Animal husbandry livestock health management and early warning method and system based on Internet of Things

Abstract

The invention relates to the technical field of livestock health management and early warning, and particularly discloses a livestock health management and early warning method and system for animal husbandry based on the Internet of Things, and the method mainly comprises the steps: obtaining the GPS data and monitoring data of livestock, the GPS data comprising a timestamp and a GPS coordinate; calculating an activity distance of the livestock in a specified time period based on the GPS data; judging whether the activity distance belongs to a preset interval or not, and if yes, identifying the livestock from the monitoring data of the specified time period as the livestock to be monitored; and extracting behavior characteristics of the to-be-monitored livestock from the monitoring data of the specified time period, and judging state information of the behavior characteristics. According to the system, the Internet of Things and the data analysis technology are adopted, intelligent monitoring of the livestock activity state is achieved by obtaining the GPS data and the monitoring data of the livestock in real time, the activity and behavior characteristics of the livestock can be effectively monitored, abnormal conditions can be found in time, early warning is made, the burden of manual monitoring is reduced, and the overall efficiency of pasture management is improved.

Description

Technical Field

[0001] The present application relates to the field of livestock health management and early warning technology, and in particular to a livestock health management and early warning method and system based on the Internet of Things in animal husbandry. Background Art

[0002] The livestock industry's health management and early warning system uses simple monitoring methods, such as regular observation and collection of basic health data of animals (such as appetite, appearance changes, etc.), combined with historical data analysis to identify abnormal situations. Once a problem is found, the system will issue an alarm to help breeders deal with it in a timely manner and reduce disease risks. This can improve the efficiency of animal health management while controlling costs.

[0003] Relying on regular observation and basic data collection can easily lead to missing important health indicators. Regular observation and collection of basic health data of animals may miss signs of early disease, thereby delaying treatment. If comprehensive monitoring is to be achieved, it involves the collection of more physiological data, such as body temperature, heart rate, etc. The amount of data will increase significantly, which not only requires more sensors and equipment, resulting in increased costs, but also increases the complexity of data storage and processing, increasing management difficulty. As the amount of monitoring data increases, the cost of data analysis and manual intervention will also increase. Farmers may face pressure on technology investment and maintenance costs, which challenges the economic feasibility of the system. Summary of the invention

[0004] The present application provides a livestock health management and early warning method and system based on the Internet of Things, which solves the problem in the prior art that reliance on regular observation and basic data collection can easily lead to the omission of important health indicators, while comprehensive monitoring can easily lead to a significant increase in data volume and costs, and also lead to increased complexity in data storage and processing. Through the combined analysis of GPS data and monitoring data, the activities and behavioral characteristics of livestock can be effectively monitored, abnormal situations can be discovered in time and early warnings can be issued, reducing the burden of human monitoring. The introduction of the prediction model enables the system to predict potential health problems in advance, prevent them before they occur, and improve the overall efficiency of pasture management.

[0005] In the first aspect, the present application provides a livestock health management and early warning method for animal husbandry based on the Internet of Things, which includes the following steps: obtaining GPS data and monitoring data of livestock, the GPS data including timestamps and GPS coordinates; calculating the activity distance of livestock in a specified time period based on the GPS data; judging whether the activity distance belongs to a preset interval, and if so, identifying the livestock as livestock to be monitored from the monitoring data of the specified time period; extracting behavioral characteristics of the livestock to be monitored from the monitoring data of the specified time period, and judging the status information of the behavioral characteristics, the status information including abnormal, possibly abnormal and normal; if the status information is abnormal, sending an early warning message to the management personnel; if the status information is possibly abnormal, inputting the behavioral characteristics into the prediction model to obtain the predicted behavioral characteristics of the livestock to be monitored, and judging whether the predicted behavioral characteristics are abnormal, and if so, sending an early warning message to the management personnel.

[0006] Furthermore, each livestock is equipped with a GPS device, each GPS device has a unique identifier, and each unique identifier corresponds to each livestock.

[0007] Furthermore, the monitoring data covers the activity range of livestock.

[0008] Furthermore, calculating the activity distance of livestock in a specified time period based on the GPS data specifically includes: sorting the GPS coordinates according to timestamps; traversing each pair of continuous coordinate points in the specified time period, using the Haversing formula to calculate the geographical distance between each pair of continuous coordinate points, and accumulating the activity distance.

[0009] Furthermore, judging whether the activity distance belongs to the preset interval specifically includes: judging the season corresponding to the current time period; judging whether the livestock has activity data of the corresponding historical season in a healthy state, if so, executing step a, if not, executing step b; step a: calculating the average value avg1 of the livestock in the historical corresponding season in a healthy state, and taking (k1*avg1, k2*avg1) as the preset interval, wherein k1 and k2 represent coefficients used to set the lower limit and upper limit of the preset interval respectively. Step b: calculating the average value avg2 of the activity data of all livestock in the historical corresponding season in a healthy state, and taking (k1*avg2, k2*avg2) as the preset interval.

[0010] Furthermore, identifying the livestock to be monitored from the monitoring data of the specified time period specifically includes: using a video processing tool to read the video frame of the monitoring data and extracting the GPS coordinates of the livestock to be monitored; converting the GPS coordinates into pixel coordinates in the video frame; using a pre-trained deep learning model to perform object detection on each video frame, identifying the position of the livestock and marking the bounding box, and determining the center of the bounding box as the pixel coordinate of the corresponding livestock; traversing each bounding box, calculating the distance between the center point of the bounding box and the GPS coordinates, and determining the livestock corresponding to the bounding box with the smallest distance from the GPS coordinates as the livestock to be monitored.

[0011] Furthermore, the behavior characteristics include eating time, rest time and exercise time within a specified time period.

[0012] Furthermore, each video frame is classified to identify the behavior of the livestock to be monitored, including eating, resting and exercising; each video frame is traversed to record the start and end time of each behavior; the total time of each behavior is counted to obtain the eating time, resting time and exercise time of the livestock.

[0013] Further, judging the state information of the behavior characteristics specifically includes: judging the season corresponding to the current time period; judging whether the livestock has the behavior characteristics of the corresponding historical season in a healthy state, if so, executing step c, if not, executing step d; step c: calculating the mean value and standard deviation of the eating time, rest time and exercise time of the livestock in the historical season in a healthy state, and establishing the normal range interval, possible abnormal interval and abnormal interval of the eating time, rest time and exercise time based on the mean value and standard deviation; step d: calculating the mean value and standard deviation of the eating time, rest time and exercise time of all livestock in the historical season in a healthy state, and establishing the normal range interval, possible abnormal interval and abnormal interval of the eating time, rest time and exercise time based on the mean value and standard deviation; Determine the interval to which the predicted behavior feature belongs to obtain the state information of the behavior feature.

[0014] In the second aspect, the present application provides an Internet of Things-based livestock health management and early warning system for animal husbandry, which adopts the Internet of Things-based livestock health management and early warning method for animal husbandry as described in the first aspect, and includes: a data acquisition module, an active monitoring module, an abnormal livestock identification module and a health status assessment module.

[0015] The data acquisition module is used to obtain GPS data and monitoring data of livestock; the GPS data includes timestamps and GPS coordinates; the activity monitoring module is used to calculate the activity distance of livestock in a specified time period based on the GPS data; the abnormal livestock identification module is used to determine whether the activity distance belongs to a preset interval, and if so, identify the livestock as livestock to be monitored from the monitoring data of the specified time period; the health status assessment module is used to extract behavioral characteristics of livestock to be monitored from the monitoring data of the specified time period, and determine the status information of the behavioral characteristics, and the status information includes abnormal, possibly abnormal and normal; if the status information is abnormal, an early warning message is sent to the management personnel; if the status information is possibly abnormal, the behavioral characteristics are input into the prediction model to obtain the predicted behavioral characteristics of the livestock to be monitored, and determine whether the predicted behavioral characteristics are abnormal, and if so, an early warning message is sent to the management personnel.

[0016] On the third aspect, the present application also provides an Internet of Things-based livestock health management and early warning device for animal husbandry, which includes a memory and a processor.

[0017] The memory is used to store computer programs; the processor is used to implement the steps of the livestock health management and early warning method based on the Internet of Things in the first aspect when executing the computer program.

[0018] The technical solution provided by this application has at least the following technical effects or advantages: The Internet of Things and data analysis technology are used to obtain livestock GPS data and monitoring data in real time to achieve intelligent monitoring of livestock activity status. This solves the problem that the existing technology relies on regular observation and basic data collection, which easily leads to the omission of important health indicators, and comprehensive monitoring easily leads to a significant increase in data volume and costs, and also increases the complexity of data storage and processing. Through the combined analysis of GPS data and monitoring data, the activities and behavioral characteristics of livestock can be effectively monitored, abnormal situations can be discovered in time and early warnings can be issued, reducing the burden of human monitoring. The introduction of predictive models enables the system to predict potential health problems in advance, prevent them before they occur, and improve the overall efficiency of pasture management. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a flow chart of the livestock health management and early warning method for animal husbandry based on the Internet of Things in Example 1 of the present application; Figure 2 For this application Figure 1 A flow chart of determining whether the activity distance belongs to a preset interval; Figure 3 For this application Figure 1 A schematic diagram of a process for identifying livestock to be monitored from monitoring data of a specified time period; Figure 4For this application Figure 1 Schematic diagram of the process of judging the state information of behavioral characteristics. DETAILED DESCRIPTION

[0020] This application optimizes the data processing process and introduces advanced prediction models to solve the defects in the existing technology mentioned in the background technology, thereby realizing real-time monitoring and early warning of livestock health status, improving livestock health management efficiency, reducing labor costs, realizing real-time monitoring of livestock health status, timely discovering problems and taking measures, reducing the occurrence of livestock diseases, and improving breeding efficiency.

[0021] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0022] Example 1: Please refer to Figure 1-Figure 4 This embodiment provides a livestock health management and early warning method based on the Internet of Things, which includes the following steps: S100. Obtain GPS data and monitoring data of livestock, where the GPS data includes a timestamp and GPS coordinates.

[0023] GPS sensors installed on livestock can collect animal location information and activity data in real time. These sensors are connected to the Internet and can transmit data to the central management system. IoT technology allows farmers to access data through the Internet anytime and anywhere to monitor the health status and activity of livestock.

[0024] By obtaining GPS data and monitoring data of livestock, comprehensive monitoring of all livestock can be achieved. GPS technology is relatively low-cost and easy to deploy, and can obtain the location and activity status of each livestock in real time. The popularity of GPS equipment enables managers to collect data efficiently, reducing the time and labor costs of manual inspections, thereby achieving wider monitoring coverage and improving overall management efficiency.

[0025] S200. Calculate the activity distance of livestock in a specified time period based on GPS data.

[0026] S300. Determine whether the activity distance belongs to a preset interval. If so, identify livestock as livestock to be monitored from monitoring data of a specified time period.

[0027] By calculating the distance that livestock move during a specified time period, we can effectively monitor their activity level and determine their health status. Under normal circumstances, healthy livestock will have a certain amount of activity, such as walking freely in the pasture and foraging. However, when the activity distance of a livestock is significantly lower than the preset normal range, it may indicate that it has health problems, such as illness, injury or poor mental state, and requires further monitoring.

[0028] For example, if a cow has only moved half its usual distance in the past 24 hours, this may be a direct sign of physical discomfort. Through this activity-based monitoring method, managers can quickly identify potential abnormalities and take timely intervention measures, thereby improving the accuracy and efficiency of monitoring and ensuring the health of livestock.

[0029] S400. Extracting the behavior characteristics of the livestock to be monitored from the monitoring data of the specified time period, and determining the state information of the behavior characteristics, where the state information includes abnormal, possibly abnormal, and normal.

[0030] By extracting the behavioral characteristics of the livestock to be monitored from the monitoring data, it is possible to further analyze their status information, significantly improve the accuracy of the monitoring results, and provide a more comprehensive health assessment. For example, if a livestock's eating behavior decreases, it may indicate that it has potential health problems, and the livestock's status can be judged more accurately, reducing the risk of misjudgment.

[0031] S410. If the status information is abnormal, send an early warning message to the management personnel.

[0032] S420. If the status information is possibly abnormal, the behavior characteristics are input into the prediction model to obtain the predicted behavior characteristics of the livestock to be monitored, and it is determined whether the predicted behavior characteristics are abnormal. If so, an early warning message is sent to the management personnel.

[0033] By collecting and analyzing livestock activity data, health indicators and behavioral characteristics in real time, individuals that need special attention can be automatically identified, greatly reducing the workload of managers. Managers do not need to frequently check all livestock. The system will automatically issue early warning information based on preset health standards and behavioral abnormalities, making resource allocation more reasonable. Managers can focus on individuals that really need intervention, improving management efficiency.

[0034] In this embodiment, the GPS data and monitoring data of livestock are obtained, the activity distance of livestock in a specific time period is calculated using the GPS data, and whether the calculated activity distance is within a preset interval is determined based on the calculated activity distance to identify the livestock that need to be monitored, and then the behavioral characteristics of the livestock to be monitored are extracted from the monitoring data, and their status information is analyzed. If the behavioral characteristics show abnormalities, the system will promptly issue an early warning message to the management personnel. If the behavioral characteristics may be abnormal, the system will further use a predictive model to analyze the behavior to determine whether it is an abnormal state, thereby improving the accuracy and timeliness of animal husbandry management.

[0035] This embodiment scheme realizes automated health monitoring and early warning of livestock, improves the efficiency and accuracy of animal husbandry management, and through the combined analysis of GPS data and monitoring data, can effectively monitor the activities and behavioral characteristics of livestock, detect abnormal situations in time and issue early warnings, reducing the burden of human monitoring. The introduction of the prediction model enables the system to predict potential health problems in advance, prevent them before they occur, and improve the overall efficiency of pasture management.

[0036] In S100, a GPS device is installed on each livestock, each GPS device has a unique identifier, and each unique identifier corresponds to each livestock.

[0037] By installing a GPS device on each livestock, we ensure that the unique identification of each device can accurately correspond to the specific livestock, thereby achieving accurate positioning and data tracking of individual livestock. The data collected by the GPS device is combined with the monitoring system to monitor the activity status of livestock in real time, and the activity data and behavioral characteristics of each livestock are matched to it through a unique identification to ensure the accuracy and relevance of the data.

[0038] This embodiment scheme realizes accurate management of individual livestock, helps to improve the accuracy of monitoring data and the overall reliability of the system, and managers can perform accurate health management and early warning based on the GPS data of specific livestock, thereby optimizing the management process and decision-making efficiency of the ranch.

[0039] In S100 , the monitoring data covers the activity range of the livestock.

[0040] By deploying monitoring equipment within the livestock's activity range, we ensure that the collected monitoring data can fully cover the livestock's activity area, ensure that the behavioral characteristics of each livestock can be captured in real time, and correlated with GPS data for analysis, and track the behavioral characteristics of each livestock in real time, thereby improving the accuracy and timeliness of health monitoring and early warning. By covering the entire activity range of livestock, the emergence of monitoring blind spots is avoided, ensuring that the system can effectively monitor and manage all livestock in activities.

[0041] In S200, calculating the activity distance of livestock in a specified time period based on GPS data specifically includes: sorting GPS coordinates according to timestamps; traversing each pair of continuous coordinate points in the specified time period, using the Haversing formula to calculate the geographic distance between each pair of continuous coordinate points, and accumulating the activity distance.

[0042] This implementation scheme sorts the timestamps in the GPS data, then traverses each pair of consecutive GPS coordinate points, uses the Haversing formula to calculate the geographic distance between them, and accumulates the distances between all consecutive coordinate points to obtain the total activity distance of the livestock in a specified time period. This can accurately reflect the movement of the livestock in the specified time period, thereby providing a reliable data basis for subsequent behavioral feature analysis and health status judgment. Through accurate distance calculation, the activity level of livestock can be better monitored and possible health abnormalities can be discovered in a timely manner.

[0043] In S300, determining whether the activity distance belongs to the preset interval specifically includes: S310. Determine the season corresponding to the current time period.

[0044] S320. Determine whether the livestock has activity data of the corresponding historical season in a healthy state. If so, execute step a; if not, execute step b.

[0045] Step a: Calculate the average value avg1 of the livestock in the corresponding season in the history of health status, and use (k1*avg1, k2*avg1) as the preset interval, where k1 and k2 represent coefficients used to set the lower limit and upper limit of the preset interval respectively.

[0046] Step b: Calculate the average value avg2 of the historical activity data of all livestock in the healthy state in the corresponding season, and use (k1*avg2, k2*avg2) as the preset interval.

[0047] In this embodiment, the corresponding season is first determined according to the current time period to consider the impact of the season on the livestock's activity behavior. Then, by comparing the historical seasonal activity data of the livestock in a healthy state, or referring to the average historical data of all livestock, a suitable preset activity distance interval is calculated. The specific method includes calculating the average activity value and setting the upper and lower limits according to the predetermined coefficients k1 and k2, so that it can effectively determine whether the current livestock activity distance is abnormal.

[0048] By comprehensively considering seasonal factors and historical activity data, this implementation can dynamically adjust the preset interval of activity distance, improving the system's sensitivity and accuracy to abnormal behavior. The dual reference of individual historical data and group data makes the system more adaptable and flexible, and can more accurately reflect the normal activity range of different livestock in different seasons.

[0049] In other embodiments, the seasonal determination algorithm may be combined with weather data or more detailed time period division to improve the accuracy of the determination.

[0050] In S300, identifying the livestock to be monitored from the monitoring data of the specified time period specifically includes: S330. Use video processing tools to read video frames of monitoring data and extract GPS coordinates of the livestock to be monitored.

[0051] S340. Convert the GPS coordinates into pixel coordinates in the video frame; use a pre-trained deep learning model to perform object detection on each video frame, identify the location of the livestock and mark the bounding box, and determine the center of the bounding box as the pixel coordinate of the corresponding livestock.

[0052] S350. Traverse each bounding box, calculate the distance between the center point of the bounding box and the GPS coordinates, and determine the livestock corresponding to the bounding box with the smallest distance from the GPS coordinates as the livestock to be monitored.

[0053] This embodiment reads monitoring data through a video processing tool, associates the GPS coordinates of livestock with the pixel coordinates in the video frame, first extracts the GPS coordinates of the livestock to be monitored and converts them into pixel coordinates in the video frame, then uses a pre-trained deep learning model to perform object detection on each video frame, identifies the livestock in the video and marks the bounding box, and determines the bounding box closest to the GPS coordinates by calculating the distance between the center point of the bounding box and the pixel coordinates, thereby accurately identifying the livestock to be monitored.

[0054] By combining GPS data with video surveillance data, livestock to be monitored can be accurately identified without the need for further monitoring of all livestock, effectively reducing the amount of calculations and thus improving monitoring efficiency. The use of deep learning models makes object detection more accurate and efficient, and can process large amounts of monitoring data in real time, thereby improving the overall performance of the livestock health management and early warning system, ensuring that the system can still accurately identify and track the activities of each livestock in a high-density monitoring environment.

[0055] In S400 , the behavior characteristics include eating time, rest time, and exercise time within a specified time period.

[0056] Each behavioral feature is identified through a deep learning model, and the system automatically counts the total time of each behavior to evaluate the health status of livestock. It can fully grasp the daily activity patterns of livestock and help to detect health abnormalities in a timely manner. Different behavioral characteristics reflect the activities of livestock in different physiological states. These data can be used to determine whether livestock are in a normal health state, thereby improving the scientificity and accuracy of pasture management.

[0057] Classify each video frame and identify the behavior of the livestock to be monitored, including eating, resting and exercising; traverse each video frame and record the start and end time of each behavior; count the total time of each behavior to obtain the livestock's eating time, resting time and exercise time.

[0058] By classifying the video frames one by one, the specific behaviors of the livestock to be monitored are identified, and the start and end time of each behavior is recorded based on the identification results. Finally, the total time of the livestock eating, resting and exercising in the specified time period is counted. Through such classification and statistics, the system can accurately grasp the daily activity time allocation of livestock and provide data support for health status assessment.

[0059] This embodiment scheme can accurately track and record the behavior patterns of livestock, providing detailed data basis for health management. The automated processing method reduces human involvement. By counting the time of eating, resting and exercising, the health status of livestock can be better judged, abnormalities can be discovered in time and early warnings can be issued, thereby effectively reducing the risk of disease and improving the production efficiency of animal husbandry.

[0060] In S400, the state information of the behavior characteristics is determined to include: S410. Determine the season corresponding to the current time period.

[0061] S420. Determine whether the livestock has the behavioral characteristics of the corresponding historical season in a healthy state. If so, execute step c; if not, execute step d.

[0062] Step c: Calculate the mean and standard deviation of the eating time, resting time and exercise time of the livestock in historical seasons when in a healthy state, and establish normal range intervals, possible abnormal intervals and abnormal intervals for eating time, resting time and exercise time based on the mean and standard deviation.

[0063] Step d: Calculate the mean and standard deviation of the eating time, resting time and exercise time of all livestock in historical seasons when they are in a healthy state, and establish the normal range interval, possible abnormal interval and abnormal interval of eating time, resting time and exercise time based on the mean and standard deviation.

[0064] S430. Determine the interval to which the predicted behavior feature belongs to thereby obtain state information of the behavior feature.

[0065] This implementation method first determines the season of the current time period, and then analyzes the behavioral characteristics of the livestock based on the data of that season. The system compares the current behavioral characteristics with the historical data, calculates the average and standard deviation, and establishes the normal range, possible abnormal range, and abnormal range of different behavioral characteristics. The current behavioral state of the livestock can be accurately determined and classified into normal, possible abnormal, or abnormal states for further processing.

[0066] This implementation method comprehensively considers seasonal factors and historical behavior data, making the system's judgment more accurate and adaptable. Through standard deviation analysis, it can not only capture obvious abnormal behaviors, but also identify potential health risks, enhancing the system's early warning function. By comparing and analyzing historical data of groups and individuals, it can effectively improve the system's detection sensitivity to abnormal behaviors, which helps to discover problems early and intervene in time.

[0067] If a problem is found, more physiological data can be collected from the livestock, such as body temperature, heart rate, etc., so as to more accurately judge the health status of the livestock.

[0068] Embodiment 2: This embodiment provides a livestock health management and early warning system based on the Internet of Things for animal husbandry, which adopts the livestock health management and early warning method based on the Internet of Things for animal husbandry as in Embodiment 1, and includes: a data acquisition module, an active monitoring module, an abnormal livestock identification module and a health status assessment module.

[0069] The data acquisition module is used to obtain GPS data and monitoring data of livestock; the GPS data includes timestamps and GPS coordinates; the activity monitoring module is used to calculate the activity distance of livestock in a specified time period based on GPS data; the abnormal livestock identification module is used to determine whether the activity distance belongs to a preset interval, and if so, identify the livestock as livestock to be monitored from the monitoring data of the specified time period; the health status assessment module is used to extract the behavioral characteristics of the livestock to be monitored from the monitoring data of the specified time period, and determine the status information of the behavioral characteristics, and the status information includes abnormal, possibly abnormal and normal; if the status information is abnormal, an early warning message is sent to the management personnel; if the status information is possibly abnormal, the behavioral characteristics are input into the prediction model to obtain the predicted behavioral characteristics of the livestock to be monitored, and it is determined whether the predicted behavioral characteristics are abnormal, and if so, an early warning message is sent to the management personnel.

[0070] This embodiment has all the advantages of the livestock health management and early warning method for animal husbandry based on the Internet of Things in Embodiment 1, and can automatically implement all steps of the livestock health management and early warning method for animal husbandry based on the Internet of Things.

[0071] Embodiment 3: This embodiment provides an animal husbandry livestock health management and early warning device based on the Internet of Things, which includes a memory and a processor.

[0072] The memory is used to store computer programs; the processor is used to implement the steps of the livestock health management and early warning method based on the Internet of Things in Example 1 when executing the computer program.

[0073] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps of the functions specified in a box or multiple boxes. Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they know the basic creative concept. Therefore, the attached claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

Claims

1. A livestock health management and early warning method based on the Internet of Things, characterized in that: It includes the following steps: Obtain GPS data and monitoring data of livestock, GPS data includes timestamps and GPS coordinates; Calculate the movement distance of the livestock in a specified time period based on the GPS data; determining whether the activity distance belongs to a preset interval, and if so, identifying the livestock as the livestock to be monitored from the monitoring data of the specified time period; Extracting the behavior characteristics of the livestock to be monitored from the monitoring data of the specified time period, and determining the state information of the behavior characteristics, wherein the state information includes abnormal, possibly abnormal and normal; If the status information is abnormal, an early warning message is sent to the management personnel; If the status information is possibly abnormal, the behavior characteristics are input into the prediction model to obtain the predicted behavior characteristics of the livestock to be monitored, and it is determined whether the predicted behavior characteristics are abnormal. If so, an early warning message is sent to the management personnel.

2. The livestock health management and early warning method based on the Internet of Things according to claim 1, characterized in that: Each livestock is equipped with a GPS device, each GPS device has a unique identifier, and each unique identifier corresponds to each livestock.

3. The livestock health management and early warning method based on the Internet of Things according to claim 1, characterized in that: The monitoring data covers the activity range of livestock.

4. The livestock health management and early warning method based on the Internet of Things according to claim 1, characterized in that: Calculating the activity distance of livestock in a specified time period based on the GPS data specifically includes: Sort GPS coordinates by timestamp; Traverse each pair of consecutive coordinate points within the specified time period, use the Haversine formula to calculate the geographical distance between each pair of consecutive coordinate points, and accumulate them to obtain the activity distance.

5. The livestock health management and early warning method based on the Internet of Things according to claim 1, characterized in that: Determining whether the activity distance belongs to the preset interval specifically includes: Determine the season corresponding to the current time period; Determine whether the livestock has activity data of the corresponding historical season in a healthy state, if so, execute step a, if not, execute step b; Step a: Calculate the average value avg1 of the livestock in the corresponding season in the health state, and use (k1*avg1, k2*avg1) as the preset interval, where k1 and k2 represent coefficients used to set the lower limit and upper limit of the preset interval respectively; Step b: Calculate the average value avg2 of the historical activity data of all livestock in the healthy state in the corresponding season, and use (k1*avg2, k2*avg2) as the preset interval.

6. The livestock health management and early warning method based on the Internet of Things according to claim 5, characterized in that: Identifying the livestock to be monitored from the monitoring data in the specified time period specifically includes: Use video processing tools to read video frames of monitoring data and extract GPS coordinates of the livestock to be monitored; Converting the GPS coordinates into pixel coordinates in a video frame; performing object detection on each video frame using a pre-trained deep learning model, identifying the location of the livestock and marking a bounding box, and determining the center of the bounding box as the pixel coordinate of the corresponding livestock; Each bounding box is traversed, the distance between the center point of the bounding box and the GPS coordinates is calculated, and the livestock corresponding to the bounding box with the smallest distance from the GPS coordinates is determined as the livestock to be monitored.

7. The livestock health management and early warning method based on the Internet of Things according to claim 6, characterized in that: The behavioral characteristics include eating time, rest time, and exercise time within a specified time period.

8. The livestock health management and early warning method based on the Internet of Things according to claim 7, characterized in that: Classify each video frame to identify the behavior of the livestock to be monitored, including eating, resting and exercising; Traverse each video frame and record the start and end time of each behavior; The total time of each behavior is counted to obtain the livestock's eating time, resting time and exercise time.

9. The livestock health management and early warning method based on the Internet of Things according to claim 1, characterized in that: The state information for determining the behavior characteristics specifically includes: Determine the season corresponding to the current time period; Determine whether the livestock has the behavioral characteristics of the corresponding historical season in a healthy state, if so, execute step c, if not, execute step d; Step c: Calculate the mean and standard deviation of the eating time, resting time and exercise time of the livestock in historical seasons when the livestock is in a healthy state, and establish normal range intervals, possible abnormal intervals and abnormal intervals of the eating time, resting time and exercise time based on the mean and standard deviation; Step d: Calculate the mean and standard deviation of the eating time, resting time and exercise time of all livestock in the historical seasons when they are in a healthy state, and establish the normal range interval, possible abnormal interval and abnormal interval of the eating time, resting time and exercise time based on the mean and standard deviation; Determine the interval to which the predicted behavior feature belongs to obtain the state information of the behavior feature.

10. A livestock health management and early warning system based on the Internet of Things, characterized in that: include: A data acquisition module, which is used to acquire GPS data and monitoring data of livestock; the GPS data includes timestamps and GPS coordinates; An activity monitoring module, which is used to calculate the activity distance of the livestock in a specified time period based on the GPS data; An abnormal livestock identification module; used to determine whether the activity distance belongs to a preset interval, and if so, to identify the livestock as the livestock to be monitored from the monitoring data of the specified time period; A health status assessment module, which is used to extract the behavior characteristics of the livestock to be monitored from the monitoring data of the specified time period, and determine the status information of the behavior characteristics, where the status information includes abnormal, possibly abnormal and normal; If the status information is abnormal, an early warning message is sent to the management personnel; If the status information is possibly abnormal, the behavior characteristics are input into the prediction model to obtain the predicted behavior characteristics of the livestock to be monitored, and it is determined whether the predicted behavior characteristics are abnormal. If so, an early warning message is sent to the management personnel.

Images