learning algorithm; After acquiring the monitoring data of the target cow through various sensors and camera equipment, the characteristic data related to the cow's estrus is determined; Inputting the characteristic data into the estrus prediction model to determine the estrus prediction information of the target dairy cow; Determine an estrus index according to the estrus prediction information, wherein the estrus index refers to the probability of a dairy cow being in estrus according to the estrus prediction model; The estrus prediction information and the estrus index are sent to a display terminal for display.
2. The method according to claim 1, characterized in that After the step of inputting the characteristic data into the estrus prediction model to determine the estrus prediction information of the target dairy cow, the method further includes: Establishing individual reproduction, health and milk production records of the target dairy cows based on the monitoring data; Assessing the health status of the target dairy cow based on the individual reproduction, health and milk production records; Determining estrus influencing factors according to the health status and the estrus prediction information; The estrus influencing factors are sent to a display terminal for display.
3. The method according to claim 1, characterized in that: After the step of establishing an estrus prediction model between the monitoring data and estrus information of the target dairy cow according to the estrus sample data combined with the deep learning algorithm, the method further includes: Determine the interaction frequency and interaction mode between the target cows according to the camera equipment; Determining whether the target cow is in a set behavior state of proestrus according to the interaction frequency and the interaction mode combined with a preset estrus interaction frequency and estrus interaction mode; If it is in the set behavior state, the frequency of collecting monitoring data is increased and an early warning message is sent to the management end, and the early warning message is used to remind the management staff to make preparations for subsequent breeding; If it is not in the set behavior state, continue to monitor at the normal monitoring frequency.
4. The method according to claim 3, characterized in that If the state is in the set behavior state, after the step of increasing the frequency of collecting monitoring data and sending warning information to the management end, the method further includes: After receiving the question information about cow estrus sent by the management end, the question information is sent to the expert end; After receiving the suggestion plan sent by the expert end, the suggestion plan is sent to the management end for display.
5. The method according to claim 1, characterized in that After the step of sending the estrus prediction information and the estrus index to a display terminal for display, the method further includes: Obtaining the actual estrus status of the target dairy cow through a preset channel; If the estrus prediction information does not match the actual estrus situation, determining the difference factors that cause the mismatch; The estrus prediction model is optimized according to the difference factors.
6. The method according to claim 1, characterized in that After the step of sending the estrus prediction information and the estrus index to a display terminal for display, the method further includes: Obtaining the pasture environment of the target cow sent by the management end; Combining the pasture environmental factors and the estrus prediction information, evaluating the potential impact of the pasture environment on the estrus of dairy cows, and determining an evaluation result; After the corresponding pasture management measures are determined from a preset solution library according to the evaluation results, the pasture management measures are sent to a display terminal for display.
7. The method according to claim 1, characterized in that After obtaining the historical estrus records of multiple cows in the target cow group, after the step of determining the estrus sample data of the target cow group, the method further includes: After extracting representative audio features through the call samples, a cow call feature data set is constructed. The call sample is one of the physiological indicators in the monitoring data, which refers to the call samples of the target cows at different estrus stages collected in advance by audio collection equipment; Based on deep learning algorithms, we train an estrus prediction model based on the calls of cows at different estrus stages; After acquiring the target audio monitoring data of the target cow, extracting the target audio features; The target audio features are input into the trained estrus prediction model to determine the estrus stage of the target cow.
8. A server, characterized in that: The server includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the server to execute the method described in any one of claims 1-7.
9. A computer-readable storage medium comprising instructions, characterized in that: When the instructions are executed on a server, the server is caused to execute the method according to any one of claims 1 to 7.
10. A computer program product, characterized in that When the computer program product is run on a server, the server is caused to execute the method according to any one of claims 1 to 7.
Citation Information
Patent Citations
Milk cow estrus identification system and milk cow estrus identification method
CN106818532A
Time prediction method and device, storage medium and electronic equipment
CN113705528A
Dairy cow oestrus prediction device based on body temperature and exercise amount data
CN114586701A
Dairy cow behavior sign monitoring system and method, electronic equipment and storage medium
CN116616952A
Intelligent animal oestrus identification method and system
CN118020666A
Cited By
Intelligent management and control system and method for animal living place
CN121433004A
