Animal husbandry monitoring method and device based on multi-sensor fusion

Abstract

The invention discloses a livestock monitoring method and device based on multi-sensor fusion, and relates to the technical field of livestock breeding. Comprising the following steps: firstly, collecting boundary information of a livestock place, determining a motion range and establishing a plane-coordinate system; motion data, including motion coordinates, acceleration and timestamps, is obtained by a motion monitoring unit carried by the livestock. And converting the acceleration into displacement by using a double-integral method, aligning a motion track by using a dynamic time migration algorithm, and outputting fusion data. And based on the motion fusion data, setting a monitoring time window, and extracting motion features and physiological parameters. And generating a comprehensive health judgment coefficient in a weighted coupling mode, and dynamically correcting a health judgment threshold value according to the environment temperature to obtain a self-adaptive health judgment threshold value. And comparing the comprehensive health judgment coefficient with a self-adaptive health judgment threshold value to obtain a health monitoring result. And comprehensive and accurate monitoring and evaluation of the health state of the livestock are realized.

Description

Technical Field

[0001] The present invention relates to the field of animal husbandry technology, and in particular to a animal husbandry monitoring method and device based on multi-sensor fusion. Background Art

[0002] In modern animal husbandry, animal health monitoring and management are critical for ensuring the quality of livestock products and improving farming efficiency. Traditional animal husbandry management relies primarily on manual observation and empirical judgment. This approach is not only inefficient but also susceptible to human influence, resulting in inaccurate and incomplete monitoring of animal health. With advances in science and technology, particularly the development of sensor technology and big data analytics, an increasing number of intelligent monitoring systems are being introduced into the livestock industry. However, existing technologies often focus on a single aspect, such as monitoring a single physiological parameter or environmental monitoring. This often results in incomplete monitoring results, making it difficult to fully reflect the health of the animal.

[0003] Existing technologies also have some shortcomings in motion monitoring. While some studies have investigated animal movement monitoring, most employ static or simplified monitoring methods, failing to fully capture the dynamics of an animal's movements within its range of motion. Furthermore, existing motion monitoring systems often lack sophisticated processing of data timing, resulting in low accuracy of motion data and insufficient correlation with physiological data, further impacting the accuracy of health assessments. Therefore, effectively integrating multiple data types to achieve a comprehensive assessment of animal health remains a pressing technical challenge.

[0004] In the prior art, publication number CN119523438A discloses a method, apparatus, device, and storage medium for monitoring livestock health status. The specific monitoring method includes: obtaining sensor monitoring data within a livestock farm; constructing a sample data set based on the sensor monitoring data; using the sensor sample data to perform model pre-training and model tuning training on a pre-constructed monitoring network model until the preset model training end conditions are met, and using the trained monitoring network model as the livestock health monitoring model. However, this solution only monitors through the model and lacks real-time dynamic monitoring of motion data, which may result in insufficient timeliness and accuracy of the data. At the same time, the solution mainly focuses on the construction of the data model and lacks effective integration of different data types, such as motion and physiological data. This may result in the model being effective only under certain specific conditions and unable to fully reflect the health status of the livestock. Therefore, the real-time and effectiveness of the monitoring system are reduced.

[0005] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not form the prior art that is already known to a person of ordinary skill in the art. Summary of the Invention

[0006] The purpose of the present invention is to provide a livestock monitoring method and device based on multi-sensor fusion to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A livestock monitoring method based on multi-sensor fusion, the specific steps include:

[0009] Collecting boundary information of the livestock farm to be detected, determining the movement range of the livestock farm to be detected based on the boundary information, establishing a plane coordinate system within the movement range, and obtaining movement data of the corresponding livestock within the movement range during the monitoring period through the movement monitoring unit carried by the livestock in the livestock farm to be detected. The movement data includes movement coordinates, movement acceleration and corresponding timestamp tags;

[0010] Based on the obtained motion data, the double integral of the acceleration data in the motion data is converted into motion displacement. Through the dynamic time migration algorithm involving penalty terms, the acceleration data is overall translated in time series, the motion coordinate positioning trajectory is aligned, and spatially consistent motion fusion data is output;

[0011] Based on the motion fusion time series data within the monitoring period, a monitoring time window is set, motion feature data of each monitoring time window is extracted, and physiological characteristic parameters of the corresponding livestock within the monitoring time window are collected, wherein the physiological characteristic parameters include average body temperature and average heart rate;

[0012] Characterize the behavioral health index based on the movement characteristic data, generate the physiological health index of the corresponding livestock based on the physiological characteristic parameters, and generate the comprehensive health judgment coefficient through weighted coupling based on the obtained behavioral health index and physiological health index;

[0013] Set a health judgment threshold and collect the ambient temperature around the livestock farm to be tested. Dynamically modify the health judgment threshold based on the collected ambient temperature to obtain an adaptive health judgment threshold. Compare the comprehensive health judgment coefficient with the adaptive health judgment threshold. Based on the comparison results, issue the corresponding health monitoring results.

[0014] Furthermore, the movement range of the livestock farm to be detected is determined based on the boundary information, and a plane coordinate system is established within the movement range, wherein the specific method of establishing the plane coordinate system is: using positioning technology to collect coordinate points at the boundary of the livestock farm, specifically including boundary corners, boundary turning points and evenly spaced points on the boundary, recording the latitude and longitude information of each point, and converting it into a unified measurement unit, removing outliers, and connecting the remaining boundary coordinate points to form a closed polygon, representing the boundary of the livestock farm to be detected, and the area enclosed by the closed polygon is the movement range, and the coordinates of a corner randomly selected within the boundary of the movement range are used as the origin, and the plane coordinate system is established with the east direction as the positive direction of the X axis and the north direction as the positive direction of the Y axis;

[0015] The motion monitoring unit carried by the livestock in the livestock farm to be tested, wherein the motion monitoring unit specifically includes a position information sensor and a motion data sensor, which is used to monitor the motion coordinates and motion acceleration of the corresponding livestock location, and timestamp all data to obtain the motion data of the corresponding livestock within the motion range during the monitoring period, wherein a cyclic monitoring method is adopted, specifically the monitoring period and the interval time together constitute a cycle, and the livestock are cyclically monitored.

[0016] Furthermore, based on the obtained motion data, the acceleration data in the motion data is double-integrated and converted into motion displacement, wherein the motion displacement is calculated based on the formula:

[0017] Q t =∫∫IMU t+z dt 2

[0018] Where Q t IMU is the movement displacement data of livestock during the monitoring period. t+z is the acceleration of the livestock at the t+z moment, t0 is the initial moment of a monitoring period, t k is the end time of a monitoring period, and t is the time variable within a monitoring period;

[0019] The acceleration data is shifted in time series by a dynamic time migration algorithm involving a penalty term. Specifically, the optimal time delay is calculated by the dynamic time migration algorithm. The formula for calculating the optimal time delay is:

[0020]

[0021] Where z is the optimal time delay, Indicates the objective function minimization, UWB t is the motion coordinate of the livestock at time t, is the position coordinate of the livestock at time t determined by the motion displacement data, λ is the penalty coefficient, Represents UWB at time t t and The Euclidean distance between the generated displacements.

[0022] Furthermore, the acceleration data is shifted as a whole in time series to align the motion coordinate positioning trajectory, wherein the specific method steps for performing the overall translation are: updating the timestamp of each motion acceleration to correspond to the timestamp after the optimal time delay z, wherein if the optimal time delay z is a positive value, it indicates that the motion acceleration data is shifted toward the back of the time gradient; if the optimal time delay z is a negative value, it indicates that the motion acceleration data is shifted toward the front of the time gradient;

[0023] Integrate the motion acceleration data obtained through translation and calculation with the motion displacement data, and output them as a unified data set, remembering motion fusion data.

[0024] Furthermore, based on the motion fusion time series data within the monitoring period, a monitoring time window is set, wherein the starting time of the monitoring time window is marked as L0, and the starting time of the monitoring time window is marked as L k , while the duration of a single monitoring time window is less than the duration of the monitoring period;

[0025] Extracting motion feature data of each monitoring time window, wherein the motion feature data specifically includes average motion displacement, maximum acceleration, and average motion speed within all monitoring time windows;

[0026] The behavioral health index is characterized based on the motion characteristic data. The specific calculation formula of the behavioral health index is as follows:

[0027]

[0028] Where, SCI is behavioral health index, wy mean is the mean motion displacement, wy ref is the reference motion displacement, V mean is the average speed, V ref is the reference motion speed, is the maximum acceleration in all monitoring time windows, and the maximum acceleration is in the i-th monitoring time window, is the maximum acceleration in the i-th monitoring time window, where i is the index of the monitoring time window.

[0029] Furthermore, a physiological health index of the corresponding livestock is generated based on the physiological characteristic parameters, wherein the specific formula for calculating the physiological health index is:

[0030]

[0031] In the formula, ECI is the physiological health index, and T mean is the average body temperature, XL mean is the average heart rate, T ref and XL ref are the reference body temperature and the reference heart rate respectively, where the average body temperature and the average heart rate both refer to the average values of the body temperature and the heart rate within all monitoring time windows.

[0032] Furthermore, a comprehensive health judgment coefficient is generated by a weighted coupling method, and the specific formula based on which the comprehensive health judgment coefficient is calculated is:

[0033]

[0034] In the formula, ZH is the comprehensive health judgment coefficient, ω1 and ω2 are the weight coefficients of the physiological health index and the behavioral health index respectively, where ω1≥ω2 and both ω1 and ω2 are greater than 0;

[0035] Based on the collected environmental temperature, the health judgment threshold is dynamically corrected to obtain an adaptive health judgment threshold, and the specific formula based on which the adaptive health judgment threshold is calculated is:

[0036]

[0037] In the formula, yz is the adaptive health judgment threshold, yz0 is the health judgment threshold, HT s is the environmental temperature, specifically referring to the average environmental temperature within the movement range, SJ is the development days of the monitored livestock, β is the scaling coefficient of the development days of the monitored livestock, and HT0 is the reference environmental temperature.

[0038] Furthermore, the comprehensive health judgment coefficient is compared with the adaptive health judgment threshold, and the specific logic for issuing the corresponding effect evaluation according to the comparison result is:

[0039] When the comprehensive health judgment coefficient ZH of the monitored livestock ≥ yz, it is judged that the health state of the monitored livestock is abnormal, and a warning signal is issued to further determine its health state;

[0040] When the comprehensive health judgment coefficient ZH of the monitored livestock < yz, it is judged that the health state of the monitored livestock is normal.

[0041] The present invention also provides a livestock monitoring device based on multi-sensor fusion. The livestock monitoring device based on multi-sensor fusion is used to execute the above-mentioned livestock monitoring method based on multi-sensor fusion, and includes:

[0042] A motion range planning module is used to collect boundary information of the livestock farm to be detected, determine the motion range of the livestock farm to be detected based on the boundary information, establish a plane coordinate system within the motion range, and obtain the motion data of the corresponding livestock within the motion range during the monitoring period through the motion monitoring units carried by the livestock in the livestock farm to be detected. The motion data includes motion coordinates, motion acceleration and corresponding timestamp tags;

[0043] The motion information correction and alignment module is used to convert the double integral of the acceleration data in the motion data into motion displacement based on the obtained motion data. Through the dynamic time migration algorithm involving penalty terms, the acceleration data is overall translated in time sequence to align the motion coordinate positioning trajectory and output spatially consistent motion fusion data;

[0044] A feature data acquisition module is used to set monitoring time windows based on the motion fusion time series data within the monitoring period, extract motion feature data in each monitoring time window, and collect physiological characteristic parameters of the corresponding livestock within the monitoring time window, wherein the physiological characteristic parameters include average body temperature and average heart rate;

[0045] A comprehensive coupling judgment analysis module is used to characterize the behavioral health index based on the motion characteristic data, generate the physiological health index of the corresponding livestock based on the physiological characteristic parameters, and generate a comprehensive health judgment coefficient through a weighted coupling method based on the obtained behavioral health index and physiological health index;

[0046] The dynamic threshold comparison module is used to set the health judgment threshold and collect the ambient temperature of the livestock farm to be tested. The health judgment threshold is dynamically corrected based on the collected ambient temperature to obtain an adaptive health judgment threshold. The comprehensive health judgment coefficient is compared with the adaptive health judgment threshold. According to the comparison result, the corresponding health monitoring result is issued.

[0047] Compared with the prior art, the present invention has the following beneficial effects:

[0048] First, a multi-sensor fusion-based monitoring method can acquire livestock movement data and physiological characteristics in real time, forming a comprehensive health assessment system. Double integration is used to convert acceleration data into movement displacement, and combined with a dynamic time offset algorithm, the movement data is time-aligned to ensure data consistency and reliability. This high-precision data processing capability effectively reduces health assessment bias caused by data errors, thereby improving the accuracy of health monitoring.

[0049] Secondly, by extracting movement characteristic data and physiological characteristic parameters, behavioral health index and physiological health index are established. This comprehensive evaluation mechanism makes the monitoring results more comprehensive and can better reflect the actual health status of livestock.

[0050] Furthermore, setting health assessment thresholds and dynamically adjusting them based on ambient temperature makes health assessment more flexible and adaptable to the needs of farming under varying climate conditions. This adaptive mechanism effectively addresses the impact of the external environment on animal health, further improving the accuracy and applicability of health monitoring. This allows for timely adjustments to feeding and management strategies, thereby reducing disease incidence and improving farming efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 Schematic diagram of the overall method flow of the present invention;

[0052] Figure 2 Comparative point and line graphs for judging livestock health status;

[0053] Figure 3 This is the fitting curve of physiological health index-comprehensive health judgment coefficient;

[0054] Figure 4 Calculate statistical charts for comprehensive health judgment coefficients;

[0055] Figure 5 It is a dot-line graph comparing the comprehensive health judgment coefficient and the adaptive health judgment threshold;

[0056] Figure 6 It is a schematic diagram of the overall structure of the device of the present invention. DETAILED DESCRIPTION

[0057] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to specific embodiments.

[0058] It should be noted that, unless otherwise defined, the technical or scientific terms used in the present invention should have the usual meanings understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative position relationships. When the absolute position of the object being described changes, the relative position relationship may also change accordingly.

[0059] Example:

[0060] See also Figure 1-Figure 5 , the present invention provides a technical solution:

[0061] A livestock monitoring method based on multi-sensor fusion, the specific steps include:

[0062] Step 1: Collect the boundary information of the livestock farm to be detected, determine the movement range of the livestock farm to be detected based on the boundary information, establish a plane coordinate system within the movement range, and obtain the movement data of the corresponding livestock within the movement range during the monitoring period through the motion monitoring unit carried by the livestock in the livestock farm to be detected. The motion data includes motion coordinates, motion acceleration and corresponding timestamp tags.

[0063] Based on the boundary information, the movement range of the livestock farm to be inspected is determined, and a plane coordinate system is established within the movement range. The specific method for establishing the plane coordinate system is: using positioning technology, such as GPS, DGPS positioning technology, etc., to collect coordinate points at the boundary of the livestock farm, specifically including boundary corners, boundary turning points and evenly spaced points on the boundary, record the latitude and longitude information of each point, and convert it into a unified measurement unit, remove outliers, wherein the removal of outliers can be specifically screened by the Z-Score method or the IQR method, and connect the remaining boundary coordinate points to form a closed polygon, representing the boundary of the livestock farm to be inspected. The area enclosed by the closed polygon is the movement range, and the coordinates of a corner randomly selected within the boundary of the movement range are used as the origin, and the plane coordinate system is established with the east direction as the positive direction of the X axis and the north direction as the positive direction of the Y axis.

[0064] The motion monitoring unit carried by the livestock in the livestock farm to be tested, wherein the motion monitoring unit specifically includes a position information sensor and a motion data sensor, which is used to monitor the motion coordinates and motion acceleration of the corresponding livestock location, and timestamp all data to obtain the motion data of the corresponding livestock within the motion range during the monitoring period, wherein a cyclic monitoring method is adopted, specifically the monitoring period and the interval time together constitute a cycle, and the livestock are cyclically monitored.

[0065] The location information sensor can use an ultra-wideband (UWB) positioning sensor. UWB technology enables precise positioning indoors or in environments with many obstacles. Its high temporal resolution and anti-interference capabilities make it superior in complex environments. Alternatively, a Global Positioning System (GPS) sensor can be used. GPS sensors can provide real-time location information for livestock and are suitable for large-scale livestock farms. Satellite signals can be used to obtain the latitude and longitude coordinates of livestock, making it suitable for large outdoor environments.

[0066] Motion data sensors are specifically accelerometers, which can monitor livestock acceleration changes in real time, helping to calculate motion status and dynamic behavior. A three-axis accelerometer can be used to obtain motion data in both horizontal and vertical directions.

[0067] Step 2: Based on the obtained motion data, the double integral of the acceleration data in the motion data is converted into motion displacement. Through the dynamic time migration algorithm involving penalty terms, the acceleration data is overall translated in time series, the motion coordinate positioning trajectory is aligned, and spatially consistent motion fusion data is output.

[0068] Based on the obtained motion data, the acceleration data in the motion data is double-integrated and converted into motion displacement. The specific formula for motion displacement calculation is:

[0069] Q t =∫∫IMU t+z dt 2

[0070] Where Q t IMU is the movement displacement data of livestock during the monitoring period. t+z is the acceleration of the livestock at the t+z moment, t0 is the initial moment of a monitoring period, t k is the end time of a monitoring period, and t is the time variable within a monitoring period.

[0071] It's important to note that acceleration is the most direct measurement parameter in motion monitoring. Converting acceleration to displacement through double integration allows the displacement of livestock to be derived from the detected acceleration data. Integrating the time variable t in the formula accurately reflects the movement over a specific time period, especially when acceleration may vary at different moments. This method adapts to monitoring needs in dynamic environments, converting acceleration into displacement through double integration by leveraging the relationship between acceleration, velocity, and displacement in physics.

[0072] The acceleration data is shifted in time series by a dynamic time migration algorithm involving a penalty term. Specifically, the optimal time delay is calculated by the dynamic time migration algorithm. The formula for calculating the optimal time delay is:

[0073]

[0074] Where z is the optimal time delay, Indicates the objective function minimization, UWB t is the motion coordinate of the livestock at time t, is the position coordinate of the livestock at time t determined by the motion displacement data, λ is the penalty coefficient, Represents UWB at time t t and The Euclidean distance between the generated displacements.

[0075] It should be noted that the purpose of this formula is to minimize the value of the objective function by adjusting the value of z. Specifically, the objective function consists of two parts: the distance between the actual position and the estimated position, which reflects the accuracy of the prediction, and the penalty for time delay, which controls the delay size.

[0076] In dynamic time series, due to measurement errors, inaccuracies in the motion model, and other factors, there may be temporal deviations between the actual measured motion trajectory and the trajectory derived based on acceleration. By calculating the optimal time delay, the two can be effectively aligned, improving the matching accuracy.

[0077] The penalty term λ*|z| is set to address overfitting and prevent the selection of unreasonable time delays. By introducing a penalty mechanism, the model can be encouraged to choose a reasonable time delay, achieving a balance between accuracy and complexity.

[0078] The penalty coefficient λ can usually be selected with an initial value through empirical rules, and then adjusted according to actual conditions. For specific application scenarios, the distribution characteristics of acceleration data and displacement data are analyzed, and a λ value that can reflect these characteristics is selected to ensure that the penalty term moderately affects the results when adjusting the time delay. The penalty coefficient λ is generally between 0.1 and 0.5.

[0079] Step 3: Based on the motion fusion time series data within the monitoring period, set the monitoring time window, extract the motion feature data of each monitoring time window, and collect the physiological characteristic parameters of the corresponding livestock within the monitoring time window, wherein the physiological characteristic parameters include average body temperature and average heart rate.

[0080] Based on the motion fusion time series data within the monitoring period, a monitoring time window is set, wherein the starting moment of the monitoring time window is marked as L0 and the starting moment of the monitoring time window is marked as L k , while the duration of a single monitoring time window is less than the duration of the monitoring period;

[0081] Extracting motion feature data of each monitoring time window, wherein the motion feature data specifically includes average motion displacement, maximum acceleration, and average motion speed within all monitoring time windows;

[0082] Physiological parameters of the corresponding livestock are collected within the monitoring time window, including average body temperature and average heart rate. The physiological parameters are collected by measuring body temperature using an infrared sensor or skin surface sensor, which is attached to the animal's ear, armpit, or other suitable area. A heart rate monitoring belt is attached around the animal's chest, ensuring good contact between the sensor and the skin. The monitoring device is then activated to begin real-time heart rate monitoring.

[0083] Step 4: Characterize the behavioral health index based on the movement characteristic data, generate the physiological health index of the corresponding livestock based on the physiological characteristic parameters, and generate a comprehensive health judgment coefficient through weighted coupling based on the obtained behavioral health index and physiological health index;

[0084] The behavioral health index is characterized based on the motion characteristic data. The specific calculation formula of the behavioral health index is as follows:

[0085]

[0086] Where, SCI is behavioral health index, wy mean is the mean motion displacement, wy ref is the reference motion displacement, V mean is the average speed, V ref is the reference motion speed, is the maximum acceleration in all monitoring time windows, and the maximum acceleration is in the i-th monitoring time window, is the maximum acceleration in the i-th monitoring time window, where i is the index of the monitoring time window.

[0087] It should be noted that the behavioral health index (SCI) is used to characterize health status through movement behavior. The larger the value of the behavioral health index (SCI), the greater the difference between the livestock's movement behavior and its movement behavior in a healthy state, and therefore the greater the probability of sub-health.

[0088] Among them, wy mean Reflects the overall movement of livestock during the monitoring period. mean -wy ref The introduction of | provides a measure of the deviation between actual movement and reference movement for the behavioral health index by calculating the relative difference The degree of variability in livestock movement can be quantified. When the average movement displacement of livestock is below the reference value, it indicates insufficient activity levels. This can be due to a variety of reasons, such as illness, injury, environmental factors such as unsuitable breeding conditions, or psychological factors such as stress or anxiety. When the average movement displacement of livestock is above the reference value, it indicates excessive activity levels. This may be caused by behaviors such as compulsive movement or chasing due to shock or sudden injury. Therefore, both insufficient and excessive exercise can lead to poor animal health. Healthy livestock should maintain an appropriate range of movement to meet their physiological and psychological needs.

[0089] Speed ​​changes are closely related to health status. Healthy livestock should have normal movement speed. Speed ​​deviation will affect the health index through the natural logarithm function. Larger speed differences have a more significant impact on the index. Use ln[1+|Vmean -V ref |] is used to smooth out the effects of speed, preventing excessive speed deviations from causing nonlinear fluctuations in the health index. The natural logarithm function effectively controls variations over a wide range, ensuring that smaller speed changes have a relatively large impact on the health index, while larger speed changes have a smaller impact. High movement speeds may indicate that the animal is stressed, anxious, or threatened. In these cases, the animal may be in a "fight or flight" state, exhibiting abnormal activity. Low movement speeds indicate insufficient activity, possibly due to illness, pain, or other factors.

[0090] High acceleration generally indicates active activity. A high maximum acceleration means the animal is active, while low acceleration may indicate poor health or insufficient exercise. Health status is assessed by averaging the maximum acceleration with the acceleration of the next time window, avoiding potential misjudgments caused by a single window's maximum acceleration. By averaging the maximum acceleration of the current and next windows, the animal's activity trends over time are better captured. This dynamic assessment more accurately reflects the animal's behavioral habits, rather than relying solely on momentary high values.

[0091] The reference motion speed and reference motion displacement can be set specifically according to expert experience.

[0092] The physiological health index of the corresponding livestock is generated based on the physiological characteristic parameters, wherein the specific formula for calculating the physiological health index is:

[0093]

[0094] Where, ECI is the physiological health index, T mean For average body temperature, XL mean is the average heart rate, T ref and XL ref They are reference body temperature and reference heart rate respectively, wherein the average body temperature and average heart rate refer to the average body temperature and heart rate in all monitoring time windows.

[0095] It should be noted that the physiological health index (ECI) indicates the degree to which livestock deviate from a healthy state through the difference between monitored physiological parameters and healthy physiological parameters. The larger the ECI value, the worse the livestock's condition.

[0096] Body temperature is an important indicator of an animal's metabolic activity. A body temperature within the normal range is usually associated with a healthy state. A high body temperature may indicate infection or other health problems, while a low body temperature may indicate a failure in thermoregulation. mean Close to T refWhen the heart rate is too high or too low, the physiological health index is low; on the contrary, a larger deviation will lead to a higher ECI, indicating an increased health risk. Heart rate is directly related to the animal's activity level and health status. Too high or too low heart rate usually indicates stress, disease or other physiological abnormalities. Similar to body temperature, XL mean and XL ref The smaller the difference, the lower the ECI, indicating better health; the opposite indicates potential health issues. The formula uses absolute values ​​to calculate the deviation between average body temperature and average heart rate and the reference value. The purpose of using absolute values ​​is to eliminate the influence of positive and negative signs, so that whether it is above or below the reference value, the magnitude of the deviation is reflected. Using the square root function can standardize the results. The square root also helps to reduce the impact of large deviations on the health index and maintain the sensitivity of the index.

[0097] The reference body temperature and reference heart rate can be set by referring to relevant information and combining expert experience.

[0098] The comprehensive health judgment coefficient is generated through weighted coupling, and the specific formula for calculating the comprehensive health judgment coefficient is:

[0099]

[0100] In the formula, ZH is the comprehensive health judgment coefficient, ω1 and ω2 are the weight coefficients of the physiological health index and behavioral health index, respectively, where ω1 ≥ ω2 and ω1 and ω2 are both greater than 0;

[0101] Since the correspondence between the physical health index and behavioral health index and health status has been specifically explained above, it will not be repeated here.

[0102] The purpose of using the exponential function is to emphasize that the impact of physiological health is nonlinear. A higher ECI will significantly reduce ZH, reflecting the strong impact of physiological problems on overall health.

[0103] Physical health is fundamental to animal survival and production. Threats to physical health, such as abnormal body temperature or heart rate, can directly impact the animal's survival. Therefore, physical health is generally considered a more important factor, and behavioral health is often considered a complement to physical health. When an animal is in good physical condition, positive behavioral performance contributes to overall health. Behavioral health reflects an animal's adaptability to the environment and its psychological state, but changes in behavioral health often occur after changes in physical health. Therefore, ω1 ≥ ω2, with both ω1 and ω2 greater than 0.

[0104] Step 5: Set the health judgment threshold and collect the ambient temperature around the livestock farm to be tested. Dynamically modify the health judgment threshold based on the collected ambient temperature to obtain an adaptive health judgment threshold. Compare the comprehensive health judgment coefficient with the adaptive health judgment threshold. Based on the comparison result, issue the corresponding health monitoring result.

[0105] The health judgment threshold is dynamically modified based on the collected ambient temperature to obtain an adaptive health judgment threshold. The specific calculation formula of the adaptive health judgment threshold is:

[0106]

[0107] Where yz is the adaptive health judgment threshold, yz0 is the health judgment threshold, HT s is the ambient temperature, specifically the average ambient temperature within the range of motion, SJ is the number of days of development of the monitored livestock, β is the scaling factor of the number of days of development of the monitored livestock, and HT0 is the reference ambient temperature.

[0108] It should be noted that hot or cold weather will affect the movement behavior of livestock, so the health judgment threshold is dynamically adjusted based on the difference between the ambient temperature and the reference ambient temperature.

[0109] At the same time, young livestock generally show higher activity levels and like to play and explore, especially in the early stages of growth. Their movement patterns are usually characterized by jumping, running, and quick turns. Adult livestock have relatively low activity levels and are usually focused on maintaining daily activities such as foraging and resting. Adult livestock's movements are more steady and slow. The heart rate of young livestock is usually higher than that of adult livestock, depending on the age, health status, and activity level of the young livestock. The heart rate of adult livestock is generally lower. With age, heart development and weight gain cause the resting heart rate of adult cattle to decrease. Therefore, the adaptive health judgment threshold is proportional to the number of days (SJ) of development of the monitored livestock.

[0110] The comprehensive health judgment coefficient is compared with the adaptive health judgment threshold. The specific logic for issuing the corresponding effect evaluation based on the comparison result is as follows:

[0111] When the comprehensive health judgment coefficient ZH of the monitored livestock is ≤ yz, the health status of the monitored livestock is judged to be abnormal, and an early warning signal is issued to further determine its health status;

[0112] When the comprehensive health judgment coefficient ZH of the monitored livestock is greater than yz, the health status of the monitored livestock is judged to be normal. Table 1 shows the statistical data of some health judgment parameters.

[0113]

[0114]

[0115] See also Figure 6 The present invention also provides a livestock monitoring device based on multi-sensor fusion, which is used to execute the above-mentioned livestock monitoring method based on multi-sensor fusion, including:

[0116] A motion range planning module is used to collect boundary information of the livestock farm to be detected, determine the motion range of the livestock farm to be detected based on the boundary information, establish a plane coordinate system within the motion range, and obtain the motion data of the corresponding livestock within the motion range during the monitoring period through the motion monitoring units carried by the livestock in the livestock farm to be detected. The motion data includes motion coordinates, motion acceleration and corresponding timestamp tags;

[0117] The motion information correction and alignment module is used to convert the double integral of the acceleration data in the motion data into motion displacement based on the obtained motion data. Through the dynamic time migration algorithm involving penalty terms, the acceleration data is overall translated in time sequence to align the motion coordinate positioning trajectory and output spatially consistent motion fusion data;

[0118] A feature data acquisition module is used to set monitoring time windows based on the motion fusion time series data within the monitoring period, extract motion feature data in each monitoring time window, and collect physiological characteristic parameters of the corresponding livestock within the monitoring time window, wherein the physiological characteristic parameters include average body temperature and average heart rate;

[0119] A comprehensive coupling judgment analysis module is used to characterize the behavioral health index based on the motion characteristic data, generate the physiological health index of the corresponding livestock based on the physiological characteristic parameters, and generate a comprehensive health judgment coefficient through a weighted coupling method based on the obtained behavioral health index and physiological health index;

[0120] The dynamic threshold comparison module is used to set the health judgment threshold and collect the ambient temperature of the livestock farm to be tested. The health judgment threshold is dynamically corrected based on the collected ambient temperature to obtain an adaptive health judgment threshold. The comprehensive health judgment coefficient is compared with the adaptive health judgment threshold. According to the comparison result, the corresponding health monitoring result is issued.

[0121] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0122] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software depends on the specific application and design constraints of the technical solution.

[0123] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, and may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment as needed.

[0124] The above is only a specific implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be covered by the scope of protection of the present application.

Claims

1. A livestock monitoring method based on multi-sensor fusion, characterized in that: The specific steps include: Collecting boundary information of the livestock farm to be detected, determining the movement range of the livestock farm to be detected based on the boundary information, establishing a plane coordinate system within the movement range, and obtaining movement data of the corresponding livestock within the movement range during the monitoring period through the movement monitoring unit carried by the livestock in the livestock farm to be detected. The movement data includes movement coordinates, movement acceleration and corresponding timestamp tags; Based on the obtained motion data, the double integral of the acceleration data in the motion data is converted into motion displacement. Through the dynamic time migration algorithm involving penalty terms, the acceleration data is overall translated in time series, the motion coordinate positioning trajectory is aligned, and spatially consistent motion fusion data is output; Based on the motion fusion time series data within the monitoring period, a monitoring time window is set, motion feature data of each monitoring time window is extracted, and physiological characteristic parameters of the corresponding livestock within the monitoring time window are collected, wherein the physiological characteristic parameters include average body temperature and average heart rate; Characterize the behavioral health index based on the movement characteristic data, generate the physiological health index of the corresponding livestock based on the physiological characteristic parameters, and generate the comprehensive health judgment coefficient through weighted coupling based on the obtained behavioral health index and physiological health index; Set a health judgment threshold and collect the ambient temperature around the livestock farm to be tested. Dynamically modify the health judgment threshold based on the collected ambient temperature to obtain an adaptive health judgment threshold. Compare the comprehensive health judgment coefficient with the adaptive health judgment threshold. Based on the comparison results, issue the corresponding health monitoring results.

2. The livestock monitoring method based on multi-sensor fusion according to claim 1, characterized in that: The motion range of the livestock farm to be inspected is determined based on the boundary information, and a plane coordinate system is established within the motion range. The specific method of establishing the plane coordinate system is as follows: using positioning technology to collect coordinate points at the boundary of the livestock farm, specifically including boundary corners, boundary turning points, and evenly spaced points on the boundary, recording the latitude and longitude information of each point, converting it into a unified measurement unit, removing outliers, and connecting the remaining boundary coordinate points to form a closed polygon representing the boundary of the livestock farm to be inspected. The area enclosed by the closed polygon is the motion range. The coordinates of a corner randomly selected within the boundary of the motion range are used as the origin, and the plane coordinate system is established with the east direction as the positive direction of the X axis and the north direction as the positive direction of the Y axis; The motion monitoring unit carried by the livestock in the livestock farm to be tested, wherein the motion monitoring unit specifically includes a position information sensor and a motion data sensor, which is used to monitor the motion coordinates and motion acceleration of the corresponding livestock location, and timestamp all data to obtain the motion data of the corresponding livestock within the motion range during the monitoring period, wherein a cyclic monitoring method is adopted, specifically the monitoring period and the interval time together constitute a cycle, and the livestock are cyclically monitored.

3. The livestock monitoring method based on multi-sensor fusion according to claim 1, characterized in that: Based on the obtained motion data, the acceleration data in the motion data is double-integrated and converted into motion displacement. The specific formula for motion displacement calculation is: Q t =∫∫IMU t+z dt 2 Where Q t IMU is the movement displacement data of livestock during the monitoring period. t+z is the acceleration of the livestock at the t+z moment, t0 is the initial moment of a monitoring period, t k is the end time of a monitoring period, and t is the time variable within a monitoring period; The acceleration data is shifted in time series by a dynamic time migration algorithm involving a penalty term. Specifically, the optimal time delay is calculated by the dynamic time migration algorithm. The formula for calculating the optimal time delay is: Where z is the optimal time delay, Indicates the objective function minimization, UWB t is the motion coordinate of the livestock at time t, is the position coordinate of the livestock at time t determined by the motion displacement data, λ is the penalty coefficient, Represents UWB at time t t and The Euclidean distance between the generated displacements.

4. The livestock monitoring method based on multi-sensor fusion according to claim 3, characterized in that: Perform an overall translation of the acceleration data in time series to align the motion coordinate positioning trajectory. The specific method and steps for performing the overall translation are as follows: update the timestamp of each motion acceleration to correspond to the timestamp after the optimal time delay z. If the optimal time delay z is a positive value, it means that the motion acceleration data is translated backward in time gradient; if the optimal time delay z is a negative value, it means that it is translated forward in time gradient; Integrate the motion acceleration data obtained through translation and calculation with the motion displacement data, and output them as a unified data set, remembering motion fusion data.

5. The livestock monitoring method based on multi-sensor fusion according to claim 4, characterized in that: Based on the motion fusion time series data within the monitoring period, a monitoring time window is set, wherein the starting moment of the monitoring time window is marked as L0 and the starting moment of the monitoring time window is marked as L k , while the duration of a single monitoring time window is less than the duration of the monitoring period; Extracting motion feature data of each monitoring time window, wherein the motion feature data specifically includes average motion displacement, maximum acceleration, and average motion speed within all monitoring time windows; The behavioral health index is characterized based on the motion characteristic data. The specific calculation formula of the behavioral health index is as follows: Where, SCI is behavioral health index, wy mean is the mean motion displacement, wy ref is the reference motion displacement, V mean is the average speed, V ref is the reference motion speed, is the maximum acceleration in all monitoring time windows, and the maximum acceleration is in the i-th monitoring time window, is the maximum acceleration in the i-th monitoring time window, where i is the index of the monitoring time window.

6. The livestock monitoring method based on multi-sensor fusion according to claim 5, characterized in that: The physiological health index of the corresponding livestock is generated based on the physiological characteristic parameters, wherein the specific formula for calculating the physiological health index is: Where, ECI is the physiological health index, T mean For average body temperature, XL mean is the average heart rate, T ref and XL ref They are reference body temperature and reference heart rate respectively, wherein the average body temperature and average heart rate refer to the average body temperature and heart rate in all monitoring time windows.

7. The livestock monitoring method based on multi-sensor fusion according to claim 6, characterized in that: The comprehensive health judgment coefficient is generated through weighted coupling, and the specific formula for calculating the comprehensive health judgment coefficient is: Where ZH is the comprehensive health judgment coefficient, ω1 and ω2 are the weight coefficients of the physiological health index and behavioral health index, respectively, where ω1 ≥ ω2 and ω1 and ω2 are both greater than 0; The health judgment threshold is dynamically modified based on the collected ambient temperature to obtain an adaptive health judgment threshold. The specific calculation formula of the adaptive health judgment threshold is: Where yz is the adaptive health judgment threshold, yz0 is the health judgment threshold, HT s is the ambient temperature, specifically the average ambient temperature within the range of motion, SJ is the number of days of development of the monitored livestock, β is the scaling factor of the number of days of development of the monitored livestock, and HT0 is the reference ambient temperature.

8. The livestock monitoring method based on multi-sensor fusion according to claim 7, characterized in that: The comprehensive health judgment coefficient is compared with the adaptive health judgment threshold. The specific logic for issuing the corresponding effect evaluation based on the comparison result is as follows: When the comprehensive health judgment coefficient ZH of the monitored livestock is ≤yz, the health status of the monitored livestock is judged to be abnormal, and an early warning signal is issued to further determine its health status; When the comprehensive health judgment coefficient ZH of the monitored livestock is greater than yz, it is judged that the health status of the monitored livestock is normal.

9. A livestock monitoring device based on multi-sensor fusion, characterized by: The livestock monitoring device based on multi-sensor fusion is used to execute the livestock monitoring method based on multi-sensor fusion according to any one of claims 1 to 8, comprising: A motion range planning module is used to collect boundary information of the livestock farm to be detected, determine the motion range of the livestock farm to be detected based on the boundary information, establish a plane coordinate system within the motion range, and obtain the motion data of the corresponding livestock within the motion range during the monitoring period through the motion monitoring units carried by the livestock in the livestock farm to be detected. The motion data includes motion coordinates, motion acceleration and corresponding timestamp tags; The motion information correction and alignment module is used to convert the double integral of the acceleration data in the motion data into motion displacement based on the obtained motion data. Through the dynamic time migration algorithm involving penalty terms, the acceleration data is overall translated in time sequence to align the motion coordinate positioning trajectory and output spatially consistent motion fusion data; A feature data acquisition module is used to set monitoring time windows based on the motion fusion time series data within the monitoring period, extract motion feature data in each monitoring time window, and collect physiological characteristic parameters of the corresponding livestock within the monitoring time window, wherein the physiological characteristic parameters include average body temperature and average heart rate; A comprehensive coupling judgment analysis module is used to characterize the behavioral health index based on the motion characteristic data, generate the physiological health index of the corresponding livestock based on the physiological characteristic parameters, and generate a comprehensive health judgment coefficient through a weighted coupling method based on the obtained behavioral health index and physiological health index; The dynamic threshold comparison module is used to set the health judgment threshold and collect the ambient temperature of the livestock farm to be tested. The health judgment threshold is dynamically corrected based on the collected ambient temperature to obtain an adaptive health judgment threshold. The comprehensive health judgment coefficient is compared with the adaptive health judgment threshold. According to the comparison result, the corresponding health monitoring result is issued.

Images