A livestock physiological behavior monitoring method and system

Abstract

The present invention relates to the technical field of livestock monitoring, and in particular, to a method and system for monitoring the physiological behavior of livestock. The method includes the following steps: collecting a three-region pressure matrix through a pressure sensing pad; performing dynamic path recognition and camera adjustment based on the three-region pressure matrix to obtain dynamic path tracking data; performing livestock mechanical gait analysis based on the dynamic path tracking data to obtain a gait mechanical feature map; obtaining livestock basic information; performing individual gait time series segmentation based on the gait mechanical feature map to obtain segmented and aligned gait data; calculating an adaptive gait threshold based on the livestock basic information and the segmented and aligned gait data to obtain an individualized abnormal threshold set; and performing abnormal gait pattern recognition based on the individualized abnormal threshold set to obtain a gait deviation index matrix. The present invention significantly improves the scientificity and reliability of livestock management and livestock production efficiency through regional pressure sensing technology and biological tracking camera technology.

Description

Technical Field

[0001] The present invention relates to the technical field of livestock monitoring, and in particular, to a method and system for monitoring livestock physiological behavior. Background Art

[0002] Existing early identification technologies for livestock lameness have significant limitations. The main manifestations are insufficient detection sensitivity, inability to capture minor gait changes in the subclinical stage, resulting in missed best intervention opportunities and the development of minor lameness into serious problems.

[0003] Secondly, the assessment of lameness severity highly depends on the experience and subjective judgment of assessors. The scoring differences between different assessors can reach more than 30%. There is a lack of precise quantification criteria and objective evaluation systems, making it difficult to achieve standardized management and accurate comparison.

[0004] In addition, traditional methods for analyzing the causes of lameness are difficult to distinguish different mechanisms with similar manifestations. For example, lameness caused by laminitis and arthritis is similar in appearance but has different causes and coping solutions. The lack of multi-dimensional parameter analysis leads to low judgment accuracy and insufficient intervention pertinence.

[0005] In summary, the problems of lag in lameness detection, subjectivity in assessment, and ambiguity in cause judgment that generally exist in the existing technologies urgently need to be solved. Summary of the Invention

[0006] Based on this, it is necessary to provide a method and system for monitoring livestock physiological behavior to solve at least one of the above technical problems.

[0007] To achieve the above object, a method for monitoring livestock physiological behavior includes the following steps:

[0008] Step S1: Collect a three-region pressure matrix through a pressure sensing pad; perform dynamic path recognition and camera adjustment based on the three-region pressure matrix to obtain dynamic path tracking data; perform livestock mechanical gait analysis based on the dynamic path tracking data to obtain a gait mechanical feature map.

[0009] Step S2: Obtain livestock basic information; perform individual gait time series segmentation based on the gait mechanical feature map to obtain segmented aligned gait data; calculate an adaptive gait threshold based on the livestock basic information and the segmented aligned gait data to obtain an individualized abnormal threshold set; perform abnormal gait pattern recognition based on the individualized abnormal threshold set to obtain a gait deviation index matrix.

[0010] Step S3: Perform an analysis of the differences in limb weight-bearing ratios based on the gait deviation index matrix and the gait mechanics feature map to obtain limb weight-bearing characteristics; identify the characteristics of the weight transfer pattern based on the limb weight-bearing characteristics; perform pain behavior feature recognition based on the characteristics of the weight transfer pattern to obtain pain feature correlation data; perform limb health assessment based on the pain feature correlation data to obtain the limb functional health index;

[0011] Step S4: Perform time-series monitoring of the behavior pattern based on the limb functional health index to obtain the livestock health trend map.

[0012] The present invention provides comprehensive and accurate basic data that reflects the physical characteristics of livestock gaits by synchronously collecting the ground reaction force distribution and limb joint movement trajectories with high precision during livestock walking, and extracting and integrating preliminary mechanical and kinematic parameters. The fine capture and synchronous processing of this multi-modal data lay a solid foundation for subsequent in-depth analysis of how livestock bear weight, coordinate movements, and the dynamic changes of gaits, significantly improving the accuracy and information content of gait analysis. By constructing and utilizing the historical gait records of individual livestock, personalized gait baselines and adaptive anomaly thresholds are established. This analysis method based on individual historical data can fully consider the physiological differences and historical habits among individuals, enabling the identification of abnormal gaits to no longer rely on the group average level, but precisely target the normal state of the livestock itself. The calculated gait deviation index matrix can quantify the degree of deviation of the current gait of the livestock from its own normal gait, significantly improving the specificity of anomaly detection and the early warning ability. By deeply analyzing key mechanical and kinematic characteristics such as the load distribution, symmetry, support time, and joint mobility of livestock limbs, combined with the identified abnormal patterns, pain-related behavioral manifestations (such as load reduction, avoidance transfer) are accurately identified. This analysis method can infer the pain state (location, intensity, acute or chronic) of livestock from objective gait data and quantify the degree of limb dysfunction. The finally generated limb functional health index comprehensively reflects the load-bearing capacity, movement coordination, and potential pain impact of the limb, providing a direct and valuable basis for evaluating the health of livestock limbs and locating problems. By constructing and monitoring the time series of the livestock limb functional health index and performing correlation analysis with environmental monitoring data and ranch management records, continuous tracking and comprehensive assessment of the health status of livestock are achieved. This method can not only identify current anomalies in the health index, but also calculate the health fluctuation baseline based on historical data to predict future health trends and potential disease risks. By revealing the correlation between health changes and external factors, early warning information and decision-making support are provided for ranch managers, helping to timely adjust feeding and management measures, prevent the occurrence or spread of diseases, and improve the overall health level of the herd. Therefore, the present invention provides a method for monitoring the physiological behavior of livestock. By combining regional pressure sensing technology and biological tracking camera technology, high-precision acquisition of gait mechanical characteristics is achieved; the introduction of individualized baseline comparison and adaptive threshold algorithms greatly improves the early identification ability of abnormal gaits; and through the comprehensive analysis of joint angles, ground reaction forces, and load patterns, an objective and quantitative lameness assessment system is established, effectively solving the core problems of difficult early detection of lameness, subjective severity assessment, and inaccurate etiology analysis in traditional technologies, providing a scientific basis for the health management of the livestock industry.

[0013] Preferably, the present invention also provides a livestock physiological behavior monitoring system for performing the livestock physiological behavior monitoring method as described above. The livestock physiological behavior monitoring system includes:

[0014] The gait data acquisition module is used to collect a three - area pressure matrix through a pressure - sensing pad; perform dynamic path recognition and camera adjustment based on the three - area pressure matrix to obtain dynamic path tracking data; perform livestock mechanical gait analysis based on the dynamic path tracking data to obtain a gait mechanical feature map;

[0015] The abnormal feature recognition module is used to obtain livestock basic information; perform individual gait time - series segmentation on the gait mechanical feature map to obtain segmented and aligned gait data; calculate an adaptive gait threshold based on the livestock basic information and the segmented and aligned gait data to obtain an individualized abnormal threshold set; perform abnormal gait pattern recognition based on the individualized abnormal threshold set to obtain a gait deviation index matrix;

[0016] The pain status assessment module is used to perform limb load - bearing ratio difference analysis based on the gait deviation index matrix and the gait mechanical feature map to obtain limb load - bearing characteristics; recognize the characteristics of the load - transfer pattern based on the limb load - bearing characteristics; perform pain behavior feature recognition based on the characteristics of the load - transfer pattern to obtain pain feature correlation data; perform limb health assessment based on the pain feature correlation data to obtain a limb function health index;

[0017] The health trend monitoring module is used to perform time - series monitoring of the behavior pattern based on the limb function health index to obtain a livestock health trend map.

[0018] This livestock physiological behavior monitoring system, by integrating four core modules of gait data acquisition, abnormal feature recognition, pain status assessment, and health trend monitoring, realizes comprehensive, objective, continuous monitoring and in - depth analysis of livestock physiological behavior. This system can accurately capture the mechanical and kinematic details of livestock gaits, establish an individualized health baseline, accurately identify and quantify subtle gait abnormalities, and then objectively evaluate the limb function status and pain degree, and predict the health trend in combination with environmental and management information. This systematic monitoring and analysis ability significantly improves the ability to detect potential health problems early, prevent and intervene in advance, and provides scientific and reliable technical support for livestock management and livestock production efficiency. Brief Description of the Drawings

[0019] Figure 1 It is a schematic diagram of the step - by - step process of a livestock physiological behavior monitoring method.

[0020] The realization of the purpose, functional characteristics, and advantages of the present invention will be further described with reference to the embodiments and the accompanying drawings. Detailed Embodiments

[0021] The technical method of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those skilled in the art within the scope of the present invention without creative work belong to the protection scope of the present invention.

[0022] In the embodiments of the present invention, reference is made to Figure 1 As shown, it is a schematic flow chart of the steps of the livestock physiological behavior monitoring method of the present invention. In this example, the livestock physiological behavior monitoring method includes the following steps:

[0023] Step S1: Collect a three-region pressure matrix through a pressure sensing pad; perform dynamic path recognition and camera adjustment based on the three-region pressure matrix to obtain dynamic path tracking data; perform livestock mechanical gait analysis based on the dynamic path tracking data to obtain a gait mechanical feature map;

[0024] In the embodiments of the present invention, focus is on collecting pressure and motion data of livestock during walking and performing preliminary mechanical and kinematic analyses. The key lies in using a pressure sensing pad with a three-region division and an independent sensor array to collect a fine pressure matrix; using multiple high-speed cameras to synchronously capture motion images with joint markers; identifying the dynamic path of hoof prints through pressure data and adjusting the camera in real time to maintain the best tracking; extracting the ground reaction force from the synchronously collected pressure data and reconstructing the three-dimensional motion trajectory of joints from the camera data; dividing the gait cycle based on force and motion data; and finally integrating these data to generate a gait mechanical feature map containing mechanical, kinematic, time, and space parameters.

[0025] Step S2: Obtain livestock basic information; perform individual gait time series segmentation based on the gait mechanical feature map to obtain segmented and aligned gait data; calculate an adaptive gait threshold based on the livestock basic information and the segmented and aligned gait data to obtain an individualized abnormal threshold set; perform abnormal gait pattern recognition based on the individualized abnormal threshold set to obtain a gait deviation index matrix;

[0026] In the embodiments of the present invention, emphasis is placed on establishing a personalized reference baseline based on the historical gait data of individual livestock, and on this basis, identifying abnormal patterns in the current gait. The key lies in constructing and utilizing an individual historical gait database, calculating the individual baseline mean and variability of various gait parameters; standardizing and time-aligning the current gait parameters with the individual baseline (e.g., using DTW); dynamically calculating an adaptive abnormal threshold set for the individual livestock according to the livestock basic information, historical variability, and parameter importance; comparing the current gait parameters with the individualized thresholds to detect abnormal points in aspects such as joint angles, ground reaction forces, and lateral movements; analyzing the co-occurrence and temporal correlation of these abnormal points within the gait cycle; and finally generating a quantified gait deviation index matrix to reflect the degree of deviation of various gait parameters from the individual normal state.

[0027] Step S3: Perform limb load-bearing ratio difference analysis based on the gait deviation index matrix and the gait mechanics feature map to obtain limb load-bearing characteristics; identify the load transfer pattern characteristics according to the limb load-bearing characteristics; perform pain behavior feature recognition according to the load transfer pattern characteristics to obtain pain feature correlation data; perform limb health assessment according to the pain feature correlation data to obtain the limb functional health index;

[0028] In the embodiments of the present invention, it mainly focuses on limb load-bearing and movement coordination, aiming to identify pain-related behavior characteristics and evaluate the limb functional health status. The key lies in extracting limb load-bearing data from the gait mechanics feature map and the gait deviation index matrix, calculating the load-bearing ratio distribution and the ipsilateral load-bearing difference rate (symmetry index); analyzing time parameters such as the duration of the stance phase of each limb; comprehensively identifying abnormal load transfer patterns (such as avoidance type) based on load-bearing, symmetry, and support time data; evaluating the range of motion of each joint according to the gait deviation index matrix and giving a score; matching characteristics such as load-bearing distribution, symmetry, support time, and load transfer patterns with a preset pain behavior feature library to identify pain-related feature combinations, locate the pain site, distinguish acute and chronic pain types, and calculate the pain intensity score; and finally integrating the pain intensity score, pain site location, and limb functional status score calculated based on load-bearing, symmetry, support time, and joint range of motion to generate a limb functional health index reflecting the health level of each limb.

[0029] Step S4: Perform temporal monitoring of the behavior pattern according to the limb functional health index to obtain the livestock health trend map.

[0030] In the embodiments of the present invention, the limb function health index obtained in the foregoing steps is integrated into the time series monitoring system, and comprehensive analysis is performed in combination with environmental and management data to achieve disease risk early warning and health trend prediction. The key lies in constructing a time series database of health index with livestock individuals and time as dimensions; obtaining and associating environmental monitoring data and ranch management records with health index data according to timestamps to form a health-environment-management association dataset; analyzing historical data in the time series database of health index, and calculating the baseline of health fluctuations (mean, standard deviation, smooth curve, prediction range); using the health-environment-management association dataset, applying a machine learning model to identify the association patterns between health anomalies and environmental / management factors for disease risk prediction; and finally generating a visual livestock health trend map to display the historical health curve, current status, future prediction, and relevant risk early warning information.

[0031] Particularly importantly, step S4 is specifically as follows:

[0032] Construct a time series database of health index according to the limb function health index and livestock individual identification information;

[0033] Obtain environmental monitoring data and ranch management records and perform environmental and management data association according to the time series database of health index to obtain a health-environment-management association dataset;

[0034] Calculate the health fluctuation baseline graph of the time series database of health index;

[0035] Perform disease risk early warning and trend map generation according to the health-environment-management association dataset to obtain a livestock health trend map;

[0036] In the embodiment of the present invention, the operation of constructing a health index time series database based on the limb function health index and livestock individual identification information is as follows: Receive the limb function health index of each limb (left front, right front, left rear, right rear) output in step S38 (for example, the left rear limb health index is 75, the right rear limb is 90, etc.), and the livestock individual identification information obtained in step S2 (for example, the ear tag number AB123). Use the livestock individual identification information as the primary key to create a health index time series database. The database adopts the structure of a time series database (such as InfluxDB or TimescaleDB), which is specifically used for efficiently storing and querying data sequences with timestamps. Create a record entry for each livestock in the database, and this entry contains the unique identifier (ear tag number) of the livestock and the health index data points arranged in chronological order. Each data point contains an accurate timestamp (for example, the time when the collection is completed, accurate to the second) and the health index value of each limb. For example, the database can be stored as: [timestamp, ear tag number, left front health index, right front health index, left rear health index, right rear health index]. Each time a gait monitoring is completed and the limb function health index is calculated, the system writes this data together with the current timestamp and the livestock ear tag number as a new record into this time series database. As time goes by and the number of monitoring times increases, the historical time series data of the health index of this livestock will accumulate in the database.

[0037] The operation of obtaining environmental monitoring data and ranch management records and associating environmental and management data according to the health index time series database to obtain a health-environment-management association dataset is as follows: The system obtains data from different data sources through interfaces. These data sources include environmental monitoring systems (e.g., recording livestock barn temperature, humidity, ammonia concentration, ventilation, etc.) and ranch management systems (e.g., recording feed batches, feeding amounts, drinking water conditions, vaccination records, veterinary treatment records, transfer information, calving dates, milking times and yields, etc.). Environmental monitoring data is usually stored in time series form, containing timestamps and various environmental parameter values. Ranch management records include event types, occurrence times, livestock involved (associated through ear tags), and relevant description information. Access the constructed health index time series database, and according to the timestamps of health index data points, find the corresponding environmental parameter values in the environmental monitoring data within the same time period (e.g., within 1 hour before and after). Then associate the environmental parameters with the health index data points. At the same time, find the management events related to the livestock that occurred within a certain period of time (e.g., within 24 hours before and after) the health index collection time point in the ranch management records, and associate the management event information with the health index data points. For example, if the health index of the left hind limb significantly decreases at a certain time point, the system will find information such as the temperature and humidity in the livestock barn, whether the feed batch has been changed, and whether vaccinations have been carried out at that time point, and attach this information to the health index record. This association operation forms a comprehensive dataset containing livestock identification, timestamp, limb health index, environmental parameters, and management events, that is, the health-environment-management association dataset.

[0038] The operations for calculating the health fluctuation baseline graph of the health index time series database are as follows: Access the constructed health index time series database. For each livestock, extract its historical health index time series data (at least a certain length of historical data, such as records for the past 30 days). Analyze the time series variation patterns of the health indices of each limb (left front, right front, left rear, right rear). Calculate the mean and standard deviation of the health index of each limb in the historical records as the health baseline value and normal fluctuation range for that limb. At the same time, identify the intraday fluctuation patterns of the health index (e.g., whether the health index regularly decreases during peak activity periods) and weekly / monthly trends (e.g., whether the health index shows specific changes as the lactation cycle progresses). Use a smoothing algorithm (such as moving average or exponential smoothing) to remove short-term noise and generate a smoothed historical curve of the health index. Based on the historical data, predict the normal fluctuation range of the health index for a future period (e.g., the next 7 days), for example, using a time series prediction model (such as an ARIMA model or an LSTM network) to predict the mean and confidence interval. Visualize the historical mean, standard deviation, smoothed curve, and predicted normal fluctuation range of the health index of each limb to form the health fluctuation baseline graph for that livestock. This baseline graph is an important reference for evaluating whether the current health status is abnormal.

[0039] For disease risk early warning and trend atlas generation based on the health-environment-management associated dataset, the operations to obtain the livestock health trend atlas are as follows: Utilize the health-environment-management associated dataset to analyze the statistical associations between abnormal fluctuations in health indices and environmental factors and management events. For example, if the hind limb health indices of multiple livestock generally decline after changing a certain feed batch, the system will identify the association between "feed batch change" and "decline in hind limb health index" and mark it as a potential risk factor. If the temperature in the livestock shed continuously exceeds 30°C and the health indices of multiple livestock (especially the overall health index or indices related to activity) generally decline, then "high temperature" is identified as an environmental risk. Establish a disease risk prediction model that takes as input the current health indices of livestock (the degree of deviation from the baseline), the identified abnormal gait patterns, as well as the current environmental parameters and recent management events. The model can use machine learning algorithms (e.g., decision trees, random forests, or neural networks), which are trained on the historical health-environment-management associated dataset to learn to identify patterns highly correlated with the occurrence of specific diseases (e.g., lameness, mastitis, respiratory diseases). For example, the model will learn that the combined pattern of "left hind limb health index lower than 60", "hind limb weight-bearing difference rate greater than 20%", and "recent group transfer" is highly correlated with the risk of lameness. The model outputs the risk probabilities of various potential diseases currently occurring in the livestock. Meanwhile, combining the historical trends in the health index time series database, the health fluctuation baseline graph, and the risk probabilities output by the prediction model, a livestock health trend atlas is generated. The trend atlas visually displays the historical change curves of the health indices of each limb, the degree of deviation of the current health status from the baseline, the predicted trends of future health indices (including the normal fluctuation range and risk intervals), as well as the environmental risk factors and management events associated with abnormal fluctuations in health indices. For example, the atlas can show that the left hind limb health index has been continuously declining in the past 3 days, the current value is already below the normal range of the baseline, and it is predicted to further decline in the next 7 days. At the same time, it marks that this decline is related to the recent high-temperature environment or feed adjustment and gives the early warning level of the lameness risk (e.g., high risk). This trend atlas provides an overview of the health status of individual and group livestock, potential risk alerts, and decision-making support for ranch managers.

[0040] Preferably, step S1 includes the following steps:

[0041] Step S11: Divide the pressure sensing pad into three functional areas, namely the front area, the middle area, and the rear area, according to the force characteristics of the livestock hoof. Each area is configured with an independent pressure sensor array, and a three-area pressure matrix is collected through the pressure sensor array;

[0042] Step S12: Set 6 high-speed cameras around the pressure sensing pad, calibrate the spatial coordinates of the camera system through a calibration board, and obtain a spatial coordinate mapping dataset;

[0043] Step S13: Set the livestock joint marking points for the livestock to be measured to obtain joint marking position data;

[0044] Step S14: Trigger synchronous data acquisition according to the three-region pressure matrix, the spatial coordinate mapping data set, and the joint marking position data to obtain a synchronous trigger time series;

[0045] Step S15: Identify the dynamic path and adjust the camera according to the three-region pressure matrix and the synchronous trigger time series to obtain dynamic path tracking data;

[0046] Step S16: Extract the ground reaction force data from the three-region pressure matrix according to the synchronous trigger time series and the dynamic path tracking data to obtain the mechanical distribution time series data;

[0047] Step S17: Calculate the joint motion parameter set according to the spatial coordinate mapping data set, the joint marking position data, and the synchronous trigger time series;

[0048] Step S18: Divide the gait cycle according to the mechanical distribution time series data and the joint motion parameter set to obtain the gait cycle segmentation data;

[0049] Step S19: Generate a gait mechanical feature map according to the mechanical distribution time series data, the joint motion parameter set, the gait cycle segmentation data, and the dynamic path tracking data.

[0050] In the embodiment of the present invention, the pressure sensing pad is divided into three functional regions: a front region, a middle region, and a rear region according to the force characteristics of the livestock hoof. For example, a resistive pressure sensing pad with a size of 2 meters × 1 meter and a sensor density of 4 per square centimeter is used. The first 0.5 meters along its length direction is divided into the front region, the middle 1 meter is divided into the middle region, and the last 0.5 meters is divided into the rear region. Each region contains an independent pressure sensor array, such as a grid composed of 200 × 100 independent pressure sensors. The size of each sensor unit is 0.5 cm × 0.5 cm, and the sensor sampling frequency is set to 200 Hz. The instantaneous pressure distribution data applied by the three regions during the livestock walking process is simultaneously and independently collected through these sensor arrays to form a three-region pressure matrix, which is a three-dimensional array with dimensions of time, region index (front, middle, rear), and sensor index (row, column) within the region.

[0051] Six high-speed cameras are set around the pressure sensing pad. For example, high-speed cameras of model Basler acA2040-90uc with a resolution of 2048×1536 pixels and a frame rate of 120 frames per second are used and arranged around the pressure pad. For example, 2 cameras are arranged on each side (long side), and 1 camera is arranged at each end (short side). The mounting height of the cameras is 2 meters, and the distance from the edge of the pressure pad is 3 meters to ensure that the 360° field of view above the pressure pad can be covered. Using a standard checkerboard calibration board, multiple images are taken at different positions and angles above and around the pressure pad, and the Zhang Zhengyou calibration method is applied to calculate the internal and external parameters of each camera. Based on these parameters, a spatial coordinate mapping relationship between a global coordinate system (for example, with the center of the pressure pad as the origin) and the coordinate system of each camera is established to form a spatial coordinate mapping data set, which includes the distortion coefficient, internal parameter matrix of each camera, and the rotation and translation matrices from the global coordinate system to the coordinate system of each camera.

[0052] Marker points are set for the key joint points of the livestock to be measured. For example, for a cow, at positions such as its shoulder joint (the highest point of the scapula), elbow joint (olecranon of the ulna), wrist joint (lateral radiocarpal joint), hip joint (greater trochanter of the femur), knee joint (lateral condyle of the femur), and ankle joint (lateral malleolus of the fibula), passive reflective marker points with a diameter of 15 mm are used and firmly attached through non-irritating adhesives or elastic straps. Ensure that the marker points do not fall off or shift during the walking of the livestock and can be clearly captured by at least two cameras from different angles. These preset joint point positions and their corresponding marker point numbers constitute the joint marker position data.

[0053] A pressure threshold trigger mechanism is set to achieve synchronous data acquisition. For example, a central control unit monitors the front zone pressure data from the three-zone pressure matrix. When it is detected that the total pressure of any sensor or sensor group in the front zone exceeds a preset threshold (set to 50 N), the central control unit immediately sends a synchronous trigger signal to the pressure data acquisition system and the high-speed camera system. After receiving the trigger signal, the pressure data acquisition system (operating at 200 Hz) and the camera system (operating at 120 Hz) start to synchronously record the data stream and attach a high-precision timestamp, such as a nanosecond-level timestamp, to each data record to ensure that the pressure data and the camera data are strictly aligned in time, forming a synchronous trigger time series.

[0054] Dynamic path recognition and camera adjustment are performed based on the three-region pressure matrix and the synchronous trigger time series. For example, during synchronous acquisition, the pressure data from the three-region pressure matrix is processed in real time, and the area with a pressure value greater than 50 N is identified as the hoofprint point. The centroid coordinates of each detected hoofprint point are calculated, and its timestamp is recorded to form the hoofprint spatio-temporal sequence data. The centroid connection vectors of the nearest three consecutive hoofprint points are calculated, and the instantaneous walking direction vector of the current hoofprint point sequence is calculated by vector averaging. Based on the hoofprint point positions within the last 5 seconds, a short-term livestock walking trajectory prediction model is constructed using the second-order polynomial fitting method to predict the possible positions of the livestock within the next 0.5 seconds. According to the predicted future positions and walking directions, the optimal camera angles that can maximize the capture of the joint markers on the side and back of the livestock are calculated. By sending control commands to the servo motors connected to the cameras, the horizontal (Pan) and vertical (Tilt) angles of the cameras are adjusted in real time, keeping the livestock always at the center of the camera's field of view and maintaining the best observation perspective, generating dynamic path tracking data.

[0055] Based on the synchronous trigger time series and the dynamic path tracking data, the ground reaction force data is extracted from the three-region pressure matrix. For example, for the three-region pressure matrix data stream collected synchronously, according to the hoofprint positions and timestamps identified in the dynamic path tracking data, the pressure data is assigned to the corresponding limbs. For each hoof strike event, the total vertical force (Fz, obtained by summing the pressures of all sensors in the hoofprint area and multiplying by the sensor area) and the estimated anterior-posterior (Fy) and lateral (Fx) shear forces (if the pressure pad sensor array includes shear force measurement functions or can be estimated by the pressure gradient of adjacent sensors) are extracted from the pressure pad sensor data. The curves of Fz, Fx, and Fy changing with time during the stance phase of each limb are calculated, and the time integrals (impulses) and peaks of these force curves are calculated. The extracted data constitutes the mechanical distribution time series data.

[0056] Calculate the joint motion parameter set based on the spatial coordinate mapping data set, joint marker position data, and synchronous trigger time series. For example, using the synchronous trigger time series, from the image sequence captured by the high-speed camera, use a marker point tracking algorithm based on deep learning or traditional image processing to identify and extract the two-dimensional coordinates of the livestock joint point markers on the image plane of each camera. Using the internal and external camera parameters in the spatial coordinate mapping data set, through a multi-view three-dimensional reconstruction algorithm (such as Direct Linear Transformation DLT or Bundle Adjustment), reconstruct the two-dimensional image coordinates into the three-dimensional spatial coordinates of the marker points in the global coordinate system. Apply a fourth-order Butterworth low-pass filter with a cut-off frequency of 8Hz to the reconstructed three-dimensional trajectory data for smoothing to eliminate noise. Based on the smoothed three-dimensional coordinates of the marker points, define the body segments of the livestock (such as the forelimb, hind limb, etc.), and calculate the posture (orientation) of each body segment. Calculate the relative angle between adjacent body segments as the joint angle (such as the flexion and extension angle, abduction and adduction angle, internal and external rotation angle of the elbow joint and knee joint). Numerically differentiate the joint angle time series to calculate the joint angular velocity. These three-dimensional coordinates, body segment postures, joint angles, and angular velocity data constitute the joint motion parameter set.

[0057] Divide the gait cycle according to the mechanical distribution time series data and the joint motion parameter set. For example, use the change in the vertical ground reaction force (Fz) in the mechanical distribution time series data to identify the touchdown and liftoff moments of the limb. Mark the moment when Fz exceeds 50N as touchdown (InitialContact, IC), and mark the moment when Fz continuously drops below 10N as liftoff (ToeOff, TO). A complete gait cycle (GaitCycle, GC) is defined as the time period from the touchdown of a certain limb to the next touchdown of the same limb. The stance phase (StancePhase, SP) is defined as the time period from IC to TO, and the swing phase (SwingPhase, SwP) is defined as the time period from TO to the next IC. Calculate the duration of each gait cycle, stance phase, and swing phase, and calculate the proportion of the stance phase and swing phase in the entire gait cycle. These timestamp, duration, and proportion data constitute the gait cycle segmentation data.

[0058] Generate a gait mechanics feature map based on the mechanical distribution time series data, joint motion parameter set, gait cycle segmentation data, and dynamic path tracking data. For example, for each identified gait cycle, perform time normalization on the mechanical distribution time series data (Fz, Fx, Fy time series) and joint motion parameter set (joint angle, angular velocity time series) within that cycle. For example, resample to 100 data points, representing 0% to 100% of the gait cycle. Extract key scalar parameters such as the peak values, impulses, support phase duration, swing phase duration, step length (distance between the centroids of two consecutive ipsilateral footprints), step width (distance between the centroids of two consecutive contralateral footprints perpendicular to the forward direction), step frequency (steps per minute), etc. Integrate this normalized time series data and scalar parameters into a structured dataset, for example, store it in the HDF5 format. This dataset contains all relevant mechanical, kinematic, time, and space parameters for each limb and each gait cycle, forming a gait mechanics feature map.

[0059] Preferably, step S15 includes:

[0060] Extract the hoofprint spatio-temporal sequence data from the three-region pressure matrix according to the synchronous trigger time series;

[0061] Calculate the path direction vector based on the hoofprint spatio-temporal sequence data to obtain livestock motion vector data;

[0062] Calculate the path prediction curve data based on the livestock motion vector data and the hoofprint spatio-temporal sequence data;

[0063] Identify turning points based on the livestock motion vector data and the hoofprint spatio-temporal sequence data to obtain behavior mutation marker data;

[0064] Calculate the optimal observation angle based on the path prediction curve data and the behavior mutation marker data to obtain camera angle optimization parameters;

[0065] Collect dynamic path tracking data according to the camera angle optimization parameters.

[0066] In the embodiments of the present invention, hoofprint spatio-temporal sequence data in the three-region pressure matrix is extracted according to the synchronous trigger time sequence. After the central control unit receives the synchronous trigger signal, the system starts to collect the three-region pressure matrix data of the pressure sensing pad at a frequency of 200 Hz. For each frame of pressure data, the sensor array is traversed to identify the sensor units with a pressure value greater than 50 N. The adjacent sensor units with pressure values exceeding the threshold are clustered to form the identified hoofprint regions. The pressure-weighted centroid coordinates (x, y) of each hoofprint region are calculated, that is, centroid coordinates = ∑(sensor pressure × sensor coordinates) / ∑(sensor pressure). Each calculated hoofprint centroid coordinate (x, y) is associated with its corresponding acquisition timestamp and organized into a sequence data arranged in chronological order to form hoofprint spatio-temporal sequence data, for example, stored as a list containing [timestamp, x coordinate, y coordinate] tuples.

[0067] Path direction vector calculation is performed based on the hoofprint spatio-temporal sequence data. From the hoofprint spatio-temporal sequence data, the three continuously detected nearest hoofprint points are extracted , and , whose coordinates are respectively , and , and the corresponding timestamps are . The displacement vectors and between adjacent points are calculated. By performing vector average calculation on and , the current walking direction vector is obtained. At the same time, the time difference between the two nearest points is calculated. The instantaneous walking speed is calculated as the modulus of the vector divided by the time difference , that is, . The calculated walking direction vector, instantaneous speed, and the current timestamp are associated to form livestock movement vector data.

[0068] Path prediction curve data is calculated based on the livestock movement vector data and the hoofprint spatio-temporal sequence data. Using the hoofprint spatio-temporal sequence data within the last 5 seconds, its (x, y) coordinates and the corresponding timestamps are extracted. The timestamps are normalized with respect to the current time. A second-order polynomial model is used to fit the walking trajectory of the livestock, that is, assuming that the positions in the x direction and y direction are quadratic functions of time: . The least squares method is used to solve for the coefficients such that the deviation between the fitting curve and the hoofprint point data within the last 5 seconds is minimized. The fitted model is extrapolated to the time point 0.5 seconds in the future, and the predicted (x, y) coordinate sequence is calculated to form the path prediction curve data. At the same time, the curvature of the prediction curve is calculated, , and predict the changing trend of the path direction.

[0069] Identify turning points based on livestock motion vector data and hoofprint spatio-temporal sequence data. Extract a continuous sequence of walking direction vectors from the livestock motion vector data. Calculate the angle θ between adjacent direction vectors. Calculate the instantaneous change rate of this angle, i.e., Δθ / Δt. When the change rate of the angle continuously exceeds 15° / second, it is determined that the livestock is turning at the current moment and marked as a potential turning point. At the same time, analyze the hoofprint spatio-temporal sequence data and calculate the time interval between continuously detected hoofprint points (regardless of limb). If a certain time interval exceeds 1.5 times the historical average gait cycle time of this livestock (for example, take the average gait cycle in the individual gait baseline parameter set), it is determined that the livestock has a pause or hesitation and is marked as a potential pause point. Associate these marks with the corresponding timestamps to form behavior mutation mark data.

[0070] Calculate the optimal observation angle based on the path prediction curve data and behavior mutation mark data. According to the path prediction curve data, determine the predicted position and forward direction of the livestock within the next 0.5 seconds. Combine the livestock body size information and joint mark position data to calculate the vector from the current position of each camera to the predicted livestock position, and the angle between this vector and the predicted forward direction. The goal is to calculate the angle that can maximize the exposure of the side or back of the livestock (i.e., the main distribution area of the mark points) in the camera's field of view. For the area marked as a potential turning point, calculate the composite angle that can simultaneously capture the motion of the inner and outer limbs of the livestock. For the area marked as a potential pause point, calculate the angle that provides stable, multi-perspective observations (such as the front side and the back side). Considering all 6 cameras, calculate a set of horizontal (Pan) and vertical (Tilt) angles of the cameras that optimize the overall observation effect through an optimization algorithm, forming camera angle optimization parameters, such as a list containing the target Pan and Tilt angles of each camera.

[0071] Collect dynamic path tracking data according to the camera angle optimization parameters. Send the calculated camera angle optimization parameters to the servo motors connected to the high-speed cameras through the control interface. The servo motors adjust the Pan and Tilt angles of the cameras in real time according to the received parameters. The camera system continuously captures images at a frequency of 120 frames per second while adjusting the angles. By processing the captured images, such as using a mark point tracking algorithm, the position of the mark points in the adjusted camera field of view is extracted in real time. Combining the real-time internal and external parameters and pose of the camera (obtained from the servo motor angle feedback), the two-dimensional image coordinates of the mark points are reconstructed into three-dimensional coordinates in the global coordinate system. These real-time updated three-dimensional coordinate sequences of the mark points are the dynamic position and pose information of the livestock during walking, constituting the dynamic path tracking data.

[0072] Preferably, the individual gait timing segmentation in step S2 includes:

[0073] Obtain the livestock individual identification information and construct and access the historical database according to the gait mechanics feature map to obtain the individual historical gait record set;

[0074] Calculate the individual gait baseline parameter set of the individual historical gait record set;

[0075] Standardize the current gait parameters for each parameter in the gait mechanics feature map to obtain the standardized gait parameters;

[0076] Perform time series segmentation and alignment on the standardized gait parameters and the individual gait baseline parameter set to obtain the segmented and aligned gait data.

[0077] In the embodiment of the present invention, the livestock individual identification information is obtained and the historical database is constructed and accessed according to the gait mechanics feature map. After the system collects the gait mechanics feature map, it reads the livestock individual identification information contained therein, such as the unique ear tag number. Using this ear tag number as the retrieval key, the historical gait records of this livestock are searched in a structured database (such as using a PostgreSQL database). If the record of this livestock already exists in the database, all the gait mechanics feature map data within the past 30 days is extracted to form the individual historical gait record set. If the record of this livestock does not exist in the database, the currently collected gait mechanics feature map is stored in the database as the first record of this livestock and marked as the basic record. The database is designed as a relational database, including a livestock information table (ear tag number, breed, age, weight, etc.) and a gait record table (ear tag number, acquisition timestamp, storage path or serialized data of the gait mechanics feature map data). After each livestock accumulates at least 5 valid gait records (such as collected on different dates and time periods), its historical record set is used for subsequent baseline calculation.

[0078] Calculate the individual gait baseline parameter set of the individual historical gait record set. Perform statistical analysis on all the gait mechanics feature map data in the individual historical gait record set obtained by accessing the historical database. For example, for key parameters such as gait cycle time, maximum flexion and extension angles of each joint, peak vertical ground reaction force, etc., calculate their mean (μ) and standard deviation (σ) in the historical records. Before calculating the mean and standard deviation, apply a robust statistical method, such as iteratively removing data points whose distance from the mean exceeds 3 standard deviations, to exclude the influence of outliers on the baseline calculation. Store the calculated historical mean, standard deviation and coefficient of variation (CV = σ / μ×100%) of each gait parameter to form the normal gait reference range of this livestock, that is, the individual gait baseline parameter set. This baseline parameter set is stored in a structured data format (such as JSON or XML) and associated with the livestock individual identification information.

[0079] Standardize the current gait parameters for each parameter in the gait mechanics feature map. Compare each scalar parameter (such as gait cycle time, peak joint angle, peak force, etc.) in the currently collected gait mechanics feature map with the individual gait baseline parameter set of the livestock. For each parameter, calculate its standardized score using the Z-score standardization method, that is: standardized score = (current parameter value - baseline mean) / baseline standard deviation. If the baseline standard deviation is close to zero (indicating extremely small historical fluctuations), the mean of the historical absolute deviations of the parameter is used as the denominator. For time series parameters (such as normalized joint angle curves, force curves), calculate the root mean square error (RMSE) or correlation coefficient between the curve mean of the corresponding time points of the current curve and the baseline as the standardization index. These calculated standardized values or indices constitute the standardized gait parameters.

[0080] Perform time series segmentation and alignment on the standardized gait parameters and the individual gait baseline parameter set. According to the historical average gait cycle information of the livestock recorded in the individual gait baseline parameter set, divide the time series data (such as joint angle change curves, force curves) in the current standardized gait parameters according to the gait cycle. A gait cycle is divided into four stages: initial contact period (from InitialContact to LoadingResponse), support period (from LoadingResponse to MidStance), push-off period (from MidStance to TerminalStance), and swing period (from TerminalStance to InitialContact). The division points of these stages are based on key events (such as touchdown, liftoff) in the gait cycle segmentation data. To eliminate the difference between the current gait speed and the historical baseline speed, process the time series data of each stage of the current gait using the Dynamic Time Warping (DTW) algorithm to align its time axis with the time axis of the baseline gait. The DTW algorithm realizes non-linear time alignment by finding the best matching path between two time series. After alignment, compare the standardized parameters of each stage of the current gait with the parameters of the corresponding stage of the baseline point by point to form segmented aligned gait data. These data are stored in a structured format, including the standardized values of each stage and each parameter at the aligned time points.

[0081] Preferably, the adaptive gait threshold calculation in step S2 includes:

[0082] Extract the livestock feature standardized parameters according to the livestock basic information and the individual gait baseline parameter set;

[0083] Calculate the gait stability index based on the segmented aligned gait data to obtain the gait fluctuation feature spectrum;

[0084] Perform parameter importance weight allocation based on the gait fluctuation feature spectrum and livestock characteristic standardization parameters to obtain a parameter importance weight table;

[0085] Identify individual-specific gait characteristics based on the gait fluctuation feature spectrum;

[0086] Calculate the joint angle abnormality threshold according to the individual-specific gait characteristics;

[0087] Calculate the mechanical parameter abnormality threshold according to the individual-specific gait characteristics;

[0088] Calculate the time parameter abnormality threshold according to the gait fluctuation feature spectrum;

[0089] Integrate the joint angle abnormality threshold, the mechanical parameter abnormality threshold, and the time parameter abnormality threshold to obtain an individualized abnormality threshold set.

[0090] In the embodiments of the present invention, livestock characteristic standardization parameters are extracted according to the livestock basic information and the individual gait baseline parameter set. From the livestock basic information database, static characteristics of the livestock to be monitored are read, such as weight (unit: kg), age (unit: month), and breed type (e.g., Holstein cow, Simmental cow). At the same time, the standard weight and expected lifespan data of this breed of livestock are accessed. Calculate the body mass index, that is, body mass index = actual weight / breed standard weight. Calculate the age coefficient, that is, age coefficient = actual age / breed expected lifespan. These calculated values, such as a body mass index of 1.05 (indicating 5% heavier than the standard weight) and an age coefficient of 0.6 (indicating 60% of the expected lifespan), constitute the livestock characteristic standardization parameters, which are adjustment factors for subsequent threshold calculation.

[0091] Calculate the gait stability index according to the segmented aligned gait data. Using the individual historical gait record set (at least 5 records) in the historical database and the currently collected and segmented aligned gait data aligned with the baseline, calculate the variability index of each gait parameter. For parameters such as gait cycle time, stance phase duration, swing phase duration, maximum / minimum angles of each joint during the stance and swing phases, peak vertical ground reaction force, and peak anterior-posterior shear force, calculate their coefficient of variation (CV) in all historical records. The coefficient of variation calculation formula is: CV = (standard deviation / mean) × 100%, where μ is the historical mean of the parameter and σ is the historical standard deviation of the parameter. A parameter with a low coefficient of variation indicates that the historical fluctuation of this parameter of the livestock is small and the gait is stable; a parameter with a high coefficient of variation indicates large fluctuations and poor stability. Combine the coefficient of variation of each parameter to form a gait fluctuation feature spectrum, such as a list containing [parameter name, CV value] tuples.

[0092] The parameter importance weights are assigned according to the gait fluctuation feature spectrum and the livestock characteristic standardization parameters. Based on the breed type of the livestock (obtained from the livestock characteristic standardization parameters), a preset parameter importance weight table is consulted. For example, for dairy cattle breeds, the weight-bearing parameters of the hind limbs (such as the peak vertical force and the weight-bearing time) and the mobility parameters of the hind limb joints (such as the knee joint and the ankle joint) are given higher importance weights (e.g., 0.8 - 1.0), while the weights of some parameters of the forelimbs are slightly lower (e.g., 0.6 - 0.8). For beef cattle breeds, more attention is paid to the overall movement coordination and the balanced weight-bearing of the four limbs. According to the CV values of the parameters in the gait fluctuation feature spectrum, the preset weights are fine-tuned: for the key parameters with lower CV values (higher stability), their weights are further increased to emphasize their value as a stable baseline; for the parameters with extremely high CV values (poor stability), their weights are slightly decreased, indicating that the parameter itself is unstable and its change does not fully represent an abnormality. These weight values form the parameter importance weight table, such as a dictionary containing tuples of [parameter name, importance weight].

[0093] Identify the individual-specific gait characteristics according to the gait fluctuation feature spectrum. Analyze the parameters in the gait fluctuation feature spectrum with lower CV values (e.g., CV < 10%). At the same time, compare the historical mean values of these parameters with the population mean values of livestock of the same breed and the same age group. If there is a significant difference between the historical mean value of a certain parameter and the population mean value (e.g., the difference exceeds 1.5 times the population standard deviation), but the parameter shows low variability in the historical records of this individual livestock, that is, it stably deviates from the population mean, then it is determined that this deviation is the gait rhythm characteristic specific to this individual livestock, rather than an abnormality. For example, if the peak vertical force of the left hind limb of a cow is always 10% lower than the population mean value, but its CV value is only 5%, then it is marked as the characteristic of the individual-specific lower weight-bearing of the left hind limb. These identified individual-specific gait characteristics and their deviation degrees are recorded to form the individual-specific gait characteristic record.

[0094] Calculate the joint angle abnormality threshold based on the gait fluctuation feature spectrum, parameter importance weight table, and individual-specific gait feature record. For each joint angle parameter in the segmented-aligned gait data (such as the maximum angle of knee joint flexion and extension during the stance phase), set the initial threshold according to its coefficient of variation (σ) in the gait fluctuation feature spectrum and the historical mean (μ) in the individual gait baseline parameter set. For example, use the threshold based on historical standard deviation: upper threshold = μ + k × σ, lower threshold = μ - k × σ. The coefficient k is dynamically set according to the CV value of this parameter and its weight in the parameter importance weight table, e.g., k = f(CV, ImportanceWeight), where f is a decreasing function (the higher the CV, the larger k; the higher the weight, the smaller k). Then, adjust the threshold according to the individual-specific gait feature record: if a joint angle parameter is identified as an individual-specific feature, adjust its baseline mean to the individual historical mean and appropriately tighten the threshold range (reduce the value of k) according to its individual stability (low CV). These calculated upper and lower limit values constitute the joint angle abnormality threshold, e.g., a list of tuples containing [joint name, gait phase, parameter name, lower threshold, upper threshold].

[0095] Calculate the mechanical parameter abnormality threshold based on the gait fluctuation feature spectrum, parameter importance weight table, and individual-specific gait feature record. For mechanical parameters such as ground reaction force (such as peak vertical force, anteroposterior shear force impulse), also set the initial threshold based on its historical distribution characteristics (mean μ, standard deviation σ) in the gait fluctuation feature spectrum and the weight in the parameter importance weight table. The percentile-based method can be used: the lower threshold is set to the 5th percentile of the historical data, and the upper threshold is set to the 95th percentile of the historical data. Adjust the percentile range according to the CV value and importance weight of this parameter. For example, for parameters with high importance and good stability, the threshold range is tightened to the 2.5th percentile to the 97.5th percentile. If a mechanical parameter is identified as an individual-specific gait feature, such as consistently low load-bearing, calculate the threshold based on its individual historical mean and standard deviation instead of the population mean, and adjust the threshold range according to its individual stability. These calculated values constitute the mechanical parameter abnormality threshold.

[0096] Calculate the time parameter abnormality threshold based on the gait fluctuation feature spectrum and livestock characteristic standardized parameters. For time parameters such as gait cycle, stance phase duration, swing phase duration, set the initial threshold based on their historical distribution (mean μ, standard deviation σ) in the gait fluctuation feature spectrum. Adjust the threshold range according to the age coefficient and body weight index in the livestock characteristic standardized parameters. For example, for older livestock, their gait cycle will be longer and the daily fluctuation will increase, so appropriately relax the upper threshold and fluctuation range of the time parameter. Livestock with a significant deviation from the standard body weight also have different gait rhythms. For example, upper threshold = + + + , the lower threshold limit = - - - , where is an adjustment coefficient set according to experience or historical population data. These calculated values constitute the abnormal threshold of time parameters.

[0097] Integrate the abnormal threshold of joint angles, the abnormal threshold of mechanical parameters, and the abnormal threshold of time parameters. Pool together the calculated abnormal threshold of joint angles, the abnormal threshold of mechanical parameters, the abnormal threshold of time parameters, and the abnormal threshold of lateral movement amplitude (if calculated). Structurally organize all thresholds, for example, classify them according to parameter type, limb, and gait phase. Store these values as a complete set of individualized abnormal thresholds, such as a dictionary or list containing all parameter names and their corresponding upper and lower threshold values. This set is the individualized criterion for determining whether the gait of this specific livestock is abnormal in the current state.

[0098] Preferably, the abnormal gait pattern recognition in step S2 includes:

[0099] Detect abnormal changes in joint angles of the segmented and aligned gait data according to the set of individualized abnormal thresholds to obtain abnormal joint angle marker data;

[0100] Detect abnormal ground reaction forces of the segmented and aligned gait data according to the set of individualized abnormal thresholds to obtain abnormal mechanical parameter marker data;

[0101] Analyze the lateral movement amplitude of the segmented and aligned gait data according to the set of individualized abnormal thresholds to obtain abnormal lateral movement marker data;

[0102] Conduct time correlation analysis on the abnormal joint angle marker data, abnormal mechanical parameter marker data, and abnormal lateral movement marker data to obtain abnormal pattern time series correlation data;

[0103] Generate a gait deviation index matrix based on the abnormal joint angle marker data, abnormal mechanical parameter marker data, abnormal lateral movement marker data, and abnormal pattern time series correlation data.

[0104] In the embodiments of the present invention, the abnormal detection of joint angle changes is performed on the segmented and aligned gait data according to the individualized abnormal threshold set. For the time series of the standardized angle change rate of each joint (such as the knee joint) in a specific gait phase (such as the stance phase) in the segmented and aligned gait data, it is compared with the upper and lower threshold values corresponding to this parameter in the individualized abnormal threshold set. For example, if the standardized flexion-extension angle change rate of the knee joint at a certain time point in the stance phase is higher than the upper limit of its individualized abnormal threshold or lower than its lower limit, it is marked as abnormal at this time point. This process is repeated for the joint angle change rate parameters of all joints and all gait phases. All time points outside the threshold range, the corresponding abnormal parameters, and the degree of deviation are recorded to form the joint angle abnormal marking data, such as a list containing tuples of [limb, joint, gait phase, time point (percentage of gait cycle), abnormal parameter name, deviation value, abnormal level].

[0105] The abnormal detection of ground reaction forces is performed on the segmented and aligned gait data according to the individualized abnormal threshold set. For the time series of the standardized vertical ground reaction force, anterior-posterior shear force, and lateral shear force of each limb during the stance phase in the segmented and aligned gait data, as well as scalar parameters such as the peak value and impulse of these forces, they are compared with the corresponding threshold values in the individualized abnormal threshold set. For example, if the peak value of the vertical force of a certain limb during the stance phase is lower than the lower limit of its individualized abnormal threshold (indicating weight reduction), or the impulse of the anterior-posterior shear force exceeds its upper limit (indicating abnormal pushing off the ground), it is marked as abnormal. This process is repeated for the mechanical parameters of all limbs. All parameters outside the threshold range and the degree of deviation are recorded to form the mechanical parameter abnormal marking data, such as a list containing tuples of [limb, gait phase, parameter name, deviation value, abnormal level]. The abnormal level can be divided according to the degree of deviation from the threshold, such as mild (deviation of 1 - 2 standard deviations), moderate (deviation of 2 - 3 standard deviations), and severe (deviation of more than 3 standard deviations).

[0106] The analysis of the lateral movement amplitude is performed on the segmented and aligned gait data according to the individualized abnormal threshold set. In the segmented and aligned gait data, the three-dimensional position information of the center of gravity of the livestock body in the gait cycle is included. The time series of the displacement of the center of gravity in the lateral direction (perpendicular to the forward direction) is extracted. The maximum lateral swing amplitude of the center of gravity in the gait cycle is calculated. This amplitude is compared with the upper and lower threshold values corresponding to this parameter in the individualized abnormal threshold set. For example, if the lateral swing amplitude exceeds the upper threshold (indicating gait instability or pain compensation), it is marked as abnormal lateral movement. At the same time, the symmetry of the lateral movement of the center of gravity is analyzed, and the difference rate of the left and right swing amplitudes is calculated. If the difference rate exceeds the threshold, it is also marked as abnormal. These abnormal lateral movement situations are recorded to form the abnormal lateral movement marking data, such as a list containing tuples of [parameter name (such as lateral swing amplitude, left-right symmetry), deviation value, abnormal level].

[0107] Perform a temporal correlation analysis on the marked data of joint angle abnormalities, mechanical parameter abnormalities, and lateral movement abnormalities. Analyze the occurrence time points and durations within the gait cycle of the marked joint angle abnormalities, mechanical parameter abnormalities, and lateral movement abnormalities. For example, if the abnormal decrease in the peak vertical force of a certain limb occurs simultaneously with or within a very short time interval of the abnormal increase in the flexion angle of the knee joint of that limb during the stance phase, a temporal correlation is considered to exist between them. Identify whether there are frequently co-occurring patterns among different abnormal marked data, such as "decreased weight bearing of a certain limb" often accompanied by "excessive flexion of the ipsilateral wrist joint". Construct a temporal correlation graph of abnormal patterns, where the nodes represent different abnormal parameters, the edges represent their co-occurrence or sequential relationships within the gait cycle, and the weights of the edges represent the strength or frequency of the association. These association information are recorded to form temporal correlation data of abnormal patterns.

[0108] Generate a gait deviation index matrix based on the marked data of joint angle abnormalities, mechanical parameter abnormalities, lateral movement abnormalities, and temporal correlation data of abnormal patterns. Integrate all the marked abnormal data and calculate a quantified gait deviation index for each key gait parameter (e.g., peak vertical force of the left front limb, maximum flexion angle of the right hind limb knee joint, overall lateral swing amplitude). The calculation of the deviation index is based on the degree of deviation from the threshold of the parameter (e.g., percentage of deviation from the upper / lower limit of the threshold) and its abnormal level. For example, the more the deviation from the threshold and the higher the level, the higher the deviation index. The value range of the deviation index is set from 0 to 10 points, where 0 indicates that the parameter is within the normal threshold range or has only a very slight deviation, and 10 indicates that the parameter seriously exceeds the threshold and has a high abnormal level. At the same time, consider the temporal correlation data of abnormal patterns: if the abnormality of a certain parameter has a strong association with the abnormalities of multiple other important parameters, its deviation index is appropriately increased. Summarize the deviation indices of all key gait parameters into a matrix, such as a two-dimensional matrix with rows representing parameters (e.g., peak vertical force of the left front limb) and columns representing different types of deviation indices (e.g., degree of deviation from the threshold, abnormal level, final deviation index), to form a gait deviation index matrix. This matrix clearly quantifies the degree of deviation of each gait parameter from the individual's normal baseline.

[0109] Preferably, step S3 includes:

[0110] Step S31: Extract limb weight-bearing data from the gait deviation index matrix and the gait mechanical feature map to obtain the original limb weight-bearing data;

[0111] Step S32: Calculate the weight-bearing ratio distribution map based on the original limb weight-bearing data and the body weight data in the livestock basic information;

[0112] Step S33: Calculate the weight-bearing difference rates between the left front limb and the right front limb, and between the left hind limb and the right hind limb according to the weight-bearing ratio distribution map. Mark the weight-bearing difference rates exceeding 15% as asymmetric weight-bearing, then calculate the asymmetry degree index, and finally obtain the limb weight-bearing symmetry index;

[0113] Step S34: Conduct weight-bearing time analysis based on the gait mechanics characteristic map and the limb weight-bearing symmetry index to obtain the support time characteristic data;

[0114] Step S35: Identify the weight-bearing transfer pattern characteristics based on the weight-bearing ratio distribution map, the limb weight-bearing symmetry index, and the support time characteristic data;

[0115] Step S36: Evaluate the joint range of motion of the gait mechanics characteristic map according to the gait deviation index matrix to obtain the joint range of motion score;

[0116] Step S37: Identify the pain behavior characteristics based on the weight-bearing ratio distribution map, the limb weight-bearing symmetry index, the support time characteristic data, and the weight-bearing transfer pattern characteristics to obtain the pain characteristic correlation data;

[0117] Step S38: Conduct limb health assessment according to the pain characteristic correlation data to obtain the limb functional health index.

[0118] In the embodiment of the present invention, limb weight-bearing data is extracted from the gait deviation index matrix and the gait mechanics characteristic map. From the gait mechanics characteristic map, for each identified gait cycle and each limb (left front, right front, left hind, right hind), the time series data of the vertical ground reaction force (Fz) during the stance phase is extracted. Calculate the maximum value of Fz (peak weight-bearing) and the time integral of Fz (impulse, representing the force effect during the total weight-bearing time) for each limb during the stance phase. At the same time, consult the gait deviation index matrix to obtain the deviation index of the vertical ground reaction force parameters (peak value, impulse, curve shape) of each limb. Associate these extracted values and deviation indices to form the original limb weight-bearing data, such as a structured data containing fields such as [limb, gait cycle ID, vertical force peak value, vertical force impulse, vertical force peak value deviation index, vertical force impulse deviation index].

[0119] Calculate the load-bearing ratio distribution map based on the original limb load-bearing data and the body weight data in the livestock basic information. Obtain the body weight data of the current livestock from the livestock basic information database (for example: 600 kg). For each gait cycle in the original limb load-bearing data, calculate the sum of the peak vertical forces of the four limbs, which represents the total instantaneous load-bearing in this gait cycle. Divide the peak vertical force of each limb by the body weight of the livestock to calculate the relative load-bearing ratio of this limb. For example, if the peak vertical force of the left front limb is 3500 N and the body weight of the livestock is 600 kg (about 5880 N), then the relative load-bearing ratio of the left front limb is 3500 / 5880≈0.595, that is, about 59.5% of the body weight. Calculate the average load-bearing ratio of each limb in multiple gait cycles. Visualize these average load-bearing ratios, for example, using a four-quadrant graph, where each quadrant represents a limb, and the shade of color or the size of the value represents the load-bearing ratio, to form the load-bearing ratio distribution map.

[0120] Calculate the load-bearing difference rate between the left front limb and the right front limb, and between the left hind limb and the right hind limb based on the load-bearing ratio distribution map. Extract the average load-bearing ratios of the left front limb, right front limb, left hind limb, and right hind limb from the load-bearing ratio distribution map. Calculate the load-bearing difference rate between the same-side limbs (left and right front limbs, left and right hind limbs). For example, if the load-bearing ratio of the left front limb is P_LF and the load-bearing ratio of the right front limb is P_RF, then the load-bearing difference rate of the front limbs = |P_LF - P_RF| / ((P_LF + P_RF) / 2)×100%. Calculate the load-bearing difference rate of the hind limbs = |P_LH - P_RH| / ((P_LH + P_RH) / 2)×100%. If the load-bearing difference rate of the front limbs or hind limbs exceeds 15%, it is marked as asymmetric load-bearing. Calculate the asymmetry degree index. For example, map the difference rate to a score of 0 - 10. A difference rate of 15% corresponds to a score of 0, and a difference rate of 50% corresponds to a score of 10. These calculated difference rates and asymmetry degree indexes constitute the limb load-bearing symmetry index.

[0121] Conduct load-bearing time analysis based on the gait mechanics characteristic map and the limb load-bearing symmetry index. From the gait mechanics characteristic map, extract the duration of the stance phase of each limb in each gait cycle. Calculate the proportion of the stance phase duration of each limb in the total time of this gait cycle. Calculate the average stance phase duration and its proportion of each limb in multiple gait cycles. If the stance phase duration of a certain limb is significantly shortened (for example, shortened by more than 20% compared to the individual baseline average) and the limb load-bearing symmetry index shows asymmetric load-bearing of the same-side limbs, it indicates that the limb has pain reduction by shortening the support time to reduce the load-bearing. Record these stance phase durations, proportions, and the degree of deviation from the baseline to form the support time characteristic data.

[0122] Identify the characteristics of the weight-bearing transfer pattern based on the weight-bearing ratio distribution map, the limb weight-bearing symmetry index, and the support time characteristic data. Analyze the dynamic process of the change in the weight-bearing ratio of the four limbs of the livestock over time during the gait cycle. Under normal circumstances, the weight-bearing is transferred gradually and smoothly among the four limbs. Identify abnormal weight-bearing transfer sequences. For example, if, after a certain limb (such as the left hind limb) touches the ground, its vertical force curve quickly reaches a peak and then drops rapidly, while the weight-bearing of the contralateral limb (right hind limb or left forelimb) increases rapidly, this is an avoidance-type weight-bearing transfer, indicating pain in the left hind limb. If the livestock shows obvious "jumping" or "hurrying" phenomena during walking, manifested as an extremely short support time for a certain limb and a rapid transfer of weight-bearing to other limbs, this indicates acute pain. Based on the weight-bearing ratio, symmetry, and support time data, identify these abnormal weight-bearing transfer patterns through a pattern matching algorithm to form the weight-bearing transfer pattern characteristics, such as a list containing [pattern type (such as avoidance type, jumping type), involved limb, pattern intensity].

[0123] Evaluate the range of joint motion of the gait mechanics feature map according to the gait deviation index matrix. From the gait mechanics feature map, extract the range of motion parameters such as the maximum flexion and extension angles, abduction and adduction angles, and internal and external rotation angles of each joint (such as the shoulder, elbow, wrist, hip, knee, and ankle joints) during the stance phase and swing phase. Refer to the gait deviation index matrix to obtain the deviation index of these joint range of motion parameters. For example, if the maximum flexion angle of the left hind limb knee joint during the swing phase is significantly smaller than the individual baseline (high deviation index), it indicates limited joint motion. Calculate a joint range of motion score on a scale of 0 - 100 based on the deviation index of each joint range of motion parameter, where 100 represents completely normal joint motion (deviation index close to 0), and 0 represents severely limited joint motion (deviation index close to 10). Summarize the joint range of motion scores of the main joints of the four limbs to form the joint range of motion score data.

[0124] Identify pain behavior characteristics based on the load-bearing ratio distribution map, limb load-bearing symmetry index, support time characteristic data, and load transfer pattern characteristics. Based on a known animal pain ethology characteristic library, which includes pain-related gait performance characteristics described in, for example, the "Lameness Scoring Standard", such as load-bearing asymmetry, shortened stride length, shortened support time, head nodding, arched back, etc. Match the load-bearing ratio distribution map, limb load-bearing symmetry index, support time characteristic data, and load transfer pattern characteristics calculated in the previous steps with the pain behavior characteristic library. For example, if the limb load-bearing symmetry index shows that the load difference rate of the left hind limb exceeds 25%, the support time characteristic data shows that the support phase duration of the left hind limb is shortened by more than 30%, and an avoidance-type load transfer pattern involving the left hind limb is identified, then these characteristic combinations are highly correlated with left hind limb pain. Calculate a pain correlation score for each identified pain-related characteristic, with a high score indicating a strong correlation between the characteristic and pain. Record these characteristics, the involved limbs, and the pain correlation scores to form pain characteristic association data.

[0125] Conduct limb health assessment based on the pain characteristic association data and joint range of motion scores. Integrate the pain characteristic association data and joint range of motion scores to evaluate the health status of each limb (and the whole body). For example, for the left hind limb, if the pain characteristic association data shows multiple strongly correlated pain characteristics such as load-bearing asymmetry, shortened support time, and the presence of an avoidance-type load transfer, and the joint range of motion scores show severe limitations in the knee and ankle joint activities, then it is determined that the left hind limb has a relatively high degree of pain and dysfunction. Calculate a comprehensive limb function status score on a scale of 0 - 100, where 100 represents completely normal function, and a decrease in the score reflects the degree of dysfunction. At the same time, calculate a pain likelihood rating on a scale of 0 - 10 based on the strength and quantity of the pain characteristic association data, as well as the degree of joint range of motion limitation. For example, 0 indicates no pain signs, and 10 indicates extreme pain. Integrate the load-bearing ratio values, function status scores, and pain likelihood ratings of the four limbs to form a limb function health index, and clearly mark the problematic limbs and their pathological manifestations (e.g., left hind limb function score 60, pain rating 8, manifested as insufficient load-bearing, short support time, and limited knee joint activity).

[0126] Preferably, step S37 includes:

[0127] Extract a standardized pain feature vector based on the load-bearing ratio distribution map, limb load-bearing symmetry index, support time characteristic data, and load transfer pattern characteristics;

[0128] Obtain livestock breed reference data and determine characteristic thresholds based on the standardized pain feature vector to obtain a pain feature active state table;

[0129] Perform feature co-occurrence pattern analysis according to the pain feature activity status table to obtain feature co-occurrence pattern data;

[0130] Perform pain location analysis according to the feature co-occurrence pattern data to obtain a pain location map;

[0131] Perform compensatory behavior recognition according to the feature co-occurrence pattern data to obtain compensatory behavior pattern data;

[0132] Perform acute and chronic pain pattern differentiation according to the feature co-occurrence pattern data and the compensatory behavior pattern data to obtain pain type determination data;

[0133] Calculate pain intensity score data according to the feature co-occurrence pattern data and the pain location;

[0134] Calculate pain feature association data according to the pain intensity score data and the pain type determination data.

[0135] In the embodiment of the present invention, the operation of extracting the standardized pain feature vector according to the load-bearing ratio distribution map, the limb load-bearing symmetry index, the support time feature data, and the load transfer pattern feature is as follows: Obtain the average load-bearing ratio of each limb from the load-bearing ratio distribution map (for example, the load-bearing ratio of the left hind limb is 40%), and calculate its deviation from the standard average load-bearing ratio of this breed of livestock (for example, the standard average load-bearing ratio of the hind limb is 55%) (40% - 55% = -15%). Obtain the load-bearing difference rate between the ipsilateral limbs (for example, the load-bearing difference rate of the hind limb is 25%) and the asymmetry degree index (for example, the corresponding score is 7) from the limb load-bearing symmetry index. Obtain the proportion of the average support phase duration of each limb in the gait cycle from the support time feature data (for example, the support ratio of the left hind limb is 30%), and calculate its deviation from the average support ratio of the individual gait baseline (for example, the average support ratio of the left hind limb of the individual baseline is 45%) (30% - 45% = -15%). Obtain the identified load transfer pattern type (for example, avoidance type) and its intensity (for example, the intensity score is 8) from the load transfer pattern feature. Perform standardization processing on these original values. For example, use the Min-Max standardization method to standardize the load-bearing ratio deviation and the support ratio deviation to the [-1, 1] interval, and standardize the difference rate, the asymmetry index, and the transfer pattern intensity to the [0, 1] interval. The standardization formula is (original value - minimum value) / (maximum value - minimum value). Combine the standardized load-bearing ratio deviations, standardized support ratio deviations, standardized load-bearing difference rates, standardized asymmetry indexes, and standardized load transfer pattern intensity values of all limbs into a multi-dimensional vector to form a standardized pain feature vector. For example, the vector contains dimensions such as [left front load deviation, right front load deviation, left hind load deviation, right hind load deviation, front limb difference rate, hind limb difference rate, left front support deviation, right front support deviation, left hind support deviation, right hind support deviation, transfer pattern intensity].

[0136] The operation of obtaining livestock breed reference data and determining the characteristic threshold based on the standardized pain feature vector to obtain the pain feature active state table is as follows: Access the pre-established pain feature threshold database, which stores the reference thresholds for each dimension parameter in the standardized pain feature vector for different livestock breeds (e.g., Holstein cows, Simmental cows). These thresholds are obtained through statistical analysis of the gait data of a large number of healthy livestock. For example, the standardized value corresponding to the normal upper limit threshold of the hind limb weight-bearing difference rate of Holstein cows is 0.2, and the standardized value corresponding to the normal lower limit threshold of the deviation of the left hind limb support ratio is -0.3. Compare each parameter value in the currently collected standardized pain feature vector with the reference threshold corresponding to the livestock breed. If a parameter value exceeds its normal range threshold (e.g., the standardized value of the left hind limb weight-bearing deviation is -0.5, which is less than the breed reference lower threshold of -0.3), then determine that this feature is in the "active" state, indicating that this feature shows an abnormal deviation. If the parameter value is within the threshold range, it is determined to be in the "inactive" state. Record the parameters in all standardized pain feature vectors and their corresponding "active" or "inactive" states, for example, using a binary matrix or list. The rows of the matrix represent different pain features, and the columns represent their active states (1 for active, 0 for inactive), forming the pain feature active state table.

[0137] The operation of performing feature co-occurrence pattern analysis based on the pain feature active state table to obtain feature co-occurrence pattern data is as follows: Analyze which features in the pain feature active state table are simultaneously in the "active" state. Use an association rule mining algorithm (such as the Apriori algorithm) or a predefined pattern matching rule library. The predefined pattern matching rule library contains known feature combination patterns related to specific pain locations or types. For example, a pattern is defined as: If the "left hind limb weight-bearing deviation" is active and negative (indicating a decrease in weight-bearing), and at the same time the "left hind limb support ratio deviation" is active and negative (indicating a shortening of the support time), and the "hind limb weight-bearing difference rate" is active and positive (indicating that the left hind limb weight-bearing is lower than the right hind limb), then it is determined that there is a "left hind limb weight reduction pattern". Traverse the pain feature active state table to identify which predefined pattern feature combinations are active in the current observation, or identify which feature sets are frequently active together. Record the identified feature co-occurrence patterns and their matching degrees (e.g., all features in the pattern being active is a complete match, and some features being active is a partial match), forming the feature co-occurrence pattern data, such as a structured data containing [pattern name, involved limb, matching degree, active feature list].

[0138] The operation of pain location analysis based on feature co-occurrence pattern data to obtain a pain location map is as follows: Based on the pattern names and their related limb information identified in the feature co-occurrence pattern data, determine the body parts where pain exists. Each predefined feature co-occurrence pattern is highly associated with one or more specific pain locations. For example, if the "left hind limb load reduction pattern" is identified and the matching degree is high, the pain location is preferentially located in the left hind limb. If the "forelimb asymmetry pattern" is identified and it involves a decrease in the load of the right forelimb, the pain location is preferentially located in the right forelimb. If the pattern involves multiple limbs, such as the "diagonal load reduction pattern" (simultaneous load reduction of the left hind limb and the right forelimb), it indicates a more complex pain cause or compensation. According to the matching degree of the pattern and the strength of the association between the pattern itself and the pain location (preset weight), calculate the confidence score for each potential pain location. Visualize the pain location results, for example, mark the most painful limb or area on a livestock body contour map, and use the shade of color or a numerical value to represent the confidence of the pain location, forming a pain location map.

[0139] The operation of compensatory behavior recognition based on feature co-occurrence pattern data to obtain compensatory behavior pattern data is as follows: In the feature co-occurrence pattern data, in addition to identifying the patterns pointing to the pain location, it is also necessary to identify the features that manifest as abnormal increased load or changed movement patterns of other limbs. These are usually compensatory behaviors adopted by livestock to reduce the burden on the painful limb. For example, if the main pattern points to pain in the left hind limb (left hind limb load reduction, short support time), and at the same time, the feature co-occurrence pattern data shows that the "right hind limb load deviation" is active and positive (indicating an increase in the load of the right hind limb), or the "left forelimb support time deviation" is active and positive (indicating an extension of the left forelimb support time), then it is determined that the increase in the load of the right hind limb and the extension of the left forelimb support time are compensatory behaviors for the pain in the left hind limb. Identify these compensatory features and their activity levels, and associate them with the already located pain location. Record the identified types of compensatory behaviors, the involved limbs, the degree of compensation, and the pain location they are directed at, forming compensatory behavior pattern data, such as a structured data containing [compensatory limb, compensatory type (e.g., increased load, extended support time), compensatory degree score, pain location targeted].

[0140] The operation of differentiating acute and chronic pain patterns based on co-occurrence pattern data of features and compensatory behavior pattern data to obtain pain type determination data is as follows: Combine the currently recognized co-occurrence pattern data of features and compensatory behavior pattern data, and conduct a comparative analysis with the historical gait records of the livestock. Access the historical records of the co-occurrence pattern data of features and compensatory behavior pattern data of the livestock. Analyze whether the currently recognized pattern appears for the first time, or whether its occurrence frequency and intensity have changed significantly compared to the historical records. If a certain pain pattern (such as the left hind limb weight-bearing reduction pattern) and the corresponding compensatory behavior (such as increased weight-bearing on the right hind limb) persist in the historical records and show relatively stable performance, for example, it is detected 80% of the time in the gait collection in the past 30 days, it is determined as a chronic pain pattern. If a strong pain pattern suddenly appears, and the compensatory behavior shows a high degree of variability or incoordination, for example, it has never appeared in the historical records, and a high-intensity left hind limb weight-bearing reduction and jumping transfer pattern is suddenly detected in this collection, it is determined as an acute pain pattern. Record the pain type determination result (acute or chronic) and the basis for the determination (such as pattern occurrence frequency, intensity change rate, compensatory stability) to form pain type determination data.

[0141] The operation of calculating pain intensity score data based on co-occurrence pattern data of features and pain location is as follows: Based on the pattern type, matching degree, and standardized values of the features recognized in the co-occurrence pattern data of features, combined with the pain location result, calculate a quantitative pain intensity score. Set basic pain intensity scores for different co-occurrence patterns of features in advance. For example, the basic score of the "slight asymmetry pattern" is relatively low, and the basic score of the "severe weight-bearing reduction with jumping transfer pattern" is very high. The higher the matching degree of the pattern (i.e., the more active features in the pattern), the higher the final score. At the same time, incorporate the standardized deviation degree of the active features in the pattern into the calculation. For example, the standardized value of the left hind limb weight-bearing deviation is -0.8 (severe deviation), which will contribute more to the pain intensity score than the standardized value of -0.4 (moderate deviation). Use a weighted summation model or a rule-based scoring system to combine the pattern basic score, pattern matching degree, deviation degree of active features, and confidence level of pain location to calculate a pain intensity score on a scale of 0-10, where 0 indicates no pain sign and 10 indicates extreme pain. Record the calculated pain intensity score to form pain intensity score data.

[0142] The operation of calculating pain characteristic correlation data based on pain intensity score data and pain type determination data is as follows: Integrate the pain intensity score data (for example, the pain intensity score is 8.5), pain type determination data (for example, the pain type is chronic), pain location determination result (for example, the pain location is the left hind limb), identified feature co-occurrence pattern (for example, left hind limb load reduction pattern), compensatory behavior pattern (for example, increased load on the right hind limb), etc. calculated in the previous steps. Construct a structured data object that contains all key information related to pain. For example, it can be organized as a JSON object: {"Pain Location": "Left Hind Limb", "Pain Intensity": 8.5, "Pain Type": "Chronic", "Detection Pattern": "Left Hind Limb Load Reduction Pattern", "Compensatory Behavior": {"Limb": "Right Hind Limb", "Type": "Increased Load"}}. This integrated data object comprehensively describes the pain-related physiological behavior characteristics detected in this gait monitoring, that is, the final pain characteristic correlation data is formed.

[0143] Preferably, step S38 includes:

[0144] Evaluate the pain status based on the pain characteristic correlation data and joint range of motion to obtain the limb functional health index;

[0145] Calculate the pain intensity score based on the pain characteristic correlation data and joint range of motion score;

[0146] Evaluate the limb function based on the load ratio distribution map, limb load symmetry index, support time characteristic data and joint range of motion score to obtain the limb function status score;

[0147] Integrate the load ratio distribution map, limb load symmetry index, pain intensity score and limb function status score to generate the limb functional health index.

[0148] In the embodiments of the present invention, the operation of evaluating the pain state based on the pain characteristic correlation data and the joint range of motion score to obtain the limb function health index is as follows: The pain characteristic correlation data (including the pain probability rating of each limb, for example, on a scale of 0 - 10) is associated and analyzed with the joint range of motion score (including the range of motion scores of the main joints of each limb, for example, on a scale of 0 - 100). The pain characteristic correlation data indicates the degree and location of the pain-related gait characteristics of the livestock, while the joint range of motion score directly reflects the degree of limitation of the limb motor ability. If the pain probability rating of a certain limb (such as the left hind limb) is very high (such as 8 points), and the range of motion scores of the joints related to this limb (such as the left hind knee joint, left hind ankle joint) are very low (such as an average of 40 points), this strong pain sign corroborates the significant functional limitation, indicating a high degree of functional impairment of this limb due to pain. A pain impact on function assessment model is constructed. This model takes the pain probability rating and the joint range of motion score as inputs and outputs a pain impact on function score (for example, on a scale of 0 - 100, the higher the score, the greater the impact of pain on function) of this limb due to pain. The model can use a rule-based mapping table or a polynomial function: Pain impact on function score = f(pain probability rating, joint range of motion score). For example, the rule can be defined as: If the pain probability rating > 7 and the joint range of motion score < 50, then the pain impact on function score is 80; if the pain probability rating < 3 and the joint range of motion score > 80, then the pain impact on function score is 20. Apply this model to each limb to calculate its pain impact on function score. These scores constitute the key indicators for evaluating the pain state and are used as the inputs for calculating the limb function health index in the subsequent steps.

[0149] The operation of calculating the pain intensity score based on pain characteristic correlation data and joint range of motion score is as follows: Use the initial pain intensity score (e.g., on a scale of 0 - 10) calculated from the pain characteristic correlation data as a basis, and adjust and refine it in combination with the joint range of motion score (on a scale of 0 - 100). The pain characteristic correlation data mainly depends on the abnormal patterns of the load-bearing and time parameters, while the joint range of motion score provides kinematic evidence. If the initial pain intensity score of a certain limb (e.g., the left hind limb) is relatively high (e.g., 7 points), but the joint range of motion score of its related joint shows that the functional limitation is not severe (e.g., an average of 70 points), this indicates that the pain does not completely restrict movement, or the source of pain is not the joint itself. On the contrary, if the initial pain intensity score is moderate (e.g., 5 points), but the joint range of motion score is extremely low (e.g., an average of 30 points), this strongly suggests the existence of significant dysfunction, that the pain intensity is underestimated, or the nature of the pain is more focused on movement limitation. Establish a pain intensity correction model that receives the initial pain intensity score and the joint range of motion score of the related limb as inputs, and outputs a corrected pain intensity score (still on a scale of 0 - 10) that more accurately reflects the actual pain level. The model can adopt an adjustment factor: Corrected pain intensity score = Initial pain intensity score × Correction factor (joint range of motion score). The correction factor function can be designed as follows: when the joint range of motion score is very low, the correction factor is slightly greater than 1 (e.g., 1.1); when the joint range of motion score is very high, the correction factor is slightly less than 1 (e.g., 0.9); when the joint range of motion score is moderate, the correction factor is equal to 1. Calculate the corrected pain intensity score for each limb, and these scores are the pain intensity score data used for the final health assessment.

[0150] The operation of performing limb function assessment based on the load-bearing ratio distribution graph, limb load-bearing symmetry index, support time characteristic data, and joint range of motion score to obtain the limb function status score is as follows: Collect the load-bearing ratio distribution graph (including the average load-bearing ratio of each limb), the limb load-bearing symmetry index (including the load-bearing difference rate and asymmetry degree index of the ipsilateral limbs), the support time characteristic data (including the support time ratio of each limb and its deviation from the baseline), and the joint range of motion score. These data directly reflect the objective capabilities of the limb in terms of support, movement, and load-bearing. For example, if the load-bearing ratio of the left hind limb is significantly lower than normal, the load-bearing difference rate of the hind limbs is high and the left hind limb is the load-reducing side, the support time ratio of the left hind limb is significantly shortened, and the joint range of motion score of the left hind limb is low, all these indicate that the function of the left hind limb is impaired. Define a function status score model for each limb, which is a multi-input and single-output evaluation function. The inputs include the average load-bearing ratio of the limb, the load-bearing difference rate compared with the contralateral limb (if applicable), the deviation of the support time ratio, and the average range of motion score of the related joint. The model uses the weighted summation method: Limb function status score = + + + Among them, is a preset weight, reflecting the importance of each parameter to the function (for example, the weights of joint range of motion and load-bearing ratio are relatively high); is a function that maps the original parameters to the 0-100 score range (for example, the lower the load-bearing ratio, the lower the score, the higher the difference rate, the lower the score, the greater the deviation, the lower the score, and the range of motion score is directly used). Calculate the functional status score of each limb (on a 0-100 scale, with 100 being normal function), and these scores constitute the limb functional status score data.

[0151] The operation of integrating the load-bearing ratio distribution map, limb load-bearing symmetry index, pain intensity score, and limb functional status score to generate the limb functional health index is as follows: gather the load-bearing ratio distribution map, limb load-bearing symmetry index, pain intensity score data, and limb functional status score data. The limb functional health index aims to provide a single indicator for comprehensively measuring the health status of the limb, which needs to consider both the objective functional status (load-bearing, motor ability) and subjective feelings (pain level) of the limb at the same time. Construct a limb functional health index calculation model, which takes the limb functional status score and pain intensity score as the main inputs and is fine-tuned by combining key original indicators such as load-bearing ratio and symmetry. The model can adopt the following structure: limb functional health index = g(limb functional status score, pain intensity score, load-bearing ratio, load-bearing symmetry). The function g is designed as: the higher the functional status score, the higher the health index; the higher the pain intensity score, the lower the health index; a significant deviation in the load-bearing ratio or poor symmetry will further reduce the health index. For example, using the method of weighted combination and penalty terms: limb functional health index = - - - . is the weight coefficient. The load-bearing ratio deviation penalty can be calculated based on the deviation of the load-bearing ratio from the normal range, and the greater the deviation, the greater the penalty. The symmetry penalty can be calculated based on the load-bearing difference rate, and the higher the difference rate, the greater the penalty. Finally, calculate the limb functional health index of each limb (left front, right front, left rear, right rear) (for example, on a 0-100 scale, with 100 being completely healthy), and the set of these indexes forms the final limb functional health index data. This index intuitively reflects the current health level of each limb and provides basic data for subsequent monitoring of the overall health trend.

[0152] Preferably, the present invention also provides a livestock physiological behavior monitoring system for performing the livestock physiological behavior monitoring method described above. The livestock physiological behavior monitoring system includes:

[0153] The gait data acquisition module is used to collect a three - area pressure matrix through a pressure - sensitive pad; perform dynamic path recognition and camera adjustment based on the three - area pressure matrix to obtain dynamic path tracking data; perform livestock mechanical gait analysis based on the dynamic path tracking data to obtain a gait mechanical feature map;

[0154] The abnormal feature recognition module is used to obtain livestock basic information; perform individual gait time - series segmentation based on the gait mechanical feature map to obtain segmented and aligned gait data; calculate an adaptive gait threshold based on the livestock basic information and the segmented and aligned gait data to obtain an individualized abnormal threshold set; perform abnormal gait pattern recognition based on the individualized abnormal threshold set to obtain a gait deviation index matrix;

[0155] The pain status assessment module is used to perform limb weight - bearing ratio difference analysis based on the gait deviation index matrix and the gait mechanical feature map to obtain limb weight - bearing characteristics; identify the characteristics of the weight - transfer pattern based on the limb weight - bearing characteristics; perform pain behavior feature recognition based on the characteristics of the weight - transfer pattern to obtain pain feature correlation data; perform limb health assessment based on the pain feature correlation data to obtain a limb function health index;

[0156] The health trend monitoring module is used to perform time - series monitoring of the behavior pattern based on the limb function health index to obtain a livestock health trend map.

[0157] Therefore, from any perspective, the embodiments should be regarded as exemplary and non - restrictive. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the application documents are intended to be encompassed within the present invention.

[0158] The above - mentioned are only the specific embodiments of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but rather to the broadest scope consistent with the principles and novel features disclosed herein.

Claims

1. A livestock physiological behavior monitoring system, characterized in that, It includes the following modules: A gait data acquisition module, which is used to collect a three-region pressure matrix through a pressure sensing pad; perform dynamic path recognition and camera adjustment based on the three-region pressure matrix to obtain dynamic path tracking data; perform livestock mechanical gait analysis based on the dynamic path tracking data to obtain a gait mechanical feature map; An abnormal feature recognition module, which is used to obtain livestock basic information; perform individual gait time series segmentation based on the gait mechanical feature map to obtain segmented and aligned gait data; Calculate an adaptive gait threshold based on the livestock basic information and the segmented and aligned gait data to obtain an individualized abnormal threshold set; Perform abnormal gait pattern recognition based on the individualized abnormal threshold set to obtain a gait deviation index matrix; A pain status assessment module, which is used to analyze the difference in limb weight-bearing ratio based on the gait deviation index matrix and the gait mechanical feature map to obtain limb weight-bearing characteristics; recognize the characteristics of the weight transfer pattern based on the limb weight-bearing characteristics; recognize the pain behavior characteristics based on the characteristics of the weight transfer pattern to obtain pain feature correlation data; Perform limb health assessment based on the pain feature correlation data to obtain a limb function health index; A health trend monitoring module, which is used to perform time series monitoring of the behavior pattern based on the limb function health index to obtain a livestock health trend map.

2. The livestock physiological behavior monitoring system according to claim 1, characterized in that, The gait data acquisition module includes the following steps: Step S11: Divide the pressure sensing pad into three functional regions, namely the front region, the middle region, and the rear region, according to the force characteristics of the livestock hoof. Each region is configured with an independent pressure sensor array, and a three-region pressure matrix is collected through the pressure sensor array; Step S12: Set 6 high-speed cameras around the pressure sensing pad, and calibrate the spatial coordinates of the camera system through a calibration board to obtain a spatial coordinate mapping data set; Step S13: Set livestock joint marking points for the livestock to be measured to obtain joint marking position data; Step S14: Perform synchronous data acquisition triggering based on the three-region pressure matrix, the spatial coordinate mapping data set, and the joint marking position data to obtain a synchronous trigger time series; Step S15: Perform dynamic path recognition and camera adjustment based on the three-region pressure matrix and the synchronous trigger time series to obtain dynamic path tracking data; Step S16: Extract ground reaction force data from the three-region pressure matrix based on the synchronous trigger time series and the dynamic path tracking data to obtain mechanical distribution time series data; Step S17: Calculate a joint motion parameter set based on the spatial coordinate mapping data set, the joint marking position data, and the synchronous trigger time series; Step S18: Perform gait cycle division based on the mechanical distribution time series data and the joint motion parameter set to obtain gait cycle segmented data; Step S19: Generate a gait mechanical feature map based on the mechanical distribution time series data, the joint motion parameter set, the gait cycle segmented data, and the dynamic path tracking data.

3. The livestock physiological behavior monitoring system according to claim 2, wherein Step S15 includes: Extract hoof print spatio-temporal sequence data from the three-region pressure matrix based on the synchronous trigger time series; Calculate the path direction vector based on the hoof print spatio-temporal sequence data to obtain livestock motion vector data; Calculate path prediction curve data based on the livestock motion vector data and the hoof print spatio-temporal sequence data; Identify turning points based on livestock movement vector data and spatio-temporal sequence data of hoof prints to obtain behavior mutation marker data; Calculate the optimal observation angle based on the path prediction curve data and behavior mutation marker data to obtain camera angle optimization parameters; Collect dynamic path tracking data according to the camera angle optimization parameters.

4. The livestock physiological behavior monitoring system according to claim 1, characterized in that, The individual gait time series segmentation in the abnormal feature recognition module includes: Obtain livestock individual recognition information and construct and access a historical database based on the gait mechanics feature map to obtain an individual historical gait record set; Calculate the individual gait baseline parameter set of the individual historical gait record set; Standardize the current gait parameters for each parameter in the gait mechanics feature map to obtain standardized gait parameters; Perform time series segmentation and alignment on the standardized gait parameters and the individual gait baseline parameter set to obtain segmented and aligned gait data.

5. The livestock physiological behavior monitoring system according to claim 1, characterized in that The adaptive gait threshold calculation in the abnormal feature recognition module includes: Extract livestock feature standardization parameters based on livestock basic information and the individual gait baseline parameter set; Calculate the gait stability index based on the segmented and aligned gait data to obtain a gait fluctuation feature spectrum; Allocate parameter importance weights according to the gait fluctuation feature spectrum and livestock feature standardization parameters to obtain a parameter importance weight table; Identify individual-specific gait features based on the gait fluctuation feature spectrum; Calculate the joint angle abnormality threshold based on the individual-specific gait features; Calculate the mechanical parameter abnormality threshold based on the individual-specific gait features; Calculate the time parameter abnormality threshold based on the gait fluctuation feature spectrum; Integrate the joint angle abnormality threshold, mechanical parameter abnormality threshold, and time parameter abnormality threshold to obtain an individualized abnormality threshold set.

6. The livestock physiological behavior monitoring system according to claim 1, characterized in that The abnormal gait pattern recognition in the abnormal feature recognition module includes: Perform abnormal detection of joint angle changes on the segmented and aligned gait data according to the individualized abnormality threshold set to obtain joint angle abnormality marker data; Perform abnormal detection of ground reaction force on the segmented and aligned gait data according to the individualized abnormality threshold set to obtain mechanical parameter abnormality marker data; Perform lateral movement amplitude analysis on the segmented and aligned gait data according to the individualized abnormality threshold set to obtain lateral movement abnormality marker data; Perform time correlation analysis on the joint angle abnormality marker data, mechanical parameter abnormality marker data, and lateral movement abnormality marker data to obtain abnormal pattern time series correlation data; Generate a gait deviation index matrix based on the joint angle abnormality marker data, mechanical parameter abnormality marker data, lateral movement abnormality marker data, and abnormal pattern time series correlation data.

7. The livestock physiological behavior monitoring system according to claim 1, characterized in that, The pain status assessment module includes the following steps: Step S31: Extract limb load data from the gait deviation index matrix and the gait mechanics feature map to obtain raw limb load data; Step S32: Calculate a load ratio distribution map based on the raw limb load data and the weight data in the livestock basic information; Step S33: Calculate the load difference rate between the left front limb and the right front limb, and between the left hind limb and the right hind limb based on the load ratio distribution map. Mark the load difference rate exceeding 15% as asymmetric load. Then calculate the asymmetry degree index, and finally obtain the limb load symmetry index; Step S34: Perform weight-bearing time analysis based on the gait mechanics feature map and the limb weight-bearing symmetry index to obtain support time feature data; Step S35: Identify the weight-bearing transfer pattern features based on the weight-bearing ratio distribution map, the limb weight-bearing symmetry index, and the support time feature data; Step S36: Evaluate the joint range of motion of the gait mechanics feature map according to the gait deviation index matrix to obtain the joint range of motion score; Step S37: Identify the pain behavior features based on the weight-bearing ratio distribution map, the limb weight-bearing symmetry index, the support time feature data, and the weight-bearing transfer pattern features to obtain pain feature correlation data; Step S38: Perform limb health assessment based on the pain feature correlation data to obtain the limb functional health index.

8. The livestock physiological behavior monitoring system according to claim 7, wherein Step S37 includes: Extract the standardized pain feature vector based on the weight-bearing ratio distribution map, the limb weight-bearing symmetry index, the support time feature data, and the weight-bearing transfer pattern features; Obtain the livestock breed reference data and perform feature threshold determination according to the standardized pain feature vector to obtain the pain feature active status table; Perform feature co-occurrence pattern analysis according to the pain feature active status table to obtain the feature co-occurrence pattern data; Perform pain location analysis according to the feature co-occurrence pattern data to obtain the pain location map; Identify the compensatory behavior according to the feature co-occurrence pattern data to obtain the compensatory behavior pattern data; Distinguish between acute and chronic pain patterns according to the feature co-occurrence pattern data and the compensatory behavior pattern data to obtain the pain type determination data; Calculate the pain intensity score data according to the feature co-occurrence pattern data and the pain location; Calculate the pain feature correlation data according to the pain intensity score data and the pain type determination data.

9. The livestock physiological behavior monitoring system according to claim 7, characterized in that Step S38 includes: Perform pain status assessment according to the pain feature correlation data and the joint range of motion to obtain the limb functional health index; Calculate the pain intensity score according to the pain feature correlation data and the joint range of motion score; Perform limb function assessment according to the weight-bearing ratio distribution map, the limb weight-bearing symmetry index, the support time feature data, and the joint range of motion score to obtain the limb function status score; Integrate the weight-bearing ratio distribution map, the limb weight-bearing symmetry index, the pain intensity score, and the limb function status score to generate the limb functional health index.

Images