A method and system for beef full-chain traceability and quality evaluation based on identification sensing

Abstract

The application provides a beef full-chain traceability and quality evaluation method and system based on identification sensing, relates to the technical field of intelligent breeding and food quality traceability, and solves the technical problem that the existing technology cannot realize multi-source data fusion analysis and cross-stage quality mapping in the whole breeding process. The method specifically comprises: acquiring multi-source data of cattle; judging the behavior state of the cattle by calculating the activity intensity feature based on the behavior data; estimating the core temperature of the cattle based on the body surface temperature, activity intensity feature and temperature and humidity; calculating the animal welfare index based on the core temperature, activity intensity feature and environmental data; obtaining the fat content and protein content of the beef sample through Raman spectrum detection, and constructing the mapping relationship between the animal welfare index and the fat content and protein content; obtaining the meat quality evaluation index by weighted summation of the fat content and protein content, generating beef product full-link traceability information and visualizing the information. The application is used for beef quality traceability.

Description

Technical Field

[0001] This application relates to the fields of smart farming and food quality traceability technology, and in particular to a method and system for full-chain traceability and quality evaluation of beef based on identification sensing. Background Technology

[0002] With the development of IoT technology, smart farming has become an important means to improve livestock production efficiency and product quality. Achieving precise monitoring of the entire cattle farming process and scientific assessment of meat quality is crucial for ensuring food safety. Current technologies typically use independent ear tags, environmental monitoring stations, and subsequent chemical testing for information collection and evaluation. However, existing technologies suffer from limitations such as single data collection dimensions and a lack of multi-source data fusion and analysis mechanisms. Specifically, traditional ear tags only serve as identification markers and cannot accurately capture real-time physiological and behavioral parameters of cattle; body surface temperature monitoring is susceptible to environmental interference and cannot accurately reflect core body temperature; environmental monitoring data is not effectively correlated with individual physiological states; and there is a lack of cross-stage mapping between farming process data and final meat quality testing data, making it impossible to achieve precise traceability and quality prediction across the entire chain from the source of farming to the final product. Therefore, existing technologies face technical challenges in achieving multi-source data fusion analysis and cross-stage quality mapping throughout the entire farming process. Summary of the Invention

[0003] This application provides a method and system for whole-chain traceability and quality evaluation of beef based on identification sensing, which solves the technical problem that existing technologies cannot achieve multi-source data fusion analysis and cross-stage quality mapping throughout the entire breeding process.

[0004] To achieve the above objectives, this application adopts the following technical solution: Firstly, a method for full-chain traceability and quality assessment of beef based on identification sensing is provided, comprising: acquiring multi-source data during the cattle breeding process; the multi-source data includes body surface temperature, behavioral data, and environmental data; the environmental data includes parameters such as carbon dioxide, ammonia, temperature and humidity, and light intensity in the breeding environment; judging the behavioral state of cattle by calculating activity intensity characteristics based on the behavioral data; the activity intensity characteristics are the resultant acceleration of the cattle's three-axis accelerations; estimating the core temperature of cattle based on the body surface temperature, the activity intensity characteristics, and the temperature and humidity; calculating an animal welfare index based on the core temperature, the activity intensity characteristics, and the environmental data; obtaining the fat and protein content of beef samples by Raman spectroscopy detection, and constructing a mapping relationship between the animal welfare index and the fat and protein content; obtaining meat quality evaluation indicators by weighted summation based on the fat and protein content, generating full-chain traceability information for beef products, and visualizing the results.

[0005] In conjunction with the first aspect above, in one possible implementation, the step of determining the behavioral state of cattle by calculating activity intensity features based on the behavioral data includes: calculating the magnitude of the triaxial acceleration in the behavioral data to obtain the resultant acceleration; performing a smoothed average filter on the resultant acceleration, and then performing exponentially weighted moving average smoothing after filtering; calculating the mean and fluctuation amplitude of the smoothed resultant acceleration sequence within a unit time window as the activity intensity feature of the cattle; and determining the behavioral state of cattle based on the mean and fluctuation amplitude using a preset threshold rule.

[0006] In conjunction with the first aspect above, in one possible implementation, the estimation of the cattle's core temperature based on the body surface temperature, the activity intensity characteristics, and the temperature and humidity satisfies the following formula:

[0007] in, The core body temperature, The body surface temperature, The activity intensity characteristic, For ambient humidity, The coefficient of temperature difference. This is the activity intensity coefficient. For environmental factors, The coefficient of temperature change. For timestamps.

[0008] In conjunction with the first aspect above, in one possible implementation, the calculation of the animal welfare index based on the core temperature, the activity intensity characteristics, and the environmental data satisfies the following formula:

[0009] in, For animal welfare index, As an indicator of physiological stability, As a behavioral activity indicator, As an indicator of environmental comfort, This is the physiological stability coefficient. This represents the behavioral activity coefficient. This refers to the environmental comfort coefficient.

[0010] In conjunction with the first aspect above, in one possible implementation, the step of obtaining the fat and protein content of beef samples through Raman spectroscopy detection and constructing a mapping relationship between the animal welfare index and the fat and protein content includes: obtaining spectral data of beef samples using Raman spectroscopy detection, predicting the fat and protein content through least squares regression; constructing a multiple linear regression model based on the fat content, protein content, animal welfare index, and auxiliary variables, and fitting the regression coefficients; the auxiliary variables include cattle breed and cattle growth stage.

[0011] In conjunction with the first aspect above, in one possible implementation, the multiple linear regression model satisfies the following formula:

[0012]

[0013] in, For fat content, Protein content, For cattle breeds, This is the growth stage of cattle. , , , The regression coefficient for fat content is . , , , The regression coefficient for protein content is given.

[0014] Secondly, a beef full-chain traceability and quality assessment system based on identifier sensing is provided, comprising: a multimodal perception module, a physiological state inversion and welfare assessment module, a meat quality detection and correlation module, and a full-chain traceability visualization module; wherein, the multimodal perception module is used to acquire multi-source data during the cattle breeding process; the multi-source data includes body surface temperature, behavioral data, and environmental data; the environmental data includes parameters of carbon dioxide, ammonia, temperature, humidity, and light intensity in the breeding environment; the physiological state inversion and welfare assessment module is used to determine the behavioral state of cattle by calculating activity intensity characteristics based on the behavioral data; the activity intensity characteristics... The system is characterized by the resultant acceleration of the cattle's three axes of acceleration; the core temperature of the cattle is estimated based on the body surface temperature, activity intensity characteristics, and temperature and humidity; the animal welfare index is calculated based on the core temperature, activity intensity characteristics, and environmental data; the meat quality detection and correlation module is used to obtain the fat and protein content of beef samples through Raman spectroscopy detection, and to construct a mapping relationship between the animal welfare index and the fat and protein content; based on the fat and protein content, the meat quality evaluation index is obtained by weighted summation; the full-chain traceability visualization module is used to generate full-chain traceability information for beef products and to visualize it.

[0015] In conjunction with the second aspect above, in one possible implementation, the multimodal sensing module includes an intelligent identification sensing terminal and an environmental monitoring terminal; the intelligent identification terminal is an intelligent ear tag, including: an RFID tag for storing the unique identification of the cattle; a body temperature monitoring module, including a thermistor, for collecting the surface temperature data of the cattle; a behavior monitoring module, including a triaxial accelerometer, for collecting the movement data of the cattle; and a wireless communication module for uploading the collected data to a data processing platform; the environmental monitoring terminal consists of multiple distributed sensing nodes deployed within the farm, each integrating temperature and humidity sensors, light sensors, ammonia sensors, and carbon dioxide sensors, for collecting environmental physical quantities and gas concentration data at different spatial locations within the farm shed.

[0016] In conjunction with the second aspect above, in one possible implementation, the physiological state inversion and welfare assessment module includes a core body temperature inversion submodule and a welfare index calculation submodule; wherein, the core body temperature inversion submodule is used to estimate the core temperature of the cattle based on the body surface temperature, the activity intensity characteristics, and the temperature and humidity; the welfare index calculation submodule is used to calculate the animal welfare index based on the core temperature, the activity intensity characteristics, and the environmental data.

[0017] In conjunction with the second aspect above, in one possible implementation, the meat quality detection and correlation module includes a Raman spectroscopy analysis submodule and a welfare status quality correlation submodule; the full-chain traceability visualization module includes a data storage correlation submodule and a consumer traceability display submodule; wherein, the Raman spectroscopy analysis submodule is used to obtain spectral data of beef samples using Raman spectroscopy detection, and predict fat content and protein content through least squares regression; the welfare status quality correlation submodule is used to construct a multiple linear regression model based on fat content, protein content, animal welfare index, and auxiliary variables, and fit the regression coefficients; the auxiliary variables include cattle breed and cattle growth stage; the data storage correlation submodule is used to uniformly store and correlate physiological data, behavioral data, environmental data, and meat quality detection data during the breeding process based on the unique identification of cattle, construct a traceability file for the entire life cycle of cattle, and generate a unique QR code for each meat product to be sold; the consumer traceability display submodule is used to respond to QR code access requests and display traceability information in a visual form.

[0018] This application provides a method and system for full-chain traceability and quality assessment of beef based on identification sensing. By acquiring multi-source data during the cattle breeding process, and calculating activity intensity characteristics based on behavioral data to determine behavioral status, the system combines body surface temperature, activity intensity characteristics, and temperature and humidity to estimate core temperature, and then calculates the animal welfare index, achieving a unified quantitative characterization of the entire cattle breeding process. Furthermore, Raman spectroscopy is used to detect the chemical composition of meat products, and a mapping relationship between the animal welfare index and meat quality is constructed, realizing cross-stage mapping from breeding status to final meat quality. This solves the technical problem of existing technologies being unable to achieve multi-source data fusion analysis and cross-stage quality mapping throughout the breeding process, providing reliable technical support for accurate traceability and quality assessment of beef products.

[0019] It should be understood that the descriptions of technical features, technical solutions, beneficial effects, or similar language in this application do not imply that all features and advantages can be achieved in any single embodiment. Rather, it is understood that the description of a feature or beneficial effect means that a specific technical feature, technical solution, or beneficial effect is included in at least one embodiment. Therefore, the descriptions of technical features, technical solutions, or beneficial effects in this specification do not necessarily refer to the same embodiment. Furthermore, the technical features, technical solutions, and beneficial effects described in this embodiment can be combined in any suitable manner. Those skilled in the art will understand that embodiments can be implemented without one or more specific technical features, technical solutions, or beneficial effects of a particular embodiment. In other embodiments, additional technical features and beneficial effects may be identified in specific embodiments that do not embody all embodiments. Attached Figure Description

[0020] Figure 1 A flowchart illustrating a method for full-chain traceability and quality evaluation of beef based on identification sensing, provided for an embodiment of this application; Figure 2 A flowchart illustrating another method for whole-chain traceability and quality evaluation of beef based on identification sensing, provided for an embodiment of this application; Figure 3 A flowchart illustrating another method for whole-chain traceability and quality evaluation of beef based on identification sensing, provided for an embodiment of this application; Figure 4 A system architecture diagram of a beef full-chain traceability and quality evaluation system based on identification sensing provided in this application embodiment; Figure 5 This is a system architecture diagram of another beef whole-chain traceability and quality evaluation system based on identification sensing provided in an embodiment of this application. Detailed Implementation

[0021] In the description of this application, unless otherwise stated, " / " means "or," for example, A / B can mean A or B. The "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone. Furthermore, "at least one" means one or more, and "multiple" means two or more. The terms "first," "second," etc., do not limit the quantity or order of execution, and "first," "second," etc., do not necessarily imply differences.

[0022] It should be noted that, in this application, the terms "exemplary" or "for example" are used to indicate that something is being described as an example, illustration, or illustration. Any embodiment or design described as "exemplary" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or design solutions. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0023] To address the technical problem of existing technologies failing to achieve multi-source data fusion analysis and cross-stage quality mapping throughout the entire cattle farming process, this application provides a method for full-chain traceability and quality evaluation of beef based on identifier sensing. This method includes: acquiring multi-source data on cattle body surface temperature, behavior, and environment during the cattle farming process; calculating activity intensity characteristics to determine the cattle's behavioral state; estimating the cattle's core temperature by combining body surface temperature, activity intensity characteristics, and temperature and humidity; calculating the animal welfare index based on core temperature, activity intensity characteristics, and environmental data; obtaining beef fat and protein content through Raman spectroscopy and constructing a mapping relationship between the animal welfare index and meat quality; and generating and visualizing full-chain traceability information for beef after weighted meat quality evaluation indicators. Based on this, the method achieves multi-source data fusion analysis throughout the entire cattle farming process and cross-stage mapping from farming status to final meat quality. It can provide a unified quantitative characterization of the cattle's entire life cycle status, providing reliable technical support for accurate traceability, scientific quality assessment, and information transparency of beef products.

[0024] Example 1: like Figure 1 As shown in the embodiments of this application, the method for full-chain traceability and quality evaluation of beef based on identification sensing includes: S101. Obtain multi-source data during the cattle breeding process.

[0025] Multi-source data refers to heterogeneous data sets collected from different sources and dimensions, used to comprehensively characterize the survival status and environmental conditions of cattle.

[0026] In this embodiment, various sensor terminals deployed within the farm collect data on cattle's body surface temperature, behavior, and environmental parameters such as carbon dioxide, ammonia, temperature, humidity, and light intensity in the farming environment in real time or periodically. Body surface temperature data is obtained through the thermistor built into the smart ear tag worn by the cattle; behavior data is obtained through the triaxial accelerometer built into the smart ear tag to record the cattle's movement trajectory and posture changes; environmental data is collected through distributed sensor nodes located within the farm shed, which integrate temperature and humidity sensors, light sensors, and gas concentration sensors. The system associates and stores the above multi-source data with a unified timestamp and the unique identifier of each cattle, forming a time-series dataset.

[0027] As an example, the smart ear tag collects body surface temperature data every 5 minutes and triaxial acceleration data every 1 minute; the environmental monitoring terminal collects temperature, humidity, light, and gas concentration data every 10 minutes, and all data are uploaded to the data processing platform via a wireless communication module.

[0028] Based on the above steps, a comprehensive perception and aggregation of multi-dimensional information on the entire cattle breeding process was achieved, providing a data foundation for subsequent physiological state inversion and quality correlation analysis.

[0029] S102. Determine the behavioral state of cattle by calculating activity intensity characteristics based on behavioral data.

[0030] Among them, the activity intensity characteristic is the resultant acceleration of the cattle's three-axis acceleration, which is used to quantitatively characterize the intensity of the cattle's movement per unit time.

[0031] In this embodiment, the raw data collected by the triaxial accelerometer is first preprocessed to calculate the magnitude of the triaxial accelerations to obtain the resultant acceleration, thus eliminating the directional dependence of single-axis data. Subsequently, based on this resultant acceleration sequence, its statistical characteristics within a time window, such as mean, variance, or fluctuation amplitude, are extracted as activity intensity features. The calculated activity intensity features are matched with preset behavioral state determination rules to identify the current behavioral state of the cattle, such as lying down, standing, walking, or vigorous activity. Accurate determination of behavioral state helps distinguish between the cattle's resting and active periods, thereby providing behavioral context for subsequent body temperature estimation and welfare assessment.

[0032] It should be noted that the calculation method for activity intensity features is not limited to the mean or fluctuation amplitude of the resultant acceleration. In other embodiments, frequency domain features, energy features, or high-dimensional features extracted based on machine learning models can also be used. Any feature that can characterize the intensity of cattle movement is applicable.

[0033] As an example, the mean and standard deviation of the combined acceleration sequence over the past 10 minutes are calculated. When the mean is less than a preset threshold and the standard deviation is less than a preset threshold, the cattle are determined to be in a stationary state. When the mean is greater than the preset threshold and shows periodic fluctuations, the cattle are determined to be in a walking state.

[0034] Based on the above steps, by transforming the raw triaxial acceleration data into physically meaningful activity intensity features, the automatic identification of cattle behavioral states is realized, providing key behavioral dimension inputs for subsequent physiological state inversion.

[0035] S103. Estimate the core temperature of cattle based on body surface temperature, activity intensity characteristics, and temperature and humidity.

[0036] Core temperature refers to the deep temperature inside a cow's body. Rectal temperature or blood temperature is usually used as the gold standard and is a key indicator that reflects the true physiological state of a cow.

[0037] In this embodiment, a physiological state inversion model coupling body surface temperature, environment, and behavior is constructed. Since body surface temperature is easily affected by ambient temperature, humidity, and the cattle's own heat production during activity, directly using body surface temperature to characterize core temperature results in significant errors. Therefore, activity intensity characteristics and ambient temperature and humidity are introduced as correction factors. Through model calculation, the body surface temperature is dynamically corrected, thereby estimating a core temperature closer to the true value. Specifically, when cattle are in a state of vigorous activity, muscle heat production causes a temporary increase in body surface temperature, and the model will correspondingly reduce the estimated value based on the activity intensity characteristics. When the ambient temperature is high, heat dissipation from the body surface is hindered, and the model will adjust the temperature gradient coefficient based on ambient temperature and humidity parameters, thereby improving the estimation accuracy.

[0038] It should be noted that the core temperature estimation model is not limited to the linear weighted model. In other embodiments, neural network models, support vector regression models, or mechanistic models based on physical heat conduction mechanisms can also be used. Any model that can integrate multi-source data to achieve core temperature estimation is applicable.

[0039] As an example, the current body surface temperature, the average activity intensity over the past 30 minutes, and the current ambient temperature and humidity are input into a pre-trained regression model, and the model outputs an estimated value of the cow's current core temperature.

[0040] Based on the above steps, environmental coupling errors and behavioral interference errors in body surface temperature measurement are effectively corrected, enabling non-invasive, continuous, and dynamic estimation of the core body temperature of cattle.

[0041] S104. Calculate the animal welfare index based on core temperature, activity intensity characteristics, and environmental data.

[0042] Among them, the animal welfare index is a dimensionless comprehensive evaluation index used to quantitatively characterize the overall health status, behavioral regularity, and environmental adaptability of cattle throughout their entire life cycle.

[0043] In this embodiment, a welfare index calculation model is constructed from three dimensions: physiological stability, behavioral activity, and environmental comfort. The physiological stability index is calculated based on the fluctuation characteristics of core temperature within a time window; the smaller the temperature fluctuation, the higher the physiological stability. The behavioral activity index is calculated based on the deviation between activity intensity characteristics and reference behavioral patterns; the more regular the behavioral pattern, the higher the behavioral activity score. The environmental comfort index is calculated based on the degree of deviation between environmental parameters and optimal environmental conditions; the smaller the deviation, the higher the environmental comfort. The indices from these three dimensions are weighted and fused to obtain the final animal welfare index. This index comprehensively considers both internal (physiological and behavioral) and external (environmental) factors of cattle, and can more comprehensively reflect the welfare status of cattle.

[0044] It should be noted that the calculation of the welfare index is not limited to the three dimensions mentioned above. In other embodiments, dimensions such as food intake, water intake, or body condition score can also be introduced. The weighting coefficients can be optimized by expert scoring, analytic hierarchy process, or machine learning methods based on historical data.

[0045] Based on the above steps, a unified quantitative representation of the entire process of cattle breeding was achieved, transforming scattered physiological, behavioral, and environmental data into comparable and traceable welfare indices, providing standardized input variables for subsequent quality correlation analysis.

[0046] S105. Obtain the fat and protein content of beef samples by Raman spectroscopy, and construct the mapping relationship between animal welfare index and fat and protein content.

[0047] Raman spectroscopy is a non-destructive testing technique based on the Raman scattering effect. It enables qualitative identification and quantitative analysis of chemical components by analyzing the vibrational spectral information of molecules.

[0048] In this embodiment, after cattle slaughter, beef samples are collected and spectral data is acquired using a Raman spectroscopy detection device. A laser light source illuminates the sample surface, exciting Raman scattered light, which is received and converted into spectral data by the spectrometer. The spectral data is processed using chemometric methods (least square regression) to predict the fat and protein content in the beef samples. Subsequently, a mapping model is constructed using the animal welfare index calculated during the rearing stage as the independent variable and the fat and protein content detected by Raman spectroscopy as the dependent variables. This model reveals the mechanism by which the welfare status during the rearing process affects the quality of the final meat product, achieving a cross-stage mapping from rearing status to meat quality.

[0049] Based on the above steps, an innovative correlation model between the welfare status of the breeding process and the chemical composition of the final meat product was established, filling the gap in the existing technology where breeding data and meat quality data are separated, and providing a technical path for predicting meat quality based on the breeding process.

[0050] S106. Based on fat content and protein content, obtain meat quality evaluation indicators through weighted summation, generate full-chain traceability information for beef products, and display it visually.

[0051] Among them, the meat quality evaluation index is a quantitative score that comprehensively reflects the nutritional value of meat, and the full-chain traceability information refers to the related data chain covering the entire process of cattle breeding, slaughtering, and processing.

[0052] In this embodiment, the fat content and protein content are weighted and summed according to preset weighting coefficients to calculate a meat quality evaluation index. This index comprehensively considers the contributions of fat and protein to the taste and nutritional value of meat, and can intuitively reflect the meat grade. Subsequently, based on the unique identification of each cattle, physiological data, behavioral data, environmental data, welfare index, and post-slaughter meat quality evaluation index are correlated and integrated to generate full-chain traceability information. A unique QR code is generated for each beef product for sale, and consumers can view the full-chain traceability information of the product on their terminal devices by scanning the QR code.

[0053] It should be noted that the form of displaying traceability information is not limited to QR codes. In other embodiments, technologies such as RFID tags, barcodes, or digital watermarks can also be used.

[0054] As an example, the weight of fat content is set to 0.4 and the weight of protein content is set to 0.6 to calculate the quality score. The generated traceability information includes the ear tag number of the cattle, the name of the farm, the number of days of breeding, the average welfare index, fat content, protein content and quality score, and is displayed in the form of a webpage.

[0055] Based on the above steps, a seamless process has been achieved from the collection of breeding data, welfare status assessment, meat quality testing to the display of traceability information, providing consumers with transparent and credible traceability services and enhancing market trust in beef products.

[0056] Based on the above technical solution, this embodiment acquires multi-source data during the cattle breeding process, calculates activity intensity characteristics to determine behavioral status, and estimates core temperature by combining body surface temperature, activity intensity characteristics, and temperature and humidity. This allows for the calculation of the animal welfare index, achieving a unified quantitative representation of the entire cattle breeding process. Furthermore, Raman spectroscopy is used to detect the chemical composition of meat products, and a mapping relationship between the animal welfare index and meat quality is constructed. This enables cross-stage mapping from breeding status to final meat quality, solving the technical problem of existing technologies that cannot achieve multi-source data fusion analysis and cross-stage quality mapping throughout the breeding process. This provides reliable technical support for the accurate traceability and quality assessment of beef products.

[0057] Example 2: This embodiment, based on S102 in Embodiment 1, provides a detailed explanation of the specific algorithm implementation for determining the behavior state. For example... Figure 2 As shown, the method includes steps S201 to S204.

[0058] S201. Calculate the magnitude of the triaxial acceleration in the behavioral data to obtain the resultant acceleration.

[0059] Among them, triaxial acceleration refers to the acceleration components in the X, Y, and Z axes collected by the triaxial accelerometer built into the smart ear tag, and the resultant acceleration refers to the magnitude of the vector sum of the three acceleration components.

[0060] In this embodiment of the application, the raw data sequence acquired by the triaxial accelerometer is obtained and denoted as follows: , , Through the modulus calculation formula The resultant acceleration sequence was calculated. This eliminates the directional dependence of single-axis data, transforming three-dimensional vector data into one-dimensional scalar data, which facilitates subsequent feature extraction and state determination. It should be understood that the modulus calculation is not limited to the Euclidean norm mentioned above. In other embodiments, absolute value summation or other norm calculation methods can also be used, as long as they can characterize the overall motion intensity of the cattle.

[0061] It should be noted that the resultant acceleration can comprehensively reflect the movement state of the cow's ear. When the cow is at rest or in uniform motion, the resultant acceleration is mainly affected by the acceleration due to gravity, and the value is close to 1g. When the cow is engaged in strenuous exercise, the resultant acceleration will fluctuate significantly.

[0062] As an example, triaxial acceleration data is collected at a sampling frequency of 10Hz, generating 10 resultant acceleration data points per second to form a continuous resultant acceleration time series.

[0063] Based on the above steps, the multidimensional raw acceleration data is transformed into a single-dimensional resultant acceleration sequence, reducing the data dimensionality and providing standardized input data for subsequent filtering and feature extraction.

[0064] S202. Perform smoothing average filtering on the resultant acceleration, and then perform exponential weighted moving average smoothing after filtering.

[0065] Among them, smoothing average filtering is a time-domain filtering method that suppresses high-frequency noise by calculating the mean within a sliding time window; exponentially weighted moving average smoothing is a recursive smoothing method that reduces the impact of instantaneous fluctuations by assigning higher weights to recent data.

[0066] In this embodiment, the resultant acceleration sequence is first smoothed by averaging. Specifically, a sliding time window of length N is set, and the mean of the resultant acceleration data is calculated within the window to eliminate sampling noise and high-frequency interference. Subsequently, the filtered data is smoothed by an exponentially weighted moving average, calculated using the following formula: ,in This is the filtered value at the current moment. The smoothed value at the current moment. This is the smoothed value from the previous time step. The smoothing coefficient is used. The order of these steps is crucial. First, random noise is removed by smoothing average filtering, and then instantaneous fluctuations are reduced by exponentially weighted moving average smoothing. The two work together to preserve the overall characteristics of the trend while suppressing local jitter, thus improving the stability of the data.

[0067] It should be noted that the parameters for filtering and smoothing can be adjusted according to the actual application scenario. For example, the length N of the sliding time window can be set according to the sampling frequency and the period of the behavioral feature, and the smoothing coefficient... A trade-off can be made between the requirements for real-time performance and smoothness. In other embodiments, median filtering, Kalman filtering, or low-pass filtering can be used instead of smoothing average filtering, and moving average or Gaussian smoothing can be used instead of exponentially weighted moving average smoothing.

[0068] As an example, the sliding time window length is set to N=5, meaning the mean of five consecutive resultant acceleration data points is calculated; a smoothing coefficient is set. =0.3, recursively smooth the filtered data to obtain the smoothed resultant acceleration sequence.

[0069] Based on the above steps, through dual data processing, noise and jitter in the original data are effectively suppressed, the signal-to-noise ratio of the combined acceleration sequence is improved, and data assurance is provided for the accuracy of subsequent feature extraction.

[0070] S203. Within a unit time window, calculate the mean and fluctuation amplitude of the smoothed resultant acceleration sequence as a characteristic of the cattle's activity intensity.

[0071] Among them, the unit time window refers to the time segment used for feature extraction, the mean refers to the average level of the resultant acceleration within the window, and the fluctuation amplitude refers to the difference between the maximum and minimum values ​​of the resultant acceleration within the window, which is used to characterize the intensity of the motion.

[0072] In this embodiment, the smoothed resultant acceleration sequence is divided into continuous unit time windows, for example, every 10 seconds. Within each window, the system calculates the mean of the resultant acceleration sequence, reflecting the average movement intensity of the cattle in the current time period; simultaneously, it calculates the fluctuation amplitude of the resultant acceleration sequence, reflecting the intensity of the cattle's movement in the current time period. The mean and fluctuation amplitude together constitute the activity intensity characteristic of the cattle. This characteristic considers not only the average level of movement but also the dispersion of movement, enabling a more comprehensive characterization of the cattle's behavioral state.

[0073] As an example, the mean value A of the smoothed acceleration within a 10-second time window is calculated. mean With fluctuation range A range , the binary pair (Amean A range This serves as the activity intensity characteristic of the current window.

[0074] Based on the above steps, statistically significant features were extracted from the smoothed resultant acceleration sequence, and the time series data was transformed into a feature vector, providing quantitative input for subsequent behavior state determination.

[0075] S204. Based on the mean and fluctuation range, determine the behavioral status of cattle through preset threshold rules.

[0076] Among them, the preset threshold rule refers to the judgment conditions set based on historical data statistical analysis or expert experience, which are used to map the activity intensity characteristics to specific behavioral state categories.

[0077] In this embodiment, threshold ranges corresponding to various behavioral states are preset, including lying down, standing, walking, and vigorous activity. Specifically, when the average resultant acceleration is between 0.9g and 1.2g, and the fluctuation amplitude is less than 0.3g / 10s, it is determined to be in a lying down state. At this time, the cow's ears are mainly affected by gravity, and the whole body is in a static, low-fluctuation state. When the average resultant acceleration is between 1.0g and 1.5g, and the fluctuation amplitude is less than 0.5g / 10s, it is determined to be in a standing state. At this time, the cow has slight head movements but remains stable overall. When the average resultant acceleration is between 1.5g and 2.5g, and the fluctuation amplitude is within the range of 1.0g to 1.5g / 10s and shows periodic fluctuations, it is determined to be in a walking state. At this time, the fluctuations are consistent with the gait frequency. When the average resultant acceleration is greater than 2.5g, and the fluctuation amplitude is greater than 2.0g / 10s and shows irregular and violent changes, it is determined to be in a vigorous activity state. At this time, the cow exhibits behaviors such as running and head shaking. This threshold rule takes into account both the mean and the fluctuation range, effectively avoiding misjudgments that may be caused by a single threshold and improving the accuracy of behavior state recognition.

[0078] It should be noted that the preset threshold rules can be adaptively adjusted according to the breed, age, or breeding environment of the cattle.

[0079] As an example, if the average net acceleration of a cow within a 10-second window is detected to be 1.1g with a fluctuation range of 0.2g / 10s, and the cow is determined to be in a lying-down state according to the preset threshold rule, then the cow is determined to be in a lying-down state.

[0080] Based on the above steps, through four stages—modulus calculation, filtering and smoothing, feature extraction, and threshold determination—automatic identification of behavioral state categories from raw acceleration data is achieved, providing reliable behavioral dimension input for subsequent physiological state inversion and welfare assessment.

[0081] Based on the above technical solution, this embodiment effectively suppresses the influence of noise interference and instantaneous fluctuations by performing modulus calculation, double smoothing processing and feature extraction on the resultant acceleration, and makes judgments based on dual thresholds of mean and fluctuation amplitude. This improves the accuracy and robustness of behavioral state judgment and provides reliable behavioral feature inputs for subsequent core temperature estimation and welfare index calculation.

[0082] Example 3: This embodiment, based on S103 and S104 in Embodiment 1, provides a detailed explanation of the specific implementation of the core temperature estimation formula and the welfare index calculation formula.

[0083] For S103, the core temperature of cattle is estimated based on body surface temperature, activity intensity characteristics, and temperature and humidity, according to the following formula:

[0084] in, Core body temperature Body surface temperature As a characteristic of activity intensity, For ambient humidity, The coefficient of temperature difference. This is the activity intensity coefficient. For environmental factors, The coefficient of temperature change. For timestamps.

[0085] In the formula, the temperature difference coefficient The weighting used to characterize the impact of the difference between ambient temperature and body surface temperature on core temperature; activity intensity coefficient. Environmental coefficient is used to characterize the contribution of heat generated by bovine movement to core temperature. The rate of change of temperature is used to characterize the effect of ambient humidity on heat dissipation from the body surface. It is used to characterize the degree to which changes in body surface temperature affect the dynamic adjustment of core temperature.

[0086] In this embodiment, a physiological state inversion model coupling body surface temperature, environment, and behavior is constructed. This model is not a simple linear weighting, but rather designed based on heat conduction mechanisms and biological thermoregulation mechanisms. Specifically, when the ambient temperature... Higher than body surface temperature At this time, ambient heat will be transferred to the surface of the cattle's body, causing the surface temperature to rise. The temperature difference at this time... If positive, through Adjust its positive contribution to core temperature estimation; conversely, when the ambient temperature is lower than the body surface temperature, heat dissipation from the body surface accelerates, and the temperature difference term becomes negative, through... Adjust its negative contribution to core temperature estimation. Activity intensity characteristics. This reflects the heat production in cattle muscles. When cattle are in a state of vigorous activity, muscle metabolic heat production increases, and the core temperature rises accordingly. Adjust this positive contribution. Ambient humidity. High humidity affects the efficiency of heat dissipation through sweat evaporation, hindering heat dissipation and potentially increasing core temperature. Moderating this effect. Rate of change in body surface temperature. This reflects the dynamic trend of body temperature changes. A rapid rise in surface temperature may indicate a lagging increase in core temperature. This dynamic correction term is introduced. The above coefficients... , , , The values ​​are not subjectively set, but determined through fitting historical data. The system collects a large amount of data on the body surface temperature, environmental parameters, activity intensity characteristics, and corresponding rectal temperature (as the true value of core temperature) of cattle. The least squares method is used to perform multiple regression analysis to fit the optimal values ​​of each coefficient, thereby ensuring the scientific nature and accuracy of the model.

[0087] Based on the above steps, by introducing four correction factors—ambient temperature difference, activity intensity, ambient humidity, and temperature change rate—a multi-parameter coupled core temperature estimation model was constructed. This model effectively corrects environmental coupling errors and behavioral interference errors in body surface temperature measurement, achieving non-invasive, high-precision dynamic estimation of cattle core body temperature and solving the problem of inaccurate body surface temperature measurement in existing technologies.

[0088] For S104, based on core temperature, activity intensity characteristics, and environmental data, the animal welfare index is calculated according to the following formula:

[0089] in, For animal welfare index, As an indicator of physiological stability, As a behavioral activity indicator, As an indicator of environmental comfort, This is the physiological stability coefficient. This represents the behavioral activity coefficient. This refers to the environmental comfort coefficient.

[0090] In the formula, physiological stability index Indicators used to characterize the stability and behavioral activity of the bovine thermoregulation system Indicators used to characterize the regularity and activity level of cattle behavior patterns, and environmental comfort levels. Used to characterize the degree of matching between the breeding environment and the optimal environment for cattle. , , These are the weighting coefficients for the corresponding indicators, used to adjust the contribution ratio of each dimension to the comprehensive welfare index.

[0091] In this embodiment, a welfare index calculation model is constructed from three dimensions: physiological, behavioral, and environmental. Physiological stability index. Based on the fluctuation characteristics of core temperature within a time window, the specific calculation method is as follows: ,in For the first time window i The core temperature of every moment This represents the average core temperature within that time window. N This represents the number of sampling points. This indicator reflects the relative fluctuation of body temperature; the smaller the temperature fluctuation, the lower the temperature variation. The closer a value is to 1, the higher the physiological stability. (Behavioral activity index) The calculation method is as follows: The deviation between activity intensity characteristics and reference behavioral patterns is constructed based on the following: ,in The resultant acceleration is obtained by calculating the magnitude of the triaxial acceleration to represent the activity intensity characteristics per unit time. Then in length of N within the time window Moving average filtering and smoothing are performed, and the mean resultant acceleration within the time window is calculated as a feature of activity intensity. , This is a reference activity level based on historical statistics of healthy cattle herds. This indicator reflects the degree of match between the current behavioral state and the normal behavioral level. near hour, A value close to 1 indicates normal behavioral activity. Environmental comfort index. The calculation method is based on the degree of deviation between environmental parameters and optimal environmental conditions. ,in , , These are the optimal ambient temperature, humidity, and light intensity, respectively. , These are the concentrations of ammonia and carbon dioxide, respectively. to These are the weighting coefficients for each environmental parameter. This index comprehensively reflects the impact of temperature, humidity, light intensity, and gas concentration on environmental comfort. The closer the environmental parameters are to the optimal conditions, the better. The closer a value is to 1, the higher the level of environmental comfort. (Weighting coefficient) , , The value of is determined through expert scoring or machine learning methods based on historical data. For example, if physiological stability has a more critical impact on welfare status, then is set to . Larger; if environmental comfort is more important for a specific breeding stage, then set Relatively large.

[0092] As an example, calculate the physiological stability index of a certain cow over the past 24 hours. =0.85, behavioral activity index =0.90, Environmental comfort index =0.80, and set the weighting coefficient. =0.4, =0.3, =0.3, substituting into the formula, we obtain the animal welfare index. =0.85 indicates that the overall welfare status of the cattle is good.

[0093] Based on the above steps, by constructing quantitative indicators of three dimensions—physiological stability, behavioral activity, and environmental comfort—and performing weighted fusion, a unified quantitative representation of the comprehensive state of cattle throughout their entire life cycle is achieved. This transforms scattered physiological, behavioral, and environmental data into comparable and traceable welfare indices, providing standardized input variables for subsequent quality correlation analysis and overcoming the problems of strong subjectivity and lack of quantitative standards in existing welfare assessments.

[0094] Based on the above technical solution, this embodiment demonstrates the scientific nature and rigor of the model by elaborating on the physical meaning, construction logic and parameter fitting method of the core temperature estimation formula and the welfare index calculation formula, effectively defending against the doubts of "subjective assignment", and providing reliable intermediate variable support for the cross-stage mapping from breeding status to meat quality.

[0095] Example 4: This embodiment, based on S105 in Embodiment 1, provides a detailed explanation of the Raman spectroscopy detection procedure and the construction of the multiple linear regression model. For example... Figure 3 As shown, the method includes S301 to S303.

[0096] S301. Raman spectroscopy was used to obtain spectral data of beef samples, and least squares regression was used to predict fat and protein content.

[0097] Raman spectroscopy is a non-destructive testing technique based on the Raman scattering effect. It enables qualitative identification and quantitative analysis of chemical components by analyzing the vibrational spectral information of molecules.

[0098] In this embodiment, after cattle slaughter, beef samples are collected and spectral data is acquired using a Raman spectroscopy detection device. Specifically, a laser light source illuminates the sample surface, exciting Raman scattered light, which is received by the spectrometer and converted into spectral data. The system preprocesses the raw spectral data, including noise reduction, baseline correction, and normalization, to eliminate noise interference and systematic errors. Subsequently, chemometric methods, such as least squares regression, are used to process the spectral data to predict the fat and protein content in the beef sample. This step enables rapid and non-destructive detection of the chemical components of meat products, avoiding the problems of complex operation, long detection cycle, and destructive nature of traditional chemical analysis methods.

[0099] As an example, a near-infrared semiconductor laser with a center wavelength of 785 nm was used to excite beef samples. The laser power was set to 300 mW and the integration time was set to 10 seconds. Raman spectral data were collected, and the fat content was predicted to be 15.2% and the protein content to be 20.5% by partial least squares regression model.

[0100] Based on the above steps, Raman spectroscopy was used to rapidly and non-destructively obtain the chemical components of meat products, providing accurate quality data support for the subsequent construction of a correlation model between farming status and meat quality.

[0101] S302. Construct a multiple linear regression model based on fat content, protein content, animal welfare index, and auxiliary variables.

[0102] Among them, auxiliary variables include cattle breed and cattle growth stage. The multiple linear regression model is a statistical model used to analyze the linear relationship between multiple independent variables and the dependent variable.

[0103] In this embodiment, an animal welfare index calculated during the rearing stage is used as the independent variable, and fat and protein content detected by Raman spectroscopy are used as the dependent variables to construct a mapping model. However, relying solely on the animal welfare index as a single variable often results in insufficient predictive accuracy, because meat quality is not only affected by the welfare status during the rearing process but also significantly influenced by the genetic characteristics of cattle breeds and the physiological characteristics of their growth stages. For example, different breeds of cattle exhibit natural differences in fat deposition capacity and protein synthesis efficiency; the developmental levels of muscle and adipose tissue also differ at different growth stages for the same breed. Therefore, cattle breed and growth stage are introduced as auxiliary variables to construct a multiple linear regression model to improve the model's predictive accuracy and applicability. Specifically, the system numerically encodes the auxiliary variables: for cattle breeds, a one-hot encoding method is used to convert breed categories into numerical variables; for cattle growth stages, segmented continuous mapping is performed based on the cattle's age, mapping the calf stage, growing stage, and finishing stage to different numerical intervals.

[0104] As an example, welfare index data, breed information, growth stage information, and corresponding fat and protein content data of meat products of 100 cattle were collected to construct a multiple linear regression model. The breed variable was converted into three feature bits using one-hot encoding, and the growth stage variable was mapped to continuous values ​​between 0 and 1 based on age.

[0105] Based on the above steps, by introducing auxiliary variables to construct a multiple linear regression model, the problem of insufficient accuracy of single-variable models is effectively solved, realizing cross-stage mapping from breeding status to meat quality, and providing a scientific technical path for predicting meat quality based on the breeding process.

[0106] S303, Fitting regression coefficients.

[0107] Among them, the regression coefficient refers to the weight parameters of each variable in the multiple linear regression model, which are used to characterize the degree of contribution of each variable to the dependent variable.

[0108] In this embodiment, the fat and protein content of beef samples obtained by Raman spectroscopy, the animal welfare index, and auxiliary variables are substituted into a multiple linear regression formula, and the regression coefficients are fitted using 1stOpt software. For the fat content prediction model, the regression coefficients are obtained through fitting. 、 、 、 For the protein content prediction model, the regression coefficients were obtained through fitting. 、 、 、 .in, and This reflects the weighting of the animal welfare index on the effects of fat and protein content. 、 and 、 These figures reflect the influence weights of breed and growth stage, respectively. By using the regression coefficients obtained through fitting, the system can substitute the welfare index, breed, and growth stage data of any cattle into the model to predict the fat and protein content of its meat, thereby achieving a cross-stage mapping from farming status to meat quality.

[0109] Alternatively, the multiple linear regression model satisfies the following formula:

[0110]

[0111] in, For fat content, Protein content, For cattle breeds, This is the growth stage of cattle. 、 、 、 The regression coefficient for fat content is . 、 、 、 The regression coefficient for protein content is denoted as . 、 、 、 The sum is 1. 、 、 、 The sum is 1.

[0112] Based on the above steps, by fitting regression coefficients, a quantitative relationship between animal welfare index, breed, growth stage and meat chemical composition was established, realizing a cross-stage mapping from farming status to meat quality. This fills the gap in existing technologies where farming data and meat quality data are separated, and provides reliable technical support for the accurate traceability and quality assessment of beef products.

[0113] Based on the above technical solution, this embodiment obtains the chemical composition of meat products by Raman spectroscopy detection, and introduces variety and growth stage as auxiliary variables to construct a multiple linear regression model, which effectively improves the prediction accuracy and applicability of the model and realizes cross-stage mapping from breeding status to final meat product quality.

[0114] Example 5: like Figure 4 As shown, this embodiment provides a beef full-chain traceability and quality assessment system 40 based on identifier sensing. This system is used to execute the various steps in the above method embodiments. The system includes: a multimodal perception module 401, a physiological state inversion and welfare assessment module 402, a meat quality detection and correlation module 403, and a full-chain traceability visualization module 404.

[0115] The multimodal sensing module 401, serving as the system's data acquisition front-end, is responsible for all-weather, multi-dimensional sensing of the real-time status of individual cattle and their environment. This module integrates multiple sensing mechanisms to convert analog signals from the physical world into digital signals, adding timestamps and identification tags to provide raw materials for subsequent data processing. Specifically, the multimodal sensing module 401 collects surface temperature and triaxial acceleration data through smart identification terminals worn by individual cattle, and simultaneously collects temperature, humidity, light, and gas concentration data through an environmental monitoring network deployed within the farm. This module not only handles data collection but also performs preliminary data cleaning, packaging, and uploading, ensuring data integrity and real-time performance.

[0116] The physiological state inversion and welfare assessment module 402 is the core computing engine of the system, responsible for transforming the raw data collected by the multimodal sensing module 401 into biologically meaningful state indicators. This module receives surface temperature, triaxial acceleration, and environmental parameters uploaded by the multimodal sensing module 401. First, it processes the acceleration data using a built-in behavioral analysis algorithm to extract activity intensity features and determine behavioral state. Then, it calls the physiological state inversion model, combining environmental parameters to correct the surface temperature and estimate the core temperature. Finally, it integrates the core temperature, activity intensity features, and environmental data to calculate the animal welfare index. This module realizes the transformation process from "data" to "information" and then to "knowledge," uniformly representing the scattered sensing data as the welfare state of cattle.

[0117] The meat quality testing and correlation module 403 is responsible for breaking down data barriers between the farming and consumer ends, enabling cross-stage mapping from farming status to meat quality. This module intervenes after cattle slaughter, using a Raman spectroscopy detection terminal to perform non-destructive testing on beef samples, acquiring spectral data and predicting fat and protein content. Subsequently, the module retrieves the animal welfare index calculated during the cattle's farming stage, and, combined with breed and growth stage, constructs and runs a multiple linear regression model to establish a quantitative relationship between the welfare index and the meat's chemical composition. Finally, based on preset weights, it calculates meat quality evaluation indicators, completing the digital grading of quality.

[0118] The end-to-end traceability visualization module 404 serves as the system's user-facing interactive window, responsible for transforming complex backend data into intuitive and easy-to-understand traceability information. Based on the unique identification of each cattle, this module integrates physiological, behavioral, environmental, and welfare data from the breeding process, as well as post-slaughter meat quality evaluation indicators, to generate an end-to-end traceability profile. A unique QR code is generated for each beef product for sale. Consumers can scan the QR code to view the product's end-to-end traceability information on their devices, including cattle breed, breeding cycle, welfare status curve, and meat nutritional components.

[0119] Based on the above technical solution, this embodiment achieves closed-loop management of data collection, state inversion, quality association and traceability display through the collaborative work of the multimodal perception module 401, the physiological state inversion and welfare assessment module 402, the meat quality detection and association module 403 and the full-chain traceability visualization module 404. The data flow between the modules is smooth and the functions are closely coupled, which together support the cross-stage mapping from breeding state to meat quality, and solves the technical problems of single system function and lack of cross-link data association mechanism in the prior art.

[0120] Example 6: like Figure 5 As shown, this embodiment, based on embodiment 5, provides a detailed description of the system's hardware configuration and sub-module division.

[0121] In some embodiments, the multimodal sensing module 401 includes an intelligent identification sensing terminal and an environmental monitoring terminal.

[0122] Among them, intelligent identification sensing terminals refer to integrated sensing devices worn on individual cattle, while environmental monitoring terminals refer to fixed sensing networks deployed in the physical space of livestock farming.

[0123] In this embodiment, the intelligent identification terminal is specifically an intelligent ear tag, which integrates an RFID tag, a body temperature monitoring module, a behavior monitoring module, and a wireless communication module. The RFID tag stores the unique identification of each cattle, ensuring accurate data binding to the individual. The body temperature monitoring module includes a thermistor, which collects surface temperature data by contacting the cattle's ear skin, offering advantages such as small size and fast response. The behavior monitoring module includes a triaxial accelerometer, used to collect the cattle's motion data in three-dimensional space, capable of capturing subtle changes in movement. The wireless communication module uploads the collected data to a data processing platform, enabling real-time data transmission. The environmental monitoring terminal specifically comprises multiple distributed sensing nodes deployed within the farm. These nodes integrate temperature and humidity sensors, light sensors, ammonia sensors, and carbon dioxide sensors, used to collect environmental physical quantities and gas concentration data at different spatial locations within the farm, thereby constructing a gridded environmental monitoring network.

[0124] In some embodiments, the physiological state inversion and welfare assessment module 402 includes a core body temperature inversion submodule and a welfare index calculation submodule.

[0125] Among them, the core body temperature inversion submodule is used to execute the core temperature estimation logic, and the welfare index calculation submodule is used to execute the welfare index calculation logic.

[0126] In this embodiment, the core body temperature inversion submodule receives surface temperature, activity intensity characteristics, and environmental temperature and humidity data uploaded by the multimodal sensing module 401, and calls a preset physiological state inversion model to estimate the core temperature of the cattle. This submodule encapsulates the complex model calculation process in the background and outputs high-precision core temperature data. The welfare index calculation submodule receives the output results of the core body temperature inversion submodule, as well as activity intensity characteristics and environmental data, and calls a welfare index calculation model to calculate the animal welfare index. This submodule implements weighted fusion of multi-dimensional indicators and outputs a standardized welfare status score.

[0127] In some embodiments, the meat quality inspection and correlation module 403 and the full-chain traceability visualization module 404 include a Raman spectroscopy analysis submodule and a welfare status quality correlation submodule; the full-chain traceability visualization module 404 includes a consumer traceability display submodule.

[0128] Among them, the Raman spectroscopy analysis submodule is used to process spectral data and predict chemical composition, the welfare status quality correlation submodule is used to build a mapping model between farming status and meat quality, the data storage correlation submodule is used for persistent storage and correlation of data, and the consumer traceability display submodule is used for information display for users.

[0129] In this embodiment, the Raman spectroscopy analysis submodule uses Raman spectroscopy to acquire spectral data of beef samples and predicts fat and protein content through least squares regression, achieving rapid and non-destructive detection of meat quality. The welfare status quality correlation submodule constructs a multiple linear regression model based on fat content, protein content, animal welfare index, and auxiliary variables, fitting regression coefficients to reveal the impact mechanism of the breeding process on meat quality. The data storage correlation submodule, based on the unique identification of cattle, uniformly stores and correlates physiological, behavioral, environmental, and meat quality testing data during the breeding process, constructing a full life-cycle traceability file for cattle and generating a unique QR code for each piece of meat to be sold, achieving cross-stage data integration. The consumer traceability display submodule responds to QR code access requests and displays traceability information in a visual format, achieving transparent information presentation.

[0130] Based on the above technical solution, this embodiment provides specific hardware support and logical implementation for functional limitations by describing in detail the hardware composition such as smart ear tags and environmental monitoring terminals, as well as the division of core sub-modules such as body temperature inversion and welfare index calculation. This reflects the binary binding of "structure-function" and ensures the feasibility and implementation of the technical solution.

[0131] Although this application has been described herein in conjunction with various embodiments, those skilled in the art, by reviewing the accompanying drawings, disclosure, and appended claims, will understand and implement other variations of the disclosed embodiments in carrying out the claimed application. In the claims, the word "comprising" does not exclude other components or steps, and "a" or "an" does not exclude multiple instances. A single processor or other unit can implement several functions listed in the claims. While different dependent claims may recite certain measures, this does not mean that these measures cannot be combined to produce good results.

[0132] Although this application has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made thereto without departing from the spirit and scope of this application. Accordingly, this specification and drawings are merely exemplary illustrations of this application as defined by the appended claims, and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from the spirit and scope of this application. Thus, if such modifications and modifications of this application fall within the scope of the claims of this application and their equivalents, this application is also intended to include such modifications and modifications.

Claims

1. A method for full-chain traceability and quality evaluation of beef based on identifier sensing, characterized in that, include: Acquire multi-source data during the cattle breeding process; the multi-source data includes body surface temperature, behavioral data, and environmental data; the environmental data includes parameters such as carbon dioxide, ammonia, temperature and humidity, and light intensity in the breeding environment. The behavioral state of cattle is determined by calculating activity intensity characteristics based on the behavioral data; the activity intensity characteristics are the resultant acceleration of the cattle's three-axis acceleration. The core temperature of cattle is estimated based on the body surface temperature, the activity intensity characteristics, and the temperature and humidity. Based on the core temperature, the activity intensity characteristics, and the environmental data, the animal welfare index is calculated. The fat and protein content of beef samples were obtained by Raman spectroscopy, and the mapping relationship between animal welfare index and fat and protein content was constructed. Based on the fat and protein content, a meat quality evaluation index is obtained by weighted summation, generating full-chain traceability information for beef products and displaying it visually.

2. The method according to claim 1, characterized in that, The process of determining the behavioral state of cattle by calculating activity intensity characteristics based on the behavioral data includes: The magnitude of the triaxial acceleration in the behavioral data is calculated to obtain the resultant acceleration; The combined acceleration is subjected to a smooth average filter, and then subjected to an exponentially weighted moving average smoothing process. Within a unit time window, the mean and fluctuation amplitude of the smoothed resultant acceleration sequence are calculated as characteristics of the cattle's activity intensity. Based on the mean and fluctuation range, the behavioral state of the cattle is determined by a preset threshold rule.

3. The method according to claim 1, characterized in that, The estimation of the cattle's core temperature based on the body surface temperature, activity intensity characteristics, and temperature and humidity satisfies the following formula: in, The core body temperature, The body surface temperature, The activity intensity characteristic, For ambient humidity, The coefficient of temperature difference. This is the activity intensity coefficient. For environmental factors, The coefficient of temperature change. For timestamps.

4. The method according to claim 1, characterized in that, Based on the core temperature, the activity intensity characteristics, and the environmental data, the animal welfare index is calculated according to the following formula: in, For animal welfare index, As an indicator of physiological stability, As a behavioral activity indicator, As an indicator of environmental comfort, This is the physiological stability coefficient. This represents the behavioral activity coefficient. This refers to the environmental comfort coefficient.

5. The method according to claim 1, characterized in that, The method of obtaining the fat and protein content of beef samples through Raman spectroscopy and constructing a mapping relationship between the animal welfare index and the fat and protein content includes: Raman spectroscopy was used to obtain spectral data of beef samples, and least squares regression was used to predict fat and protein content. A multiple linear regression model was constructed based on fat content, protein content, animal welfare index, and auxiliary variables, and regression coefficients were fitted. The auxiliary variables included cattle breed and cattle growth stage.

6. The method according to claim 5, characterized in that, The multiple linear regression model satisfies the following formula: in, For fat content, Protein content, For cattle breeds, This is the growth stage of cattle. , , , The regression coefficient for fat content is . , , , The regression coefficient for protein content is given.

7. A beef whole-chain traceability and quality evaluation system based on identifier sensing, characterized in that, The system includes: a multimodal perception module, a physiological state inversion and welfare assessment module, a meat quality detection and correlation module, and a full-chain traceability visualization module; The multimodal sensing module is used to acquire multi-source data during the cattle breeding process; the multi-source data includes body surface temperature, behavioral data, and environmental data; the environmental data includes parameters such as carbon dioxide, ammonia, temperature and humidity, and light intensity in the breeding environment. The physiological state inversion and welfare assessment module is used to determine the behavioral state of cattle by calculating activity intensity characteristics based on the behavioral data; the activity intensity characteristics are the resultant acceleration of the cattle's three-axis acceleration; the core temperature of cattle is estimated based on the body surface temperature, the activity intensity characteristics, and the temperature and humidity; and the animal welfare index is calculated based on the core temperature, the activity intensity characteristics, and the environmental data. The meat quality detection and correlation module is used to obtain the fat and protein content of beef samples through Raman spectroscopy detection, construct the mapping relationship between animal welfare index and fat and protein content, and obtain meat quality evaluation index by weighted summation based on the fat and protein content. The full-chain traceability visualization module is used to generate full-chain traceability information for beef products and to visualize it.

8. The system according to claim 7, characterized in that, The multimodal sensing module includes an intelligent identification sensing terminal and an environmental monitoring terminal; The intelligent identification terminal is an intelligent ear tag, comprising: an RFID tag for storing the unique identification of the cattle; a body temperature monitoring module, including a thermistor, for collecting the surface temperature data of the cattle; a behavior monitoring module, including a triaxial accelerometer, for collecting the movement data of the cattle; and a wireless communication module for uploading the collected data to a data processing platform. The environmental monitoring terminal consists of multiple distributed sensing nodes deployed within the farm. Each sensing node integrates temperature and humidity sensors, light sensors, ammonia sensors, and carbon dioxide sensors to collect environmental physical quantities and gas concentration data at different locations within the farm.

9. The system according to claim 7, characterized in that, The physiological state inversion and welfare assessment module includes a core body temperature inversion submodule and a welfare index calculation submodule; The core body temperature inversion submodule is used to estimate the core temperature of cattle based on the body surface temperature, the activity intensity characteristics, and the temperature and humidity. The welfare index calculation submodule is used to calculate the animal welfare index based on the core temperature, the activity intensity characteristics, and the environmental data.

10. The system according to claim 7, characterized in that, The meat quality testing and correlation module includes a Raman spectroscopy analysis submodule and a welfare status quality correlation submodule; the full-chain traceability visualization module includes a data storage correlation submodule and a consumer traceability display submodule. The Raman spectroscopy analysis submodule is used to obtain spectral data of beef samples by Raman spectroscopy detection and to predict fat and protein content by least squares regression. The welfare status quality correlation submodule is used to construct a multiple linear regression model based on fat content, protein content, animal welfare index and auxiliary variables, and fit the regression coefficients; the auxiliary variables include cattle breed and cattle growth stage. The data storage association submodule is used to uniformly store and associate physiological data, behavioral data, environmental data and meat quality testing data during the breeding process based on the unique identification of cattle, to build a traceability file for the entire life cycle of cattle, and to generate a unique QR code for each meat product to be sold. The consumer traceability display submodule is used to respond to QR code access requests and display traceability information in a visual form.

Images