Health detection method and equipment based on cattle stomach data and storage medium

Abstract

The invention relates to the field of animal health detection, and discloses a health detection method and equipment based on cattle stomach data and a storage medium. The method comprises the following steps: collecting characteristic time sequence data of a cattle stomach through an implantable stomach capsule; performing health state labeling processing on the feature time sequence data to obtain labeled feature data; performing parameter optimization processing on the preset linear regression model to obtain a first analysis model; performing parameter optimization processing on the preset random forest model to obtain a second analysis model; the method comprises the following steps: collecting monitoring time sequence data of a cattle stomach through an implantable stomach capsule; performing detection processing on the monitoring time sequence data to obtain a first analysis value, and performing detection processing on the monitoring time sequence data to obtain a second analysis value; and generating a health state corresponding to the cattle based on the first analysis value and the second analysis value. In the embodiment of the invention, under the condition that the deviation ratio of the analysis value is not high, the effect of detecting the accurate cattle health state based on the physiological data of the cattle stomach is realized.

Description

Technical Field

[0001] This invention relates to the field of animal health testing, and more particularly to a health testing method, device, and storage medium based on bovine stomach data. Background Technology

[0002] Monitoring the activity level of ruminants is a crucial aspect of livestock farming and health management. Currently, the following are some approaches to health management for ruminants: 1. Ankle bracelet pedometer device The device, equipped with a pedometer, is attached to the legs of cattle to monitor different steps and activity levels. However, the installation process is cumbersome, and the fit and tightness of the device are critical, requiring regular adjustments. Furthermore, most devices only monitor activity levels and cannot provide a comprehensive assessment of the cattle's health.

[0003] 2. Smart ear tag device Ear tags with motion sensors are implanted to monitor cattle's activity levels in real time via wireless transmission. However, implanted devices can cause infections or rejection reactions, and data can be distorted if the ear tag does not make proper contact (such as when the ear is bent or there is a foreign object).

[0004] 3. Smart collar device The device, equipped with a 3D accelerometer and timer, is attached to the legs of cattle to monitor different steps and activity levels. However, prolonged wear may cause discomfort to the cattle and affect their normal activities. In harsh environments, such as high temperature, high humidity, or dusty conditions, the device's performance and lifespan may be affected.

[0005] Therefore, given the instability and inaccuracy of existing technologies in monitoring cattle health, a new technology is needed to solve the current problems in monitoring cattle health. Summary of the Invention

[0006] The main objective of this invention is to solve the technical problem of unstable and inaccurate results in monitoring the health of cattle.

[0007] The first aspect of this invention provides a health detection method based on bovine stomach data, comprising the following steps: By implanting a gastric capsule, characteristic temporal data of the bovine stomach were collected; The health status annotation process is performed on the aforementioned feature time series data to obtain annotated feature data; The labeled feature data is divided according to preset division parameters to obtain first division data and second division data; Based on the first and second partition data, the parameters of the preset linear regression model are optimized to obtain the first analysis model. Based on the first partitioning data and the second partitioning data, the parameters of the preset random forest model are optimized to obtain the second analysis model. Time-series monitoring data of the bovine stomach is collected through an implanted gastric capsule; Based on the first analysis model, the monitoring time series data is processed to obtain a first analysis value, and based on the second analysis model, the monitoring time series data is processed to obtain a second analysis value. Calculate the deviation rate between the first analytical value and the second analytical value; When the deviation rate is less than the preset deviation threshold, the health status of the cow is generated based on the first analysis value and the second analysis value.

[0008] Optionally, in a first implementation of the first aspect of the present invention, the step of optimizing the parameters of a preset linear regression model based on the first partitioned data and the second partitioned data to obtain a first analytical model includes: The first partitioned data is set as the training set, and the second partitioned data is set as the validation set. The preset linear regression model is then subjected to preliminary training to obtain a preliminary linear regression model. The hyperparameters of the preliminary linear regression model are adjusted using a pre-defined grid search algorithm to obtain the first analytical model.

[0009] Optionally, in a second implementation of the first aspect of the present invention, the step of optimizing the parameters of the preset random forest model based on the first partitioning data and the second partitioning data to obtain the second analysis model includes: The second partitioned data is set as the training set, and the first partitioned data is set as the validation set. The preset random forest model is then subjected to preliminary training to obtain a preliminary random forest model. The hyperparameters of the preliminary random forest model are adjusted using a pre-defined grid search algorithm to obtain the second analysis model.

[0010] Optionally, in a third implementation of the first aspect of the present invention, the step of generating the health status of the cow based on the first analysis value and the second analysis value includes: According to the preset data acquisition algorithm, the confidence value is obtained by taking values ​​from the interval data of the first analysis value and the second analysis value. Based on the confidence value, the health status of the cow is generated.

[0011] Optionally, in a fourth implementation of the first aspect of the present invention, after the step of calculating the deviation rate between the first analytical value and the second analytical value, the method further includes: When the deviation rate is not less than the preset deviation threshold, the first analysis model and the second analysis model are fine-tuned based on the monitoring time series data to generate a fine-tuned first analysis model and a fine-tuned second analysis model.

[0012] Optionally, in a fifth implementation of the first aspect of the present invention, after the step of collecting characteristic time-series data of the bovine stomach via an implanted gastric capsule, and before the step of performing health status annotation processing on the characteristic time-series data to obtain annotated characteristic data, the method further includes: Missing values ​​in the feature time series data are filled to obtain the first preprocessed data; Calculate the normalized standard deviation for each data point in the first preprocessed data to obtain a set of standard deviations; Iterate through the set of standard deviations and determine whether the standard deviations are less than a preset deviation threshold. When the deviation is less than a preset deviation threshold, the data corresponding to the standard deviation values ​​less than the preset deviation threshold in the first preprocessed data are extracted and written into a preset second preprocessing table to obtain the second preprocessed data. The second preprocessed data is subjected to timestamp difference alignment to obtain the third preprocessed data; The third preprocessed data is normalized to obtain preprocessed feature time series data.

[0013] Optionally, in a sixth implementation of the first aspect of the present invention, the step of performing health status annotation processing on the feature time-series data to obtain annotated feature data includes: Based on a pre-set health label table, the time-series data of the features are matched and labeled to obtain labeled feature data.

[0014] Optionally, in a sixth implementation of the first aspect of the present invention, the step of performing health status annotation processing on the feature time-series data to obtain annotated feature data includes: Based on a pre-set health label table, the time-series data of the features are matched and labeled to obtain labeled feature data.

[0015] Optionally, in the seventh implementation of the first aspect of the present invention, the step of collecting characteristic temporal data of the bovine stomach via an implanted gastric capsule includes: By implanting a gastric capsule, we collected temporal data on gastric peristalsis, feeding, pH, and resting in the bovine stomach.

[0016] A second aspect of the present invention provides a health detection device based on bovine stomach data, comprising: a memory and at least one processor, wherein the memory stores instructions, and the memory and the at least one processor are interconnected via a circuit; the at least one processor invokes the instructions in the memory to cause the health detection device based on bovine stomach data to perform the aforementioned health detection method based on bovine stomach data.

[0017] A third aspect of the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the above-described health detection method based on bovine stomach data.

[0018] In this embodiment of the invention, physiological data of the bovine stomach is collected through an implanted gastric capsule. The physiological data is then labeled by tracking and statistically analyzing the correlation between bovine stomach motility data and health. The labeled physiological data is input into a linear regression model and a random forest model for training, respectively, to obtain a first analytical model and a second analytical model. These two trained analytical models are then used to perform health analysis on the newly collected bovine stomach physiological data from the implanted gastric capsule, comparing the deviations between the generated first and second analytical values. With a low deviation rate, this method achieves accurate detection of bovine health status based on bovine stomach physiological data, solving the technical problems of instability and inaccuracy in bovine health monitoring results. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of one embodiment of the health detection method based on bovine stomach data in this invention. Figure 2 This is a schematic diagram of a specific embodiment of the 104 steps of the health detection method based on bovine stomach data in this invention. Figure 3 This is a schematic diagram of a specific embodiment of the 105 steps of the health detection method based on bovine stomach data in this invention. Figure 4 This is a schematic diagram of a specific embodiment of the 109 steps of the health detection method based on bovine stomach data in this invention. Figure 5 This is a schematic diagram of one embodiment of a health detection device based on bovine stomach data in this invention. Detailed Implementation

[0020] This invention provides a health detection method, device, and storage medium based on bovine stomach data.

[0021] The embodiments of the present invention will now be described in more detail with reference to the accompanying drawings. While some embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the accompanying drawings and embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0022] In the description of the embodiments disclosed in this invention, the term "comprising" and similar terms should be understood as open-ended inclusion, i.e., "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc., may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0023] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 One embodiment of the health detection method based on bovine stomach data in this invention includes: 101. Collect characteristic temporal data of the bovine stomach by implanting a gastric capsule; In this embodiment, bovine rumen motility is a complex process and a physiological activity that follows a certain pattern. The biggest difference between cattle and ruminants, which are ruminants and monogastric animals, lies in the rumen. Regulating rumen motility to control ruminant nutrient metabolism is key to studying ruminant nutrition. Rumination is the most characteristic digestive mode of ruminants. Ruminosis is regulated by the ascending system and body fluids, and is also influenced by feed palatability, feed amount, ruminant activity, medications, and other factors. The periodic movement of the rumen allows the digesta within it to move repeatedly according to a certain pattern, ensuring sufficient fermentation of the feed entering the rumen to supply the nutrients needed for the animal's growth and development. In actual production, monitoring the frequency and intensity of rumen motility can help diagnose acute and chronic diseases and physiological abnormalities during digestion. Measuring various indicators of rumen motility can also assess the efficiency of feed fermentation in the rumen, providing a theoretical basis for the healthy breeding of ruminants. The implantable gastric capsule has an IP68 sealing performance, is corrosion-resistant, non-toxic, non-irritating, and does not cause allergic reactions. It will not get stuck in the cow's digestive tract or be expelled from the body. The signal can penetrate the cow's body and be received by the base station. It has low power consumption and long life.

[0024] The implantable gastric capsule uses a 3D motion sensor to continuously collect physiological data from the bovine stomach, and the physiological data is arranged in a time series.

[0025] Specifically, step 101 includes the following specific implementation methods: 1011. By implanting a gastric capsule, collect gastric peristalsis time-series data, feeding time-series data, pH time-series data, and resting time-series data from the bovine stomach.

[0026] In step 1011, an implantable gastric capsule is used to collect gastric peristalsis time-series data, feeding time-series data, pH time-series data, and resting time-series data from the bovine stomach. The above-mentioned relevant data are aligned with timestamps to generate a list of data.

[0027] 102. Perform health status annotation processing on the aforementioned feature time-series data to obtain annotated feature data; In this embodiment, the health status of the feature time series data can be labeled according to the following: healthy, abnormally healthy, health with a downward trend, unhealthy with poor exercise, early disease state, disease state, disease aggravation state, near death state, and death state, respectively, and given with status labels of 5, 4.5, 4, 3.5, 3, 2, 1, 0.5, and 0, respectively, to obtain labeled feature data.

[0028] Furthermore, after step 101 and before step 102, the following is also included: 1021. Fill in the missing values ​​of the feature time series data to obtain the first preprocessed data; 1022. Calculate the normalized standard deviation for each data point in the first preprocessed data to obtain a set of standard deviations; 1023. Iterate through the set of standard deviations and determine whether the standard deviations are less than a preset deviation threshold. 1024. When the deviation is less than the preset deviation threshold, extract the data corresponding to the standard deviation value less than the preset deviation threshold from the first preprocessed data and write it into the preset second preprocessing table to obtain the second preprocessed data. 1025. Perform timestamp difference alignment on the second preprocessed data to obtain the third preprocessed data; 1026. Normalize the third preprocessed data to obtain preprocessed feature time series data.

[0029] In steps 1021-1026, after obtaining the feature time series data, the feature time series data is preprocessed. The fillna() function can be used to fill missing values ​​in the feature time series data with 0, and the missing values ​​can also be copied from the previous time series data as the filling method, resulting in the first preprocessed data.

[0030] Write the first preprocessed data into a table, and then calculate the normalized standard deviation for each category of time series data, as shown in the following formula:

[0031] , where μ is the mean of time series data for each category, x is a time series data for each category, σ is the standard deviation of time series data for each category, and z is the normalized standard deviation of the output, representing the degree to which time series data for each category deviates from the mean.

[0032] Using the index, we iterate through the time series data of each category from beginning to end to analyze whether the standard deviation z is less than the deviation threshold 3. If it is less than the deviation threshold 3, we write the time series data of each category corresponding to the deviation threshold 3 into the second preprocessing table. This process continues until all data of all categories are written into the second preprocessing table, resulting in the second preprocessed data.

[0033] The second preprocessed data can use the index as a timestamp and align the differences between the timestamps. Alternatively, the second preprocessed data can be set with corresponding sequence numbers for the timestamps, and the sequence numbers can be aligned to obtain the third preprocessed data.

[0034] Finally, the StandardScaler() function is used to normalize the range of the values ​​in the third preprocessed data to obtain the preprocessed feature time series data.

[0035] Specifically, step 102 includes the following specific implementation methods: 1027. Based on the preset health label table, perform matching and annotation processing on the feature time series data to obtain labeled feature data.

[0036] In step 1027, statistical data is used to attach state labels corresponding to 5, 4.5, 4, 3.5, 3, 2, 1, 0.5, and 0 to each physiological state data, resulting in a health label table. Using this health label table, the feature time-series data is matched and labeled, and the physiological data within the feature time-series data are matched and labeled to obtain labeled feature data.

[0037] 103. According to the preset division parameters, the labeled feature data is divided to obtain the first division data and the second division data; In this embodiment, the partitioning parameter is the partitioning ratio of the labeled feature data. The first and second partitioning data will be inverted subsequently to avoid overfitting. Therefore, the partitioning parameter is best set in the range of 0.3-0.7 to avoid either partitioning value being too small or too large.

[0038] 104. Based on the first and second partitioned data, the parameters of the preset linear regression model are optimized to obtain the first analytical model; In this embodiment, either the first or second partition of the data is used as the training set, and the other partition is used as the validation set. After setting the hyperparameters, the linear regression model is trained, and the parameters of the linear regression model are optimized and adjusted to obtain the first analysis model.

[0039] For details, please refer to Figure 2 , Figure 2 This is a specific embodiment of step 104 of the health detection method based on bovine stomach data in this invention. Step 104 includes the following specific implementation methods: 1041. Set the first partitioned data as the training set and the second partitioned data as the validation set, and perform preliminary training on the preset linear regression model to obtain a preliminary linear regression model; 1042. Using a preset grid search algorithm, the hyperparameters of the preliminary linear regression model are adjusted to obtain the first analytical model.

[0040] In steps 1041-1042, the positions are fixed, the first partition of data is set as the training set, and the second partition of data is set as the validation set. The linear regression model is initialized first, and then the first partition of data is fitted to the linear regression model in the scikit-learn framework.

[0041] Then, the GridSearchCV function is introduced, and a preset grid search algorithm is used to adjust the hyperparameters of the preliminary linear regression model to obtain the first analysis model.

[0042] 105. Based on the first partitioning data and the second partitioning data, perform parameter optimization on the preset random forest model to obtain the second analysis model; In this embodiment, either the first or second partition of the data is used as the training set, and the other partition is used as the validation set. After setting the hyperparameters, the random forest model is trained, and the parameters of the random forest model are optimized and adjusted to obtain the second analysis model.

[0043] For details, please refer to Figure 3 , Figure 3 This is a specific embodiment of step 105 of the health detection method based on bovine stomach data in this invention. Step 105 includes the following specific implementation methods: 1051. Set the second partitioned data as the training set and the first partitioned data as the validation set, and perform preliminary training on the preset random forest model to obtain a preliminary random forest model; 1052. Using a pre-set grid search algorithm, the hyperparameters of the preliminary random forest model are adjusted to obtain the second analysis model.

[0044] In steps 1051-1052, to avoid overfitting between the two models, the training and validation sets of the random forest model are reversed compared to the linear regression model. The second partition of data is set as the training set, and the first partition of data is set as the validation set. Under the scikit-learn framework, the random forest model is initialized and its parameters are trained to obtain the initially trained model.

[0045] Then, the GridSearchCV function is introduced, and a preset grid search algorithm is used to adjust the hyperparameters of the initial random forest model to obtain the second analysis model.

[0046] The first and second analytical models differ in their data sources and models. They learn from the physiological data of the cow's stomach and correlate this data with its health status. The first and second analytical models can mutually verify the cow's health data, improve the accuracy of detecting the cow's health status, reduce the error in status judgment caused by data fluctuations, and improve the anti-interference ability of the detection results.

[0047] 106. Time-series monitoring data of the bovine stomach is collected through an implanted gastric capsule; In this embodiment, after the model is trained, physiological data of the bovine stomach is collected as monitoring time-series data through an implanted gastric capsule.

[0048] 107. Based on the first analysis model, the monitoring time series data is processed to obtain a first analysis value, and based on the second analysis model, the monitoring time series data is processed to obtain a second analysis value; In this embodiment, the first analysis model and the second analysis model are used to detect the monitoring time series data respectively, and the corresponding values ​​of the starting health label are obtained. For example, a first analysis value of 4.5 indicates that the cow is in an abnormal health state, and a second analysis value of 4 indicates that the cow is in a health state with a downward trend.

[0049] 108. Calculate the deviation rate between the first analytical value and the second analytical value; In this embodiment, the deviation rate is set to 25%, meaning the absolute value of the difference between the first and second analytical values ​​is at the average of the first and second analytical values. The deviation between the analytical values ​​should not be too large, otherwise two contradictory results will occur. It should be noted that the deviation rate between the first and second analytical values ​​can be calculated using other methods; the purpose is to measure the degree of deviation between the first and second analytical values.

[0050] Furthermore, following step 108, the following specific implementation methods are also included: 1081. When the deviation rate is not less than the preset deviation threshold, the first analysis model and the second analysis model are fine-tuned based on the monitoring time series data to generate a fine-tuned first analysis model and a fine-tuned second analysis model.

[0051] In this embodiment, if the calculated deviation rate exceeds the set deviation threshold, the monitoring time-series data can be manually verified and labeled with a health status. The labeled monitoring time-series data is then merged into the previously trained labeled feature data, and a time window (e.g., 6 hours or 3 hours) is added. The merged data is then split according to the time window. Using the split window time-series data, the first and second analysis models are fine-tuned to obtain the finely adjusted first and second analysis models after the data source update. This avoids limitations imposed by the training data source and improves the analysis models' ability to handle data. 109. When the deviation rate is less than the preset deviation threshold, the health status of the cow is generated based on the first analysis value and the second analysis value.

[0052] In this embodiment, if the deviation rate is less than the deviation threshold, the actual value is considered to be within the interval between the first and second analysis values. If there is no intermediate value within the interval between the first and second analysis values, the smallest value is selected as the corresponding health status of the cow. If there is an intermediate value within the interval between the first and second analysis values, the intermediate value is selected as the corresponding health status of the cow.

[0053] For details, please refer to Figure 4 , Figure 4 This is a specific embodiment of step 109 of the health detection method based on bovine stomach data in this invention. Step 109 includes the following specific implementation methods: 1091. According to the preset data acquisition algorithm, the interval data of the first analysis value and the second analysis value are taken to obtain the confidence value; 1092. Based on the confidence value, generate the corresponding health status of the cow.

[0054] In steps 1091-1092, the data retrieval algorithm can be implemented by writing all state values ​​into an ordered list, then matching the corresponding index value in the list, and analyzing whether there is a difference greater than 1 between the index values. If a difference greater than 1 exists, the difference divided by 2 and rounded down is added to the smallest index to obtain a data retrieval index. The state value corresponding to the data retrieval index is then identified as the confidence value. Finally, based on the confidence value, the corresponding health status in the health status table is retrieved, and the health status of the cow is determined.

[0055] In this embodiment of the invention, physiological data of the bovine stomach is collected through an implanted gastric capsule. The physiological data is then labeled by tracking and statistically analyzing the correlation between bovine stomach motility data and health. The labeled physiological data is input into a linear regression model and a random forest model for training, respectively, to obtain a first analytical model and a second analytical model. These two trained analytical models are then used to perform health analysis on the newly collected bovine stomach physiological data from the implanted gastric capsule, comparing the deviations between the generated first and second analytical values. With a low deviation rate, this method achieves accurate detection of bovine health status based on bovine stomach physiological data, solving the technical problems of instability and inaccuracy in bovine health monitoring results.

[0056] Figure 5 This is a schematic diagram of a health detection device 500 based on bovine stomach data, provided in an embodiment of the present invention. The health detection device 500 can vary significantly due to different configurations or performance characteristics. It may include one or more central processing units (CPUs) 510 (e.g., one or more processors) and a memory 520, and one or more storage media 530 (e.g., one or more mass storage devices) for storing application programs 533 or data 532. The memory 520 and storage media 530 can be temporary or persistent storage. The program stored in the storage media 530 may include one or more modules (not shown in the diagram), each module including a series of instruction operations on the health detection device 500 based on bovine stomach data. Furthermore, the processor 510 may be configured to communicate with the storage media 530 and execute the series of instruction operations in the storage media 530 on the health detection device 500 based on bovine stomach data.

[0057] The health monitoring device 500 based on bovine stomach data may also include one or more power supplies 540, one or more wired or wireless network interfaces 550, one or more input / output interfaces 560, and / or one or more operating systems 531, such as Windows Server, Mac OS X, Unix, Linux, Free BSD, etc. Those skilled in the art will understand that... Figure 5 The illustrated structure of a health monitoring device based on bovine stomach data does not constitute a limitation on health monitoring devices based on bovine stomach data. It may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0058] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the steps of the health detection method based on bovine stomach data.

[0059] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0060] Furthermore, although the operations are described in a specific order, this should be understood as requiring that such operations be performed in the specific order shown or in sequential order, or requiring that all illustrated operations be performed to achieve the desired result. In certain environments, multitasking and parallel processing may be advantageous. Similarly, although several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of this disclosure. Certain features described in the context of individual embodiments may also be implemented in combination in a single implementation. Conversely, various features described in the context of a single implementation may also be implemented individually or in any suitable sub-combination in multiple implementations.

[0061] Although the subject matter has been described using language specific to structural features and / or methodological logic, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are merely illustrative examples of implementing the claims.

Claims

1. A health detection method based on bovine stomach data, characterized in that, Including the following steps: By implanting a gastric capsule, characteristic temporal data of the bovine stomach were collected; The health status annotation process is performed on the aforementioned feature time series data to obtain annotated feature data; The labeled feature data is divided according to preset division parameters to obtain first division data and second division data; Based on the first and second partition data, the parameters of the preset linear regression model are optimized to obtain the first analysis model. Based on the first partition data and the second partition data, the parameters of the preset random forest model are optimized to obtain the second analysis model. Time-series monitoring data of the bovine stomach is collected through an implanted gastric capsule; Based on the first analysis model, the monitoring time series data is processed to obtain a first analysis value, and based on the second analysis model, the monitoring time series data is processed to obtain a second analysis value. Calculate the deviation rate between the first analytical value and the second analytical value; When the deviation rate is less than the preset deviation threshold, the health status of the cow is generated based on the first analysis value and the second analysis value.

2. The health detection method based on bovine stomach data according to claim 1, characterized in that, The step of optimizing the parameters of a pre-set linear regression model based on the first and second partitioning data to obtain the first analytical model includes: The first partitioned data is set as the training set, and the second partitioned data is set as the validation set. The preset linear regression model is then subjected to preliminary training to obtain a preliminary linear regression model. The hyperparameters of the preliminary linear regression model are adjusted using a pre-defined grid search algorithm to obtain the first analytical model.

3. The health detection method based on bovine stomach data according to claim 2, characterized in that, The step of optimizing the parameters of the pre-set random forest model based on the first partitioning data and the second partitioning data to obtain the second analysis model includes: The second partitioned data is set as the training set, and the first partitioned data is set as the validation set. The preset random forest model is then subjected to preliminary training to obtain a preliminary random forest model. The hyperparameters of the preliminary random forest model are adjusted using a pre-defined grid search algorithm to obtain the second analysis model.

4. The health detection method based on bovine stomach data according to claim 1, characterized in that, The step of generating the health status of the cow based on the first analysis value and the second analysis value includes: According to the preset data acquisition algorithm, the confidence value is obtained by taking values ​​from the interval data of the first analysis value and the second analysis value. Based on the confidence value, the health status of the cow is generated.

5. The health detection method based on bovine stomach data according to claim 1, characterized in that, After the step of calculating the deviation rate between the first analytical value and the second analytical value, the method further includes: When the deviation rate is not less than the preset deviation threshold, the first analysis model and the second analysis model are fine-tuned based on the monitoring time series data to generate a fine-tuned first analysis model and a fine-tuned second analysis model.

6. The health detection method based on bovine stomach data according to claim 1, characterized in that, After the step of collecting characteristic time-series data of the bovine stomach via an implanted gastric capsule, and before the step of performing health status annotation processing on the characteristic time-series data to obtain annotated characteristic data, the method further includes: Missing values ​​in the feature time series data are filled to obtain the first preprocessed data; Calculate the normalized standard deviation for each data point in the first preprocessed data to obtain a set of standard deviations; Iterate through the set of standard deviations and determine whether the standard deviations are less than a preset deviation threshold. When the deviation is less than a preset deviation threshold, the data corresponding to the standard deviation values ​​less than the preset deviation threshold in the first preprocessed data are extracted and written into a preset second preprocessing table to obtain the second preprocessed data. The second preprocessed data is subjected to timestamp difference alignment to obtain the third preprocessed data; The third preprocessed data is normalized to obtain preprocessed feature time series data.

7. The health detection method based on bovine stomach data according to claim 1, characterized in that, The step of performing health status annotation processing on the feature time series data to obtain annotated feature data includes: Based on a pre-set health label table, the time-series data of the features are matched and labeled to obtain labeled feature data.

8. The health detection method based on bovine stomach data according to claim 1, characterized in that, The step of collecting characteristic temporal data of the bovine stomach via an implanted gastric capsule includes: By implanting a gastric capsule, we collected temporal data on gastric peristalsis, feeding, pH, and resting in the bovine stomach.

9. A health monitoring device based on bovine stomach data, characterized in that, The health detection device based on bovine stomach data includes: a memory and at least one processor, wherein the memory stores instructions, and the memory and the at least one processor are interconnected via a circuit; The at least one processor invokes the instructions in the memory to cause the health detection device based on bovine stomach data to perform the health detection method based on bovine stomach data as described in any one of claims 1-8.

10. A computer-readable storage medium storing a computer program thereon, characterized in that, When the computer program is executed by the processor, it implements the health detection method based on bovine stomach data as described in any one of claims 1-8.

Images