An intelligent sow health early warning system and method based on ear canal core temperature monitoring

Abstract

The application provides a kind of intelligent sow health early warning system and method based on ear canal core temperature monitoring, comprising: wearable ear tag, data analysis module and health early warning module;Wearable ear tag is used to collect the core temperature data of sow ear canal;Data analysis module is used to receive core temperature data, and the time series analysis is carried out on core temperature data, and sow health assessment model is constructed;Health early warning module is used to lock the health abnormal sow and carry out abnormal early warning based on the evaluation result of sow health assessment model using preset early warning rule.The technical scheme of the application realizes the real-time and accurate monitoring of sow health condition, effectively improves the breeding management efficiency and disease prevention ability.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of intelligent breeding and health monitoring, and particularly relates to an intelligent sow health early warning system and method based on ear canal core temperature monitoring. BACKGROUND

[0002] The existing ear tags mostly adopt infrared temperature measurement technology, but the ear canal environment is complex and is easily disturbed by external temperature, resulting in insufficient precision (±0.5℃ error) and inability to obtain core temperature data. Artificial inspection relies on handheld temperature measurement equipment, which is time-consuming and has a high rate of missed detection, and cannot realize 24-hour continuous monitoring. The existing systems mostly warn based on a single temperature threshold, do not combine multidimensional data such as exercise amount, and have a false alarm rate of more than 30%. Disease prediction relies on artificial experience and lacks dynamic time series analysis capability, making it difficult to discover early sub-health states in a timely manner. The ear tags are heavy (some products exceed 15g) and have poor wearing comfort, resulting in frequent shaking and falling off of the sows.

[0003] Ear canal core temperature, as an important physiological parameter, can sensitively reflect the health status of sows (such as estrus, infection, stress response, etc.), but there is currently a lack of an intelligent sow health early warning system and method based on ear canal core temperature monitoring. SUMMARY

[0004] The application provides an intelligent sow health early warning system and method based on ear canal core temperature monitoring, which deeply integrates precise sensing technology, biological feature modeling and artificial intelligence algorithms, and solves key problems such as low monitoring precision, delayed early warning and poor device reliability in traditional breeding.

[0005] An intelligent sow health early warning system based on ear canal core temperature monitoring, comprising: a wearable ear tag, a data analysis module and a health early warning module.

[0006] The wearable ear tag is used to collect core temperature data of the sow's ear canal.

[0007] The data analysis module is used to receive the core temperature data and perform time series analysis on the core temperature data to construct a sow health evaluation model.

[0008] The health early warning module is used to lock and perform abnormal early warning on sows with health abnormalities based on the evaluation results of the sow health evaluation model and using preset early warning rules.

[0009] Preferably, the wearable ear tag comprises:

[0010] A temperature data collection module is used to collect core temperature data of the sow's ear canal based on a temperature sensor.

[0011] a control module, configured to perform primary filtering and dynamic temperature compensation on the core temperature data, and obtain compensated core temperature data;

[0012] a Bluetooth communication module, configured to transmit the compensated core temperature data;

[0013] a power supply module, configured to supply power to the temperature sensor based on a button cell.

[0014] Preferably, the process of constructing the temperature sensor comprises:

[0015] obtaining anatomical structure data of the sow ear canal at different growth stages and physiological states based on three-dimensional scanning technology, and constructing an ear canal three-dimensional model;

[0016] integrating a miniature thermistor on a flexible substrate, and wrapping it from inside to outside with a plastic material and skin-friendly silicone to obtain an initial temperature sensor;

[0017] adjusting the shape of the initial temperature sensor based on the ear canal three-dimensional model and the ear canal variation range to obtain the final temperature sensor.

[0018] Preferably, the data analysis module comprises:

[0019] a data receiving unit, configured to receive the core temperature data;

[0020] a multi-scale window analysis unit, configured to capture short-term fluctuation features of the core temperature data based on a preset time scale;

[0021] a long short-term memory network unit, configured to capture long-term dependence features of the core temperature data based on an LSTM extended architecture;

[0022] a feature classification unit, configured to extract key features of the short-term fluctuation features and the long-term dependence features, classify the key features by using an improved random forest algorithm, and obtain a classification result;

[0023] an evaluation model construction unit, configured to perform sow health grade evaluation based on the classification result, and complete construction of the sow health evaluation model.

[0024] Preferably, the process of improving the random forest algorithm comprises:

[0025] training a random forest network by using an existing feature set, and obtaining an average value of node purity reduction of each feature in all decision trees for splitting nodes during a training process;

[0026] calculating a feature importance score based on the average value of node purity reduction;

[0027] normalizing the feature importance scores to obtain feature sampling weights;

[0028] sampling the features to be sampled according to the feature sampling weights to obtain a weighted random sampling feature subset when each node is split;

[0029] selecting a feature with maximum information gain in the feature subset as a split point to complete the improvement of the random forest algorithm.

[0030] Preferably, the health warning module comprises:

[0031] an evaluation result tracking unit configured to locate a health abnormal sow based on the evaluation result and obtain an abnormal duration and an abnormal degree;

[0032] a warning unit configured to perform hierarchical warning based on the abnormal duration and the abnormal degree.

[0033] The application also provides an intelligent sow health warning method based on ear canal core temperature monitoring, which is used to realize the system and comprises the following steps:

[0034] collecting core temperature data of the sow's ear canal;

[0035] receiving the core temperature data and performing time series analysis on the core temperature data to construct a sow health evaluation model;

[0036] based on the evaluation result of the sow health evaluation model, adopting a preset warning rule to lock a health abnormal sow and perform abnormal warning.

[0037] Preferably, the method for constructing the sow health evaluation model comprises:

[0038] receiving the core temperature data;

[0039] capturing short-term fluctuation characteristics of the core temperature data based on a preset time scale;

[0040] capturing long-term dependence characteristics of the core temperature data based on an LSTM extended architecture;

[0041] extracting key features of the short-term fluctuation characteristics and the long-term dependence characteristics, classifying the key features by using an improved random forest algorithm to obtain a classification result;

[0042] based on the classification result, performing sow health grade evaluation to complete the construction of the sow health evaluation model.

[0043] Compared with the prior art, the beneficial effects of the present application are: by directly collecting the core temperature data of the sow's ear canal through the wearable ear tag, the real body temperature condition of the sow can be better reflected than the traditional body surface temperature measurement, and the accuracy of temperature measurement is improved. The system can monitor the sow's body temperature for 24 hours without interruption, timely detect abnormal body temperature, and avoid delaying the treatment opportunity. The data analysis module performs time series analysis on the core temperature data, can capture the subtle changes of the sow's body temperature, and construct a more accurate health assessment model. The health assessment model can more comprehensively evaluate the sow's health condition, reduce false positives and false negatives, and improve the reliability of early warning. The system is an important part of intelligent management of breeding industry, and can be linked with other intelligent devices (such as intelligent feeding system, environment control system) to realize comprehensive intelligent management of breeding farm. BRIEF DESCRIPTION OF DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the present application, the following briefly introduces the drawings needed to be used in the embodiments of the present application. Obviously, the drawings described in the following embodiments are only some embodiments of the present application, and all other drawings obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0045] Figure 1 The structure diagram of the intelligent sow health early warning system based on ear canal core temperature monitoring of the embodiment of the present application is shown.

[0046] Figure 2 The curve graph of the core temperature of sow's ear canal changing with time of the embodiment of the present application is shown. DETAILED DESCRIPTION

[0047] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0048] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0049] Embodiment one

[0050] As shown in Figure 1 , Figure 2 A kind of intelligent sow health early warning system based on ear canal core temperature monitoring, including: wearable ear tag, data analysis module and health early warning module.

[0051] Wearable ear tag is used to collect the core temperature data of sow's ear canal.

[0052] Further embodiments are that the wearable ear tag comprises a temperature data acquisition module, a control module, a Bluetooth communication module, and a power supply module.

[0053] The temperature data acquisition module is configured to acquire the core temperature data of the sow's ear canal based on a temperature sensor. In this embodiment, further embodiments are that the process of constructing the temperature sensor comprises:

[0054] Based on three-dimensional scanning technology, anatomical structure data of the sow's ear canal in different growth stages and physiological states is obtained to construct a three-dimensional model of the ear canal. Specifically, a structured light scanner (accuracy ±0.05 mm) and a thermal imager are used to collect data synchronously, solving the problem of motion artifacts in living bodies. The structured light scanning generates point cloud data of the sow's ear canal through grating projection and phase decoding. That is, a projector projects a coded stripe pattern onto the ear canal, and two high-resolution cameras (binocular system) capture the stripes deformed by the ear canal surface. The absolute phase value is calculated through the phase shift method to establish a pixel-three-dimensional coordinate mapping:

[0055]

[0056] (I0-I3) are light intensity images collected by four-step phase shift method.

[0057] Reflective marker points (diameter 2 mm) are pasted on the ear, and a motion trajectory equation is constructed by 6 high-speed infrared cameras (1000 fps).

[0058] Then, the improved time series ICP algorithm is used to align the point cloud data in different growth stages:

[0059]

[0060] where R t represents the rotation matrix (3x3 orthogonal matrix) at time t, t t represents the translation vector (3x1 vector) at time t, represents the coordinates of the i-th point at time t (3D vector), represents the coordinates of the corresponding point at time t+Δt, and λ represents the regularization coefficient, which controls the smoothness of the rigid body transformation and avoids sudden changes in the model between adjacent stages. ||·|| F is the Frobenius norm (square root of the sum of the squares of the matrix elements).

[0061] The differential geometric properties of the point cloud are calculated by the moving least squares method (MLS):

[0062] κ combfeature points, keeping key anatomical structures such as the helix of the ear canal, blood vessel depressions, etc. Among them, κ1 represents the first principal curvature, κ2 represents the second principal curvature, κ comb is the combined curvature, and the threshold judgment for feature preservation.

[0063] Motion compensation is achieved using the motion estimation equation and the IPC algorithm, and feature extraction is performed using the least squares method.

[0064] After obtaining the point cloud data of the sow ear canal at different growth stages through three-dimensional scanning, a bias optimization algorithm is used to solidify the modeling of the point cloud. This strategy dynamically adjusts the offset between the scanned data and the standard ear canal model, eliminating modeling errors caused by individual morphological differences, ensuring that the model can reflect the commonality of the group and retain individual characteristics. Specifically, an implicit surface function F(x) is defined, combined with a radial basis function (RBF) and growth constraints:

[0065]

[0066] φ(r) = r 3 represents a cubic radial basis function, P(x) represents a quadratic polynomial to compensate for global shape, G(t, x) = α(t) · ||x|| represents the time growth term, α(t) · ||x||, and α(t) is the time-dependent growth rate. i represents the weight coefficient of the i-th control point.

[0067] An energy minimization objective function is established:

[0068]

[0069] d k represents the signed distance from the point cloud k to the surface (external d > 0, internal d < 0), represents the gradient vector of the implicit function, n represents the point cloud normal field, β represents the smoothing term weight, and γ represents the distance constraint weight. The normal field-guided optimization algorithm improves the surface smoothness by 40%.

[0070] Discrete solution is obtained by finite element method, with adaptive mesh size (0.1mm in narrow ear canal area, 0.5mm in periphery), completing solidification modeling and obtaining sow ear canal three-dimensional model.

[0071] A parameterized growth tensor is used to establish growth dynamic constraints for the sow ear canal three-dimensional model. Specifically, the ear canal growth velocity field is defined as:

[0072]

[0073] g radial (t) represents the radial growth rate, g axial (t) represents the axial growth rate (along the depth direction of the ear canal).spiral (t) represents the spiral growth rate (controls the cochlea morphological evolution), e r , e θ , e z is the unit basis vector of the cylindrical coordinate system.

[0074] And based on the sow growth rate, the network topology is dynamically adjusted to complete the growth dynamic constraint of the sow ear canal three-dimensional model.

[0075] Finally output the sow ear canal three-dimensional entity model:

[0076] V k represents the vertex set, F k represents the face set, A k represents the attribute set (curvature, thickness and growth stage label).

[0077] The ear canal three-dimensional entity model of pigs in different physiological states such as pregnancy and lactation is clustered and analyzed, and key parameters such as ear canal volume and curvature radius are extracted to establish an ear canal morphological variation database.

[0078] Based on the genetic algorithm, the "safe threshold range" of the ear canal model is optimized to determine the boundary conditions of the sensor shape adjustment, avoiding fitting failure caused by sudden changes in ear canal size.

[0079] Integrate the miniature thermistor on the flexible substrate, and wrap it with plastic material and skin-friendly silicone from the inside to the outside to obtain the initial temperature sensor. In this embodiment, polyurethane foam and conductive polymer (such as PEDOT:PSS) composite substrate are used, and porous microstructure is constructed by 3D printing technology, which not only realizes high-precision embedding of thermistor, but also improves the ductility (stretchability up to 200%) and thermal conductivity efficiency of the substrate. Add nano-silver antibacterial coating to the outer layer of silicone, and design surface microtexture (imitate sharkskin structure) to reduce ear canal friction and allergy risk.

[0080] Based on the ear canal three-dimensional model and the ear canal variation range, the initial temperature sensor is adjusted in shape to obtain the final temperature sensor. In this embodiment, based on the sensor stiffness matrix and the sow ear canal soft tissue stiffness matrix, a contact mechanics model is constructed to solve the contact pressure distribution and perform virtual wearing simulation.

[0081] The control module is used for primary filtering and dynamic temperature compensation of the core temperature data to obtain compensated core temperature data; the built-in low-power processing unit is used for real-time processing and preliminary filtering of the collected temperature data. In this embodiment, based on the sow ear blood flow metabolic heat generation rate and the thermal expansion coupling coefficient, a heat conduction-deformation coupling equation is constructed to complete the dynamic temperature compensation of the core temperature data.

[0082] The Bluetooth communication module is used for transmitting the compensated core temperature data. Specifically, the compressed temperature data is segmented and divided by using a preset wireless communication protocol to generate a plurality of data segments. A check value of each data segment is calculated by using a redundancy check mechanism, and the check value is attached to the end of the data segment. In the transmission process, the signal interference strength is monitored in real time, and if the interference strength is higher than a preset threshold, a data retransmission mechanism is triggered. According to the packet loss information fed back by the receiving end, the lost data segment is located, and the lost data segment is retransmitted from the source end. An error correction algorithm is used to verify the check value of the received data segment, and if the check value does not match, an error correction process is triggered. The error part in the data segment is repaired through the error correction process to generate a complete and accurate data segment. The repaired data segment is recombined to restore the complete compressed temperature data.

[0083] The power supply module is used for supplying power to the temperature sensor based on the button cell.

[0084] The data analysis module is used for receiving the core temperature data and performing time series analysis on the core temperature data to construct a sow health assessment model. In a further embodiment, the data analysis module comprises:

[0085] The data receiving unit is used for receiving the core temperature data. Specifically, the uploaded temperature data is received by using a preset distributed storage architecture to obtain the original temperature data. The original temperature data is classified according to the sow individual identifier to generate classified temperature data sets. Independent data storage units are constructed for the classified temperature data sets to generate storage unit identifiers. The classified temperature data is stored in the corresponding independent storage unit through the storage unit identifier. If data loss occurs during storage, the lost data is reacquired from the source end according to the storage unit identifier. A preset data check algorithm is used to check the integrity of the temperature data in the storage unit. According to the check result, if the data is complete, the storage is completed, and if the data is not complete, a data repair process is triggered.

[0086] The multi-scale window analysis unit is used for capturing short-term fluctuation features of the core temperature data based on a preset time scale. In this embodiment, a plurality of time scale windows are preset to ensure 50% overlap between the windows. The window size is adaptively adjusted according to the data sampling frequency and signal characteristics. The short-term fluctuation features include time domain features, frequency domain features and nonlinear features. The time domain features include calculating the mean, standard deviation, skewness and kurtosis in each window. The frequency domain features include performing fast Fourier transform on the core temperature data in the window to obtain the main frequency and spectral entropy. The nonlinear features include sample entropy for measuring the complexity of time series and Lyapunov exponent for evaluating chaotic characteristics. The features of different windows are spliced into a high-dimensional vector, and the high-dimensional features are mapped to a low-dimensional space by using t-SNE to preserve the local structure information.

[0087] A long short-term memory network unit is configured to capture long-term dependency features of the core temperature data based on an LSTM extended architecture. In this embodiment, an exponential gating (sLSTM) and a new storage mixing technique are introduced into the LSTM, allowing the model to revise storage decisions and enhance control over information flow. The memory unit is extended from a scalar to a matrix (mLSTM), increasing storage capacity and introducing a covariance update rule to support parallel processing. The sLSTM and the mLSTM are integrated into a residual block, and an xLSTM architecture is constructed by stacking to improve the nonlinear expression capability of the model.

[0088] Specifically, the sLSTM and the mLSTM are integrated into the residual block to alleviate the gradient vanishing problem:

[0089] H t = LayerNorm(H t-1 + sLSTM(H t-1 ) + mLSTM(H t-1 )),

[0090] LayerNorm: layer normalization, H t represents a hidden state matrix.

[0091] An xLSTM network is constructed by stacking multiple residual blocks:

[0092] where L is the number of residual blocks (L = 6 by default).

[0093] The constructed xLSTM network is used for long-term dependency feature extraction:

[0094] A time convolution module TCN is added before the xLSTM to extract short-term local features, and a dilated convolution is used to expand the receptive field. A multi-head self-attention is introduced at each layer of the xLSTM, and the attention outputs of different layers are weighted and fused to obtain a weighted fusion feature. The long-term dependency of the weighted fusion feature is captured by the memory matrix of the mLSTM, and the global dependency relationship of the time series is modeled by the covariance matrix, completing the extraction of the long-term dependency feature.

[0095] A feature classification unit is configured to extract key features of short-term fluctuation features and long-term dependency features, and to classify the key features using an improved random forest algorithm to obtain a classification result. In this embodiment, a further implementation is that the process of improving the random forest algorithm includes:

[0096] A random forest network is trained using an existing feature set to obtain an average value of node purity reduction of each feature in all decision trees when used for splitting nodes during the training process;

[0097] Based on the average value of the node purity reduction, a feature importance score is calculated;

[0098] The feature importance scores are normalized to obtain feature sampling weights;

[0099] When each node is split, the features to be sampled are sampled according to the feature sampling weights to obtain a weighted random sampling feature subset;

[0100] In the feature subset, the feature with the maximum information gain is selected as the split point, and the improvement of the random forest algorithm is completed.

[0101] The evaluation model construction unit is configured to evaluate the sow health based on the classification result, and complete the construction of the sow health evaluation model.

[0102] The health warning module is configured to lock the sow with abnormal health based on the evaluation result of the sow health evaluation model, and perform abnormal warning by using a preset warning rule.

[0103] Further embodiments are that the health warning module comprises:

[0104] The evaluation result tracking unit is configured to locate the sow with abnormal health based on the evaluation result, and obtain the abnormal duration and the abnormal degree; specifically, the length of continuous abnormality is counted, the start and end time [t start , t end ] of each abnormal segment is recorded, and the duration is calculated. Based on the deviation of the evaluation result from the preset health threshold, the abnormal intensity is calculated, and each abnormal segment is comprehensively scored:

[0105]

[0106] Wherein τ is a decay coefficient (default τ = 10) for weighting recent abnormalities.

[0107] The warning unit is configured to perform hierarchical warning based on the abnormal duration and the abnormal degree.

[0108] Specifically, according to the abnormal duration Δt and the abnormal degree I total , three levels of warning are defined:

[0109] Low-risk warning: Δt ∈ [T1, T2) and I total <I1,

[0110] Medium-risk warning: Δt ∈ [T2, T3) or I total ∈ [I1, I2),

[0111] High-risk warning: Δt ≥ T3 or I toal ≥ I2.

[0112] The duration range T1, T2, T3 can be customized by the user.

[0113] Early warning content includes: 1. Sow ID; 2. Abnormal start time 3. Abnormal duration; 4. Abnormal degree score; 5. Early warning level.

[0114] Push early warning information in the following ways:

[0115] Mobile terminal push: real-time sending to breeding management APP;

[0116] SMS notification: high-risk early warning is sent to the administrator's mobile phone;

[0117] Acoustic light alarm: trigger acoustic light alarm equipment in the farm.

[0118] The model of the application continuously learns individual temperature data, and dynamically updates the benchmark value and the health threshold.

[0119] The system provides a manual marking function (confirming actual events such as infection, delivery, etc.), and feeds back to optimize the model.

[0120] Periodically optimize the algorithm using all individual data to ensure that the group applicability and individual accuracy are considered.

[0121] The application can accurately identify the estrus period, pre-delivery state and disease risk of sows, and provide early warning signals to breeding management personnel in a timely manner, optimize production management. Support docking with existing breeding management systems to form a health management closed loop. Data is stored in the cloud for a long time, supports multi-dimensional health report generation, and provides scientific basis for breeding optimization.

[0122] Example two

[0123] The application also provides an intelligent sow health early warning method based on ear canal core temperature monitoring, which is used to realize a system, comprising:

[0124] Collecting core temperature data of sow ear canal;

[0125] Receiving core temperature data and performing time series analysis on core temperature data to build a sow health assessment model;

[0126] Based on the evaluation results of the sow health assessment model, a preset early warning rule is used to lock the health abnormal sow and perform abnormal early warning.

[0127] Further embodiments are that the method for building a sow health assessment model comprises:

[0128] Receiving core temperature data;

[0129] Capturing short-term fluctuation characteristics of core temperature data based on a preset time scale;

[0130] Capture long-term dependence characteristics of core temperature data based on LSTM extended architecture;

[0131] Extract the key features of short-term fluctuation characteristics and long-term dependence characteristics, classify the key features by using an improved random forest algorithm, and obtain a classification result;

[0132] Based on the classification result, the sow health grade is evaluated, and the sow health evaluation model is constructed.

[0133] The above-described embodiments are only descriptions of the preferred modes of the present application and do not limit the scope of the present application. Without departing from the design spirit of the present application, various modifications and improvements to the technical solutions of the present application made by those skilled in the art shall fall within the protection scope determined by the claims of the present application.

Claims

1. An intelligent sow health early warning system based on ear canal core temperature monitoring, characterized in that: include: Wearable ear tags, data analysis modules, and health warning modules; The wearable ear tag is used to collect core temperature data of the sow's ear canal; The data analysis module is configured to receive the core temperature data, perform time series analysis on the core temperature data, and construct a sow health assessment model; The health warning module is used to identify abnormally healthy sows and issue abnormal warnings based on the evaluation results of the sow health evaluation model and using preset warning rules; The wearable ear tag comprises: Temperature data acquisition module, used to collect sow ear canal core temperature data based on temperature sensors; a control module, configured to perform primary filtering and dynamic temperature compensation on the core temperature data to obtain compensated core temperature data; Bluetooth communication module for transmitting compensated core temperature data; A power supply module, configured to supply power to the temperature sensor based on a button battery; The process of building the temperature sensor includes: Using 3D scanning technology, we obtain anatomical data of the sow's ear canal at different growth stages and physiological states, and construct a 3D model of the ear canal. The micro-thermistor is integrated on a flexible substrate and wrapped from the inside out with plastic material and skin-friendly silicone to obtain an initial temperature sensor. Based on the three-dimensional ear canal model and the ear canal variation range, the shape of the initial temperature sensor is adjusted to obtain the final temperature sensor.

2. The system according to claim 1, wherein: The data analysis module includes: a data receiving unit, configured to receive the core temperature data; a multi-scale window analysis unit, configured to capture short-term fluctuation characteristics of the core temperature data based on a preset time scale; A long short-term memory network unit, configured to capture long-term dependency features of the core temperature data based on an extended LSTM architecture; a feature classification unit, configured to extract key features of the short-term fluctuation features and the long-term dependency features, and classify the key features using an improved random forest algorithm to obtain classification results; The evaluation model construction unit is used to evaluate the sow health level based on the classification result and complete the construction of the sow health evaluation model.

3. The system according to claim 2, characterized in that The process of improving the random forest algorithm includes: Use the existing feature set to train a random forest network and obtain the average value of the reduction in node purity when each feature is used to split nodes in all decision trees during training; Calculate feature importance scores based on the average of reduced node purity; Normalizing the feature importance scores to obtain feature sampling weights; When each node is split, the features to be sampled are sampled according to the feature sampling weights to obtain a weighted randomly sampled feature subset; In the feature subset, the feature with the maximum information gain is selected as the splitting point to improve the random forest algorithm.

4. The system according to claim 1, wherein: The health warning module includes: An assessment result tracking unit, configured to locate healthy abnormal sows based on the assessment results, and obtain the duration and degree of abnormality; The early warning unit is used to provide graded early warnings based on the duration and severity of the abnormality.

5. An intelligent sow health early warning method based on ear canal core temperature monitoring, used to implement the system according to any one of claims 1 to 4, characterized in that: include: Collect core temperature data from the sow's ear canal; receiving the core temperature data, performing time series analysis on the core temperature data, and constructing a sow health assessment model; Based on the evaluation results of the sow health evaluation model, preset early warning rules are used to lock in abnormally healthy sows and issue abnormal early warnings.

6. The method according to claim 5, characterized in that The method for constructing the sow health assessment model includes: receiving the core temperature data; capturing short-term fluctuation characteristics of the core temperature data based on a preset time scale; Capturing the long-term dependency characteristics of the core temperature data based on the LSTM extended architecture; Extracting key features of the short-term fluctuation features and the long-term dependency features, and classifying the key features using an improved random forest algorithm to obtain classification results; Based on the classification results, the sow health level is assessed to complete the construction of the sow health assessment model.

Images