Meat duck breast muscle measuring system and method thereof

Abstract

The invention discloses a meat duck pectoralis determination system and method. The system comprises a multi-modal pectoralis image fusion divider module, a muscle three-dimensional feature principal component fusion model module, a heterogeneous pectoralis multi-index grading model module, a poultry meat muscle parameter management analysis platform, a meat duck pectoralis parameter acquisition module and a parameter result output module. The method comprises the following steps: acquiring parameters such as muscle fiber diameter and intermuscular fat content through a meat duck pectoralis parameter acquisition module, storing and associating the parameters through a poultry meat muscle parameter management and analysis platform, and respectively transmitting the parameters to a multi-modal pectoralis image fusion divider and a muscle three-dimensional feature principal component fusion model; the former processes a multi-source image to generate a segmented image, the latter processes parameters to generate a feature vector, the two data are transmitted into a heterogeneous pectoralis multi-index grading model to analyze and judge grades, and finally, a parameter result output module integrates an output result. The meat duck breast muscle measuring efficiency and accuracy are improved, and the quality control requirement of the meat duck industry is met.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of meat duck breast muscle measurement, and in particular to a system and method for meat duck breast muscle measurement. BACKGROUND

[0002] With the development of the meat duck breeding industry, the market's demand for meat duck breast muscle quality is gradually increasing. The muscle fiber structure, fat content, tenderness and other parameters of meat duck breast muscle directly affect the edible value and commercial value of the product. Precise measurement of these parameters is a key requirement for the industry to ensure meat quality and optimize production processes. Current meat duck breast muscle measurement work needs to cover image information collection, multi-class muscle parameter acquisition, feature analysis and grade determination. Traditional measurement methods rely on single detection equipment and manual analysis, making it difficult to achieve collaborative processing of multi-source data. In addition, the parameter storage, retrieval and result output in the measurement process lack systematic management, which cannot meet the efficient and accurate measurement requirements. Therefore, it is urgent to build a meat duck breast muscle measurement system and method that integrates multiple modules and can realize data linkage and intelligent analysis to improve measurement efficiency and result accuracy and adapt to the fine requirements of the industry for meat duck breast muscle quality control.

[0003] The existing technology has two significant shortcomings in meat duck breast muscle measurement. On the one hand, the existing measurement technology cannot effectively integrate and link multi-source data. In the image collection and muscle parameter acquisition process, there is often a gap between image information and muscle fiber diameter, moisture content and other parameter data, making it difficult to establish a deep connection between image features and muscle parameters, which affects the completeness of the measurement results. On the other hand, the analysis model and grading mechanism in the existing technology lack adaptability. They are designed for single parameters or single image types, making it difficult to process multi-modal image features and multi-dimensional muscle parameters simultaneously. Moreover, the allocation of weights for each indicator in the grading process lacks scientific basis, making it difficult to accurately reflect the comprehensive influence of different parameters on meat duck breast muscle quality, resulting in deviations between the grading results and the actual quality, which cannot meet the industry's demand for accurate grading of meat duck breast muscle quality. SUMMARY

[0004] To overcome the shortcomings and deficiencies of the existing technology, the present application provides a system and method for meat duck breast muscle measurement.

[0005] The technical solution adopted by the present application is a system for meat duck breast muscle measurement, which includes a multi-modal breast muscle image fusion segmenter module that receives visible light images, near-infrared images and ultrasonic tomographic images of meat duck breast muscle, and generates a meat duck breast muscle full-domain segmentation image through multi-source image pixel-level registration and feature layer fusion processing.

[0006] The duck breast muscle parameter acquisition module collects parameters such as muscle fiber diameter, intramuscular fat content, pH value, shear force value, and moisture content of duck breast muscle samples in real time, and transmits the collected data to the poultry muscle muscle parameter management and analysis platform. The poultry muscle muscle parameter management and analysis platform classifies, stores, and maps the received parameter data, and then transmits the image data to the multimodal breast muscle image fusion segmenter module and the parameter data to the muscle muscle three-dimensional feature principal component fusion model module. The muscle muscle three-dimensional feature principal component fusion model module performs feature extraction and principal component dimensionality reduction processing on the received parameter data.

[0007] The heterogeneous pectoral muscle multi-index grading model module receives the segmented image features output by the multimodal pectoral muscle image fusion segmenter module and the feature vector output by the principal component fusion model module of the three-dimensional features of the muscle tissue, and performs multi-dimensional index correlation analysis and grade determination processing.

[0008] The parameter result output module receives the judgment results output by the heterogeneous pectoral muscle multi-index grading model module and the raw parameter data stored in the poultry muscle muscle parameter management and analysis platform. After data integration and format conversion, it outputs the measurement results of duck pectoral muscle.

[0009] Furthermore, the multimodal pectoral muscle image fusion segmenter module performs image fusion segmentation using the following formula:

[0010]

[0011] in, The pixel values ​​of the merged image of the duck breast muscle. Image pixel coordinates, The fusion weights for visible light images, near-infrared images, and ultrasound tomography images are respectively... These are the pixel values ​​of a visible light image. These are the pixel values ​​of the near-infrared image. These are the pixel values ​​of the ultrasound tomography image. These are the feature enhancement functions for the three types of images; the three-dimensional feature principal component fusion model module for muscle tissue uses the following formula for feature fusion: ,in, This is a fusion vector of three-dimensional features of the muscle tissue. The number of parameters for the breast muscle of meat ducks. For the first Principal component weights of various parameters For the first Types of parameters Principal component analysis results, For the first A three-dimensional feature mapping matrix of various parameters.

[0012] Furthermore, the principal component fusion model module for the three-dimensional features of the sarcoplasm uses the following formula for principal component fusion: ,in, Principal component fusion value, The number of samples for a single parameter. For the first The parameter values ​​of each sample, For the first The mean of the parameter category of each sample. For the first The principal component contribution rate of each sample; the heterogeneous pectoral muscle multi-index grading model module uses the following formula for grading: Grade Among them, Grade represents the grading results of the duck breast muscle. For the number of indicators, For the first The characteristic values ​​of each indicator For the first The weight of each indicator, The total number of levels, This is the floor function.

[0013] Furthermore, the heterogeneous pectoral muscle multi-index grading model module uses the following formula for multi-index analysis: ,in, The scoring is based on a combination of multiple indicators. The number of indicators used in the scoring. For the first The scoring coefficients for each indicator For the first The actual measured value of each indicator For the first The minimum value of each indicator. For the first The maximum value of each indicator; the multimodal pectoral muscle image fusion segmenter module uses the following formula for segmentation: ,in, For the segmentation results, 1 represents the pectoral muscle region, and 0 represents the non-pectoral muscle region. The segmentation threshold is... To merge image pixel values.

[0014] Furthermore, the poultry muscle tissue parameter management and analysis platform uses the following formula for parameter correlation: Where Corr is the correlation coefficient between the two parameters. The total number of samples, For the first The first parameter value of each sample The mean of the first parameter, For the first The second parameter value of each sample, The mean of the second parameter is used; the three-dimensional feature principal component fusion model module of the sarcoplasm generates feature vectors using the following formula: ,in, For feature vectors, The number of feature dimensions. For the first Weights of each dimension For the first Feature values ​​in each dimension.

[0015] Furthermore, the poultry muscle tissue parameter management and analysis platform uses the following formula for data storage mapping: ,in, For parameters The storage address mapping value, This represents the number of sub-parameters after parameter decomposition. For the first The mapping coefficients of each sub-parameter, For hash functions, For the first Sub-parameters; the heterogeneous pectoral muscle multi-index grading model module uses the following formula to determine the rank: Rank Where Rank is the level determination value. For the number of indicators, For the first The weighting of each indicator, For the first The measured values ​​of each indicator, For the first The lower limit of each indicator, For the first The upper limit of each indicator.

[0016] Furthermore, the heterogeneous pectoral muscle multi-index grading model module includes an index feature extraction unit, a multi-dimensional weight allocation unit, a grading threshold calculation unit, and a grade output unit. The index feature extraction unit receives the segmented image output by the multimodal pectoral muscle image fusion segmenter module and the feature vector output by the muscle tissue three-dimensional feature principal component fusion model module. It extracts edge and texture features from the segmented image, performs dimensional feature parsing on the feature vector, and integrates the extracted image features and the parsed vector features into a multi-index feature set. The multi-dimensional weight allocation unit receives the multi-index feature set output by the index feature extraction unit and, based on the influence of different features in the quality determination of duck pectoral muscle, performs hierarchical weight allocation. The analysis method calculates the weight coefficients corresponding to different features and generates a weight allocation matrix. The grading threshold calculation unit receives the weight allocation matrix output by the multi-dimensional weight allocation unit and the historical parameter data transmitted by the poultry muscle muscle parameter management and analysis platform. It performs statistical analysis on the historical parameter data and determines the feature threshold range corresponding to different levels by combining the weight allocation matrix, and generates a grading threshold table. The level output unit receives the grading threshold table output by the grading threshold calculation unit and the multi-indicator feature set output by the indicator feature extraction unit. It compares the different feature values ​​in the multi-indicator feature set with the corresponding thresholds in the grading threshold table, determines the final level of the duck breast muscle based on the comparison results, and transmits it to the parameter result output module.

[0017] Furthermore, the poultry muscle muscle parameter management and analysis platform includes a parameter receiving unit, a data classification and storage unit, a parameter correlation analysis unit, and a data retrieval unit. The parameter receiving unit receives parameter data such as muscle fiber diameter, intramuscular fat content, pH value, shear force value, and moisture content transmitted from the duck breast muscle parameter acquisition module. It performs format verification on the parameter data, discards data with incorrect formats, and transmits the verified data to the data classification and storage unit. The data classification and storage unit receives the verified data output from the parameter receiving unit, establishes different data storage directories according to parameter types, and stores the different parameter data in order of acquisition time. The system retrieves data from the corresponding directory and adds a unique identifier code to each data entry. The parameter correlation analysis unit receives the stored data output by the data classification and storage unit, calculates the Pearson correlation coefficient between different parameters, analyzes the degree of correlation between different parameters, and generates a parameter correlation table. The data retrieval unit receives data retrieval requests from the multimodal pectoral muscle image fusion segmenter module, the muscle tissue three-dimensional feature principal component fusion model module, and the heterogeneous pectoral muscle multi-index grading model module. Based on the parameter type and identifier code in the request, the unit retrieves the corresponding parameter data from the data classification and storage unit and transmits the retrieved data to the corresponding module according to the request format.

[0018] Furthermore, the duck breast muscle parameter acquisition module includes a sample fixation unit, a parameter detection unit, a data conversion unit, and a data transmission unit. The sample fixation unit fixes the position of the duck breast muscle sample to be tested by adjusting the sample's posture using an adjustable clamp to maintain a preset distance and angle between the sample's detection surface and the detection probe of the parameter detection unit, ensuring that the sample position does not shift during the detection process. The parameter detection unit uses a muscle fiber measuring instrument, a near-infrared spectrometer, a pH meter, a shear force meter, and a moisture meter to detect the sample's muscle fiber diameter, intramuscular fat content, pH value, shear force value, and moisture content, respectively, to obtain the raw detection signals of different parameters. The data conversion unit receives the raw detection signals output by the parameter detection unit, filters the signals to remove noise interference, converts the filtered analog signals into digital signals, and quantizes the digital signals to generate parameter values. The data transmission unit receives the parameter values ​​output by the data conversion unit and transmits the parameter values ​​to the poultry muscle parameter management and analysis platform via wired communication, while encrypting the transmitted data to prevent tampering or leakage during transmission.

[0019] A method for measuring breast muscle in broiler ducks includes the following steps: First, the sample fixing unit in the broiler duck breast muscle parameter acquisition module adjusts the posture and fixes the position of the sample breast muscle to be tested, ensuring the sample detection surface is in a preset relative position with different detection probes. The parameter detection unit is then activated to detect the muscle fiber diameter, intramuscular fat content, pH value, shear force value, and moisture content of the sample, acquiring the raw detection signals for different parameters and transmitting them to the data conversion unit. Second, the data conversion unit filters the received raw detection signals to eliminate noise signals caused by environmental interference, converts the filtered analog signals into digital signals, performs quantization calculations on the digital signals to generate specific parameter values, and transmits the encrypted parameter values ​​to the poultry muscle muscle parameter management and analysis platform via the data transmission unit. Third, the parameter receiving unit of the poultry muscle muscle parameter management and analysis platform verifies the format of the transmitted parameter values, transmitting the verified parameter values ​​to the data classification and storage unit for storage according to parameter type. The process involves several steps: First, a unique identifier is added. The parameter correlation analysis unit calculates the correlation degree of the stored parameter data to generate a parameter correlation table. Second, the multimodal breast muscle image fusion segmenter module receives multi-source images of duck breast muscles, performs pixel-level registration and feature layer fusion processing to generate a global segmentation image. The data retrieval unit of the poultry muscle muscle parameter management and analysis platform retrieves the corresponding parameter data and transmits it to the model according to the request of the muscle muscle three-dimensional feature principal component fusion model module. Third, the muscle muscle three-dimensional feature principal component fusion model module performs feature extraction and principal component dimensionality reduction processing on the received parameter data to generate a muscle muscle three-dimensional feature fusion vector, which is then transmitted to the heterogeneous breast muscle multi-index grading model module. This grading model receives the segmented image features and feature vectors and performs multi-dimensional index correlation analysis. Fourth, the heterogeneous breast muscle multi-index grading model module determines the grade of duck breast muscles based on the analysis results. The parameter result output module receives the grade results and the original parameter data, integrates the data, converts the format, and outputs the final duck breast muscle measurement results.

[0020] Beneficial effects: The present invention provides a system and method for measuring the breast muscle of meat ducks. The multimodal breast muscle image fusion and segmentation device can integrate multi-source images of visible light, near infrared, and ultrasonic tomography. In cooperation with the meat duck breast muscle parameter acquisition module, it comprehensively collects parameters such as muscle fiber diameter and fat content. Then, through the poultry meat quality parameter management and analysis platform, it realizes the classified storage, associated mapping, and call management of image and parameter data, effectively solving the problem of fragmentation of multi-source data and difficulty in联动 processing in the prior art, providing comprehensive data support for subsequent analysis, and ensuring the integrity of the measurement results; The three-dimensional feature principal component fusion model of meat quality extracts and reduces the dimension of parameter data, combines the heterogeneous breast muscle multi-index grading model for the collaborative analysis of multimodal image features and meat quality parameters, as well as scientific index weight allocation and grading threshold calculation, solving the problems of insufficient adaptability of existing analysis models and deviation of grading results, and improving the accuracy of quality grading; The cooperation of each module and the parameter result output module realizes the full-process systematic operation from data acquisition, analysis to result output, replacing traditional single equipment and manual analysis, greatly improving the measurement efficiency and accuracy, meeting the refined requirements of the industry for the quality control of meat duck breast muscles, and helping the large-scale breeding industry optimize the production process and ensure the value of meat products. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a diagram of the system module composition of the present invention;

[0022] Figure 2 It is a flowchart of the method steps of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0023] It should be noted that, without conflict, the embodiments in this application and the features in the embodiments can be combined with each other. The following will further describe this application in detail with reference to the drawings and specific embodiments.

[0024] As Figure 1 shown, a system for measuring the breast muscle of meat ducks includes: a multimodal breast muscle image fusion and segmentation device module, a three-dimensional feature principal component fusion model module of meat quality, a heterogeneous breast muscle multi-index grading model module, a poultry meat quality parameter management and analysis platform, a meat duck breast muscle parameter acquisition module, and a parameter result output module;

[0025] The multimodal breast muscle image fusion and segmentation device module receives visible light images, near infrared images, and ultrasonic tomography images of the meat duck breast muscle, and generates a global segmentation image of the meat duck breast muscle through pixel-level registration and feature layer fusion processing of multi-source images;

[0026] Specifically, the multimodal pectoral muscle image fusion and segmentation module fuses and segments multi-source images of duck pectoral muscles. The process first receives a visible light image with a resolution of 1920×1080 pixels, a near-infrared image with a spectral range of 900-1700nm, and an ultrasound tomography image with a scanning depth of 0-50mm. Pixel-level registration technology is used to control the spatial coordinate error of the three images within ±0.1mm. Then, a feature layer fusion algorithm is used to extract edge features, texture features, and structural features from each image. For edge feature extraction, a gradient threshold of 15-25 is set. Texture feature analysis uses gray-level co-occurrence matrix parameters for calculation. Structural features focus on the orientation and distribution patterns of pectoral muscle fibers. During the fusion process, weight allocation ensures that the visible light image accounts for 40%, the near-infrared image for 35%, and the ultrasound tomography image for 25%. Finally, a full-domain segmented image is generated with a segmentation accuracy of over 95%. This module solves the problem of incomplete information from a single image, providing complete image data support for subsequent muscle mass parameter correlation analysis, ensuring accurate visual representation of the structural characteristics of duck pectoral muscles.

[0027] The duck breast muscle parameter acquisition module collects parameters such as muscle fiber diameter, intramuscular fat content, pH value, shear force value, and moisture content of duck breast muscle samples in real time and transmits the collected data to the poultry muscle parameter management and analysis platform.

[0028] Specifically, the implementation of the duck breast muscle parameter acquisition module involves fixing the duck breast muscle samples to be tested. The sample size is set to 10-15cm in length, 8-12cm in width, and 3-5cm in height to ensure the integrity of the detection area. During the implementation, a muscle fiber measuring instrument is used to collect the muscle fiber diameter, with the measurement accuracy controlled within ±1μm. 30-50 collection points are used per sample, and the average value is taken as the final data. A near-infrared spectrometer is used to collect the intramuscular fat content, with a measurement error not exceeding ±0.1%, and the detection time is 5-8 seconds per sample. A pH meter is used to collect the pH value, with a measurement accuracy of ±0. The data was collected at 0.01 kgf, with a 10-minute interval, for a total of 3 collections, and stable values ​​were taken. Shear force was collected using a shear force meter with a measurement accuracy of ±0.1 kgf and a shearing speed of 200 mm / min. Moisture content was collected using a moisture meter with a measurement error not exceeding ±0.5% and a detection temperature controlled at 105℃±2℃. All parameter collection processes were conducted in an environment with a room temperature of 25℃±2℃ and a humidity of 50%±5%. This module obtains accurate and comprehensive data on duck breast muscle parameters, providing basic data support for subsequent analysis and grading, and ensuring the accuracy of the measurement results.

[0029] The poultry muscle muscle parameter management and analysis platform classifies, stores, and maps the received parameter data, and then transmits the image data to the multimodal pectoral muscle image fusion segmenter module and the parameter data to the muscle muscle three-dimensional feature principal component fusion model module, respectively.

[0030] Specifically, the poultry muscle muscle parameter management and analysis platform receives real-time parameter data transmitted from the duck breast muscle parameter acquisition module. The data transmission rate is set at 10-20 Mbps to ensure real-time data transmission. During implementation, the parameter data is first classified and stored, establishing five independent databases based on muscle fiber diameter, intramuscular fat content, pH value, shear force value, and moisture content. Each database is designed with a storage capacity of over 1000 GB, supporting the storage of at least 100,000 sample data entries. During storage, each data entry is labeled with identification information such as acquisition time, sample number, and testing equipment number to facilitate data traceability. Then, association mapping processing is performed to establish the association between image data and parameter data. The association accuracy must ensure that the sample matching error is 0. Subsequently, based on the call requests from the multimodal breast muscle image fusion segmenter and the muscle muscle three-dimensional feature principal component fusion model, data retrieval and transmission are completed within 1-2 seconds. This module realizes the systematic management of parameter data, ensures data storage security and efficient retrieval, and provides data flow support for the collaborative work of various modules.

[0031] The three-dimensional feature principal component fusion model module of the sarcoplasm extracts features and performs principal component dimensionality reduction on the received parameter data to generate a three-dimensional feature fusion vector of the sarcoplasm and transmits it to the heterogeneous pectoral muscle multi-index grading model module.

[0032] Specifically, the implementation of the three-dimensional feature principal component fusion model of muscle tissue receives duck breast muscle parameter data transmitted from the poultry meat muscle parameter management and analysis platform. This data includes muscle fiber diameter (10-100 μm), intramuscular fat content (1%-8%), pH value (5.5-6.8), shear force value (1.5-8.0 kgf), and moisture content (60%-80%). During implementation, feature extraction is performed on each parameter. Muscle fiber diameter feature extraction focuses on fiber length and diameter distribution density; intramuscular fat content feature extraction focuses on fat particle size and distribution uniformity; pH ​​value feature extraction emphasizes the rate of change and stable range; shear force value feature extraction emphasizes the peak value and change curve; and moisture content feature extraction focuses on the spatial distribution differences. Then, principal component dimensionality reduction reduces the original 5-dimensional parameter data to a 3-dimensional feature vector. During dimensionality reduction, the cumulative principal component contribution rate must reach over 90% to ensure the retention of key feature information. This module simplifies the parameter data dimensions, reduces subsequent analysis and computation, and strengthens the correlation between parameters, providing high-value feature data for the multi-index grading model of heterogeneous breast muscle.

[0033] The heterogeneous pectoral muscle multi-index grading model module receives the segmented image features output by the multimodal pectoral muscle image fusion segmenter module and the feature vector output by the three-dimensional muscle feature principal component fusion model module, and performs multi-dimensional index correlation analysis and grade determination processing.

[0034] Specifically, the implementation of the heterogeneous pectoral muscle multi-index grading model receives the segmented image features output by the multimodal pectoral muscle image fusion segmenter and the feature vector output by the principal component fusion model of the three-dimensional features of the muscle tissue. During implementation, the segmented image features are first quantized, converting the image's edge sharpness, texture complexity, and structural integrity into quantized values ​​from 0 to 100. Then, weights are assigned to each dimension of the feature vector, with muscle fiber-related features accounting for 30%, fat content-related features accounting for 25%, pH-related features accounting for 15%, and shear force-related features accounting for 15%. Moisture content-related characteristics account for 15%, and the weighting is determined using the analytic hierarchy process (AHP) to ensure compliance with the quality standards for duck breast muscle. Subsequently, multi-dimensional index correlation analysis is performed to calculate a comprehensive score, with the score range set from 0 to 100 points. Based on the score, five levels are defined: 90-100 points is Level 1, 80-89 points is Level 2, 70-79 points is Level 3, 60-69 points is Level 4, and below 60 points is Level 5. This module enables precise grading of duck breast muscle quality, providing a clear basis for subsequent result output and meeting the industry's needs for meat quality classification.

[0035] The parameter result output module receives the judgment results output by the heterogeneous pectoral muscle multi-index grading model module and the original parameter data stored in the poultry muscle muscle parameter management and analysis platform. After data integration and format conversion, it outputs the measurement results of duck pectoral muscle.

[0036] Specifically, the parameter result output module receives the grading results from the heterogeneous pectoral muscle multi-index grading model and the raw parameter data stored in the poultry muscle muscle parameter management and analysis platform. During implementation, the two types of data are first integrated. The grading results are matched with the corresponding raw parameter data according to the sample number. The integration process must ensure a 100% data matching accuracy rate. Then, the format is converted to PDF and Excel formats. The PDF format is used to visually display the grading results and key parameters, while the Excel format is used to fully present all raw data. The conversion time is controlled within 3-5 seconds. The results are then output via wired printing or wireless transmission. The printing resolution is set to 300 dpi to ensure clear and legible text and data. The wireless transmission distance supports 0-50 meters, with a transmission rate of no less than 5 Mbps. This module presents the analysis results in an intuitive and convenient way, facilitating viewing, archiving, and use by staff, ensuring a complete closed loop in the measurement work and meeting the industry's actual needs for the application of measurement results.

[0037] Preferably, the multimodal pectoral muscle image fusion segmenter module performs image fusion segmentation using the following formula:

[0038]

[0039] in, The pixel values ​​of the merged image of the duck breast muscle. Image pixel coordinates, The fusion weights for visible light images, near-infrared images, and ultrasound tomography images are respectively... These are the pixel values ​​of a visible light image. These are the pixel values ​​of the near-infrared image. These are the pixel values ​​of the ultrasound tomography image. These are the feature enhancement functions for the three types of images; the three-dimensional feature principal component fusion model module for muscle tissue uses the following formula for feature fusion: ,in, This is a fusion vector of three-dimensional features of the muscle tissue. The number of parameters for the breast muscle of meat ducks. For the first Principal component weights of various parameters For the first Types of parameters Principal component analysis results, For the first A three-dimensional feature mapping matrix of various parameters.

[0040] Specifically, in the multimodal pectoral muscle image fusion and segmentation process, the fusion weights of the three types of images are first determined: visible light image weight is set to 0.4, near-infrared image weight to 0.35, and ultrasound tomography image weight to 0.25, ensuring that the information from the three images is fused in a reasonable proportion. Then, feature enhancement processing is performed on each image separately. Visible light image feature enhancement focuses on brightness and contrast optimization, near-infrared image feature enhancement emphasizes the grayscale difference between fat and muscle regions, and ultrasound tomography image feature enhancement focuses on clarifying the contours of deep muscle structures. Feature enhancement functions are used to improve the recognition of key information in each image, ultimately generating a fused image that ensures that pixel values ​​accurately reflect the multidimensional features of the pectoral muscle. In the principal component fusion of three-dimensional features of muscle tissue, the number of parameters of duck breast muscle was first determined to be 5 types, corresponding to parameters such as muscle fiber diameter and intramuscular fat content. Principal component weights were assigned to each type of parameter, with muscle fiber diameter weighted at 0.25, intramuscular fat content at 0.25, pH value at 0.2, shear force value at 0.15, and moisture content at 0.15. Then, principal component analysis was performed on each type of parameter to extract the principal components that reflect the core features of the parameter. Combined with the three-dimensional feature mapping matrix, the parameter features were transformed into three-dimensional vector form to ensure that the fused vector can comprehensively cover the key information of various parameters, improve the image fusion accuracy and the effectiveness of parameter feature fusion, and provide a more reliable data foundation for subsequent grading.

[0041] Preferably, the principal component fusion model module for the three-dimensional features of the skin texture uses the following formula for principal component fusion: ,in, Principal component fusion value, The number of samples for a single parameter. For the first The parameter values ​​of each sample, For the first The mean of the parameter category of each sample. For the first The principal component contribution rate of each sample; the heterogeneous pectoral muscle multi-index grading model module uses the following formula for grading: Grade Among them, Grade represents the grading results of the duck breast muscle. For the number of indicators, For the first The characteristic values ​​of each indicator For the first The weight of each indicator, The total number of levels, This is the floor function.

[0042] Specifically, when implementing principal component fusion of three-dimensional features of muscle tissue, the sample size for each parameter is first determined, with 100-200 samples per parameter category to ensure the reliability of statistical analysis. The deviation of each sample parameter value from the mean of its category is calculated. Combined with the principal component contribution rate of that sample (the contribution rate ranges from 0.1 to 0.3, allocated according to the importance of the principal components), the sum of the products of the deviation and the contribution rate is calculated and then divided by the square root of the sum of the squares of the contribution rates of all samples to obtain the principal component fusion value, ensuring that the fusion value can highlight the influence of key samples. In the multi-index grading of heterogeneous pectoral muscles, the sample size is first determined... The number of indicators involved in the grading is 8 to 10, covering image features and parametric features. Each indicator feature value is assigned a weight (the total weight is 1, allocated according to the degree of influence of the indicator on quality, such as 0.2 for muscle fiber structure features and 0.18 for fat distribution features). The sum of the products of all indicator feature values ​​and their corresponding weights is calculated, and the maximum value of the product sum is found. The sum of the products of the current sample is divided by the maximum value and then multiplied by the total number of gradings (the total number of gradings is set to 5). Finally, the grading result is obtained by rounding down. The principal component fusion calculation logic and grading judgment rules are optimized to improve the scientificity and consistency of the grading results.

[0043] Preferably, the heterogeneous pectoral muscle multi-index grading model module uses the following formula for multi-index analysis: ,in, The scoring is based on a combination of multiple indicators. The number of indicators used in the scoring. For the first The scoring coefficients for each indicator For the first The actual measured value of each indicator For the first The minimum value of each indicator. For the first The maximum value of each indicator; the multimodal pectoral muscle image fusion segmenter module uses the following formula for segmentation: ,in, For the segmentation results, 1 represents the pectoral muscle region, and 0 represents the non-pectoral muscle region. The segmentation threshold is... To merge image pixel values.

[0044] Specifically, in the comprehensive scoring calculation for the multi-indicator grading of heterogeneous pectoral muscles, the number of indicators involved in the scoring is first determined to be 6-8. A scoring coefficient is assigned to each indicator (the coefficient ranges from 0.8 to 1.2, adjusted according to the importance of the indicator; for example, the shear force value scoring coefficient is 1.1, and the water content scoring coefficient is 0.9). The difference between the actual measured value and the minimum value of each indicator is calculated, and this difference is divided by the difference between the maximum and minimum values ​​to obtain the normalized value of that indicator. The product of the scoring coefficient and the normalized value is then added to 1. Finally, the calculation results of all indicators are multiplied to obtain the comprehensive score. To ensure that the scoring reflects the synergistic effect of each indicator, when implementing multimodal pectoral muscle image fusion and segmentation, a segmentation threshold is first determined. Through multiple experiments, the threshold range is set to 80-120 (the specific value is determined based on the image grayscale distribution). Each pixel value of the fused image is compared with the threshold. Pixel values ​​greater than the threshold are determined to be pectoral muscle regions, and those less than or equal to the threshold are determined to be non-pectoral muscle regions. Binarized segmentation results are generated to ensure that the segmentation can accurately distinguish between pectoral muscle and background regions. The comprehensive scoring calculation method and image segmentation judgment criteria are improved to enhance the accuracy of scoring and segmentation precision.

[0045] Preferably, the poultry muscle tissue parameter management and analysis platform uses the following formula for parameter correlation: Where Corr is the correlation coefficient between the two parameters. The total number of samples, For the first The first parameter value of each sample The mean of the first parameter, For the first The second parameter value of each sample, The mean of the second parameter is used; the three-dimensional feature principal component fusion model module of the sarcoplasm generates feature vectors using the following formula: ,in, For feature vectors, The number of feature dimensions. For the first Weights of each dimension For the first Feature values ​​in each dimension.

[0046] Specifically, in the parameter correlation calculation of the poultry muscle muscle parameter management and analysis platform, the total number of samples is first determined to be 500-1000, covering duck breast muscle samples from different breeding conditions. The deviation of the first parameter value from the mean of the first parameter and the deviation of the second parameter value from the mean of the second parameter for each sample are calculated. The products of the two sets of deviations are summed and then divided by the product of the squares and square roots of the two sets of deviations to obtain the correlation coefficient (the correlation coefficient ranges from -1 to 1, with positive values ​​indicating positive correlation and negative values ​​indicating negative correlation), ensuring that it can accurately reflect the degree of correlation between the two parameters. In the feature vector generation of the principal component fusion of muscle muscle three-dimensional features, the number of feature dimensions is first determined to be 3-5, with each dimension corresponding to a core feature. Weights are assigned to each dimension (weights range from 0.2-0.4, allocated according to the importance of the dimension). The sum of the products of each dimension weight and the corresponding feature value is calculated and then divided by the sum of all dimension weights to obtain the feature vector, ensuring that the feature vector can condense key feature information. The parameter correlation analysis method and feature vector generation rules are established to provide a clear basis for data correlation application and feature extraction.

[0047] Preferably, the poultry muscle tissue parameter management and analysis platform uses the following formula for data storage mapping: ,in, For parameters The storage address mapping value, This represents the number of sub-parameters after parameter decomposition. For the first The mapping coefficients of each sub-parameter, For hash functions, For the first Sub-parameters; the heterogeneous pectoral muscle multi-index grading model module uses the following formula to determine the rank: Rank Where Rank is the level determination value. For the number of indicators, For the first The weighting of each indicator, For the first The measured values ​​of each indicator, For the first The lower limit of each indicator, For the first The upper limit of each indicator.

[0048] Specifically, in implementing the storage address mapping of the poultry muscle muscle parameter management and analysis platform, each parameter is first decomposed into 3-5 sub-parameters (e.g., muscle fiber diameter is decomposed into average diameter, maximum diameter, minimum diameter, etc.). A mapping coefficient is assigned to each sub-parameter (coefficients range from 0.15 to 0.3, allocated according to the importance of the sub-parameter). A hash operation is performed on each sub-parameter (a simplified version of the SHA-256 algorithm is used to ensure computational efficiency). The sum of the products of the mapping coefficient and the hash value is calculated to obtain the storage address mapping value, ensuring that each parameter can be accurately mapped to its corresponding storage address. This process is also applied to the multi-index grading of heterogeneous pectoral muscle. In the grading process, the number of indicators is first determined to be 7 to 9. A grading weight is assigned to each indicator (the total weight is 1, such as pH value 0.15 and muscle fiber density 0.2). The difference between the measured value of each indicator and its lower limit is calculated and divided by the difference between the upper limit and the lower limit of the indicator to obtain the indicator normalization result. The grading weight is multiplied by the normalization result, and then the calculation results of all indicators are summed to obtain the grading value (the grading value ranges from 0 to 1, corresponding to the division interval of 5 grades). The parameter storage mapping mechanism and grading calculation method are optimized to improve data storage efficiency and the accuracy of grading.

[0049] Preferably, the heterogeneous pectoral muscle multi-index grading model module includes an index feature extraction unit, a multi-dimensional weight allocation unit, a grading threshold calculation unit, and a grade output unit. The index feature extraction unit receives the segmented image output by the multimodal pectoral muscle image fusion segmenter module and the feature vector output by the muscle tissue three-dimensional feature principal component fusion model module. It extracts edge and texture features from the segmented image, performs dimensional feature parsing on the feature vector, and integrates the extracted image features and the parsed vector features into a multi-index feature set. The multi-dimensional weight allocation unit receives the multi-index feature set output by the index feature extraction unit and, based on the influence of different features in the quality determination of duck pectoral muscle, performs hierarchical weight allocation. The analysis method calculates the weight coefficients corresponding to different features and generates a weight allocation matrix. The grading threshold calculation unit receives the weight allocation matrix output by the multi-dimensional weight allocation unit and the historical parameter data transmitted by the poultry muscle muscle parameter management and analysis platform. It performs statistical analysis on the historical parameter data and determines the feature threshold range corresponding to different levels by combining the weight allocation matrix, and generates a grading threshold table. The level output unit receives the grading threshold table output by the grading threshold calculation unit and the multi-indicator feature set output by the indicator feature extraction unit. It compares the different feature values ​​in the multi-indicator feature set with the corresponding thresholds in the grading threshold table, determines the final level of the duck breast muscle based on the comparison results, and transmits it to the parameter result output module.

[0050] Specifically, the heterogeneous pectoral muscle multi-index grading model module includes four units. The index feature extraction unit receives the segmented image output by the multimodal pectoral muscle image fusion segmenter and the feature vector output by the muscle tissue 3D feature principal component fusion model. It then uses an edge detection algorithm to extract edge features from the segmented image, setting the gradient threshold for edge detection to 18-22 to ensure accurate capture of the pectoral muscle contour edges. The gray-level co-occurrence matrix method is used to extract texture features, calculating parameters such as texture energy and entropy. The energy value is controlled within the range of 0.01-0.1, and the entropy value is controlled within the range of 1-5 to distinguish different muscle textures. The feature vector is parsed by dimension, extracting the feature peaks and distribution intervals of each dimension, and integrating the image features and vector features into a multi-index feature set containing 8-10 feature terms. After receiving the feature set, the multi-dimensional weight allocation unit constructs a judgment matrix using the analytic hierarchy process (AHP). It compares the importance of each feature pairwise, ensuring the consistency ratio of the judgment matrix is ​​below 0.1 to guarantee the rationality of the weight allocation. This results in a weight allocation matrix with clearly defined weight proportions for each feature, where muscle fiber-related features account for 28%–32% and fat-related features account for 23%–27%. The grading threshold calculation unit receives the weight allocation matrix and 1000–2000 historical parameter data points transmitted from the poultry muscle tissue parameter management and analysis platform. It categorizes and statistically analyzes the historical data by feature, calculating the mean and standard deviation for each category. Combining this with the weight allocation matrix, it determines the threshold range for each level: Level 1 threshold is set at 85%–100% of the feature's comprehensive value, Level 2 at 70%–84%, and so on, generating a grading threshold table. The grading output unit receives the threshold table and the feature set, compares each feature value against the threshold table, completes the grading determination within 10 seconds, and transmits the result to the parameter output module. This collaborative approach among the units ensures accurate grading, solving the problem of traditional grading relying on experience.

[0051] Preferably, the poultry muscle muscle parameter management and analysis platform includes a parameter receiving unit, a data classification and storage unit, a parameter correlation analysis unit, and a data retrieval unit. The parameter receiving unit receives parameter data such as muscle fiber diameter, intramuscular fat content, pH value, shear force value, and moisture content transmitted from the duck breast muscle parameter acquisition module. It performs format verification on the parameter data, removes data with incorrect formats, and transmits the verified data to the data classification and storage unit. The data classification and storage unit receives the verified data output from the parameter receiving unit, establishes different data storage directories according to parameter types, and stores the different parameter data in order of acquisition time. The system retrieves data from the corresponding directory and adds a unique identifier code to each data entry. The parameter correlation analysis unit receives the stored data output by the data classification and storage unit, calculates the Pearson correlation coefficient between different parameters, analyzes the degree of correlation between different parameters, and generates a parameter correlation table. The data retrieval unit receives data retrieval requests from the multimodal pectoral muscle image fusion segmenter module, the muscle tissue three-dimensional feature principal component fusion model module, and the heterogeneous pectoral muscle multi-index grading model module. Based on the parameter type and identifier code in the request, the unit retrieves the corresponding parameter data from the data classification and storage unit and transmits the retrieved data to the corresponding module according to the request format.

[0052] Specifically, the poultry muscle parameter management and analysis platform comprises four units. The parameter receiving unit receives parameter data such as muscle fiber diameter and intramuscular fat content transmitted from the duck breast muscle parameter acquisition module. The data transmission protocol uses TCP / IP, with a baud rate set to 9600-115200bps to ensure real-time data reception. The unit verifies the parameter format and ensures that parameter values ​​are within a preset reasonable range (e.g., muscle fiber diameter 10-100μm, pH value 5.5-6.8), removing abnormal data outside the range. The verification pass rate must reach over 98%, and qualified data is then transmitted to the data classification and storage unit. The data classification and storage unit establishes independent storage directories according to parameter type. Each directory uses the FAT32 file system, with a maximum file size of 4GB. It supports data sorting and storage by acquisition timestamp (accurate to the second), and adds an 18-bit unique identifier code (including device number, sample number, and acquisition time information) to each data entry for easy data traceability. The parameter correlation analysis unit retrieves stored data and uses the Pearson correlation coefficient calculation method to perform correlation analysis on 100-200 sets of sample data for any two parameters. The correlation coefficient calculation result is retained to four decimal places, with a value range of -1 to 1, generating a parameter correlation table containing the pairwise correlation degrees of all parameters. After receiving the call requests from each module, the data retrieval unit parses the parameter type and identification code in the request, retrieves the data from the corresponding storage directory within 0.5 to 1.5 seconds, encapsulates it according to the request format (such as JSON or XML), and transmits it. Through these units, efficient parameter management is achieved, solving the problems of chaotic data storage and inefficient retrieval in traditional methods.

[0053] Preferably, the duck breast muscle parameter acquisition module includes a sample fixing unit, a parameter detection unit, a data conversion unit, and a data transmission unit. The sample fixing unit fixes the position of the duck breast muscle sample to be tested by adjusting the sample's posture using an adjustable clamp, ensuring that the sample's detection surface maintains a preset distance and angle with the detection probe of the parameter detection unit, thus preventing sample position shift during detection. The parameter detection unit uses a muscle fiber measuring instrument, a near-infrared spectrometer, a pH meter, a shear force meter, and a moisture meter to detect the sample's muscle fiber diameter, intramuscular fat content, pH value, shear force value, and moisture content, respectively, acquiring raw detection signals for different parameters. The data conversion unit receives the raw detection signals output by the parameter detection unit, filters the signals to remove noise interference, converts the filtered analog signals into digital signals, and quantizes the digital signals to generate parameter values. The data transmission unit receives the parameter values ​​output by the data conversion unit and transmits them to the poultry muscle parameter management and analysis platform via wired communication, while simultaneously encrypting the transmitted data to prevent tampering or leakage during transmission.

[0054] Specifically, the duck breast muscle parameter acquisition module includes four units. The sample fixation unit uses an adjustable mechanical clamp to fix the duck breast muscle sample to be tested. The clamping force is set to 5-10N to avoid sample damage. The sample posture is adjusted by the clamp adjustment knob to control the vertical distance between the sample detection surface and each detection probe to 2-3cm, and the horizontal deviation to no more than ±0.5mm, to ensure accurate detection position. In the parameter detection unit, the muscle fiber measuring instrument uses an optical microscope for imaging, with a magnification set to 200-400x, and measures the muscle fiber diameter using image analysis software, with a measurement accuracy of ±1μm; the near-infrared spectrometer has a spectral scanning range set to 900-1700nm and a scanning interval of 1nm, and collects intramuscular fat content data, with a measurement error of ±0.1%; the pH meter uses a glass electrode, with a measurement range of 0-14 and an accuracy of ±0.01, and collects pH values ​​at 3-5 points in different areas of the sample; the shear force meter has a shear probe speed set to 200mm / min, a shear force measurement range of 0-10kgf, and an accuracy of ±0.1kgf; the moisture analyzer uses the loss-on-drying method, with a heating temperature set to 105℃±2℃ and a drying time of 3-4 hours, and a moisture content measurement error of ±0.5%. The data conversion unit uses a low-pass filter to filter the raw detection signal, with a cutoff frequency set to 100Hz to remove high-frequency noise. An A / D converter then converts the analog signal into a 16-bit digital signal with a quantization error of ±1LSB. The digital signal is then scaled to generate specific parameter values. The data transmission unit uses RS485 bus communication, with a transmission distance of 0-100 meters and a transmission rate of 10-20Mbps. Transmitted data is encrypted using a CRC32 checksum algorithm to ensure data transmission security. Through these units, accurate parameter acquisition is achieved, solving the problems of large errors and poor stability in traditional data acquisition methods.

[0055] The multimodal pectoral muscle image fusion segmenter is a multi-source image pixel-level registration and feature layer fusion algorithm. Its implementation process is as follows: First, it receives a visible light image with a resolution of 1920×1080 pixels, a near-infrared image with a spectral range of 900-1700nm, and an ultrasound tomographic image with a scanning depth of 0-50mm. Pixel-level registration technology is used to control the spatial coordinate errors of the three images within ±0.1mm. Then, a feature layer fusion algorithm is employed, focusing on brightness and contrast optimization for the visible light image, enhancing the gray-level differences between fat and muscle regions in the near-infrared image, and clarifying the deep muscle structure contours in the ultrasound tomographic image. Simultaneously, the weights are set as follows: visible light image 40%, near-infrared image 35%, and ultrasound tomographic image 25%. Edges (gradient threshold 15-25), texture (gray-level co-occurrence matrix calculation), and structural features of each image are extracted and fused, ultimately generating a full-domain segmentation image with a segmentation accuracy of over 95%. The function of this module is to integrate multi-source image information, breaking the limitations of single image information and providing complete visual data support for subsequent muscle mass parameter correlation analysis. This study addresses the problem that traditional single images cannot fully reflect the structural characteristics of duck breast muscles, laying the foundation for accurate determination of duck breast muscle quality and improving the comprehensiveness and accuracy of subsequent analysis.

[0056] The three-dimensional feature principal component fusion model for muscle tissue combines feature extraction and principal component dimensionality reduction algorithms. In implementation, it first receives parameter data from a poultry muscle tissue parameter management and analysis platform, including muscle fiber diameter (10-100 μm), intramuscular fat content (1%–8%), pH value (5.5-6.8), shear force value (1.5-8.0 kgf), and moisture content (60%–80%). The feature extraction algorithm extracts features such as the length and distribution density of muscle fiber diameter, the particle size and uniformity of intramuscular fat content, the rate of change and stable range of pH value, the peak value and curve of shear force value, and the spatial distribution differences of moisture content. Then, the principal component dimensionality reduction algorithm reduces the original 5-dimensional parameter data to a 3-dimensional feature vector, ensuring that the cumulative contribution rate of principal components reaches over 90%, while retaining key feature information. The model simplifies the dimensionality of parameter data, reduces the computational load of subsequent analysis, and strengthens the correlation between parameters. This approach addresses the issues of high dimensionality, computational complexity, and weak parameter correlation in traditional parameter data, providing high-value feature data for multi-index grading models of heterogeneous pectoral muscles, thereby helping to improve the analysis efficiency and judgment accuracy of grading models.

[0057] The heterogeneous pectoral muscle multi-index grading model comprises a feature quantization algorithm, a hierarchical analysis weight allocation algorithm, and a multi-dimensional index correlation analysis algorithm. The implementation process is as follows: First, the segmented image and the fused vector of the muscle's three-dimensional features are received. The feature quantization algorithm converts the image edge sharpness, texture complexity, and structural integrity into quantized values ​​from 0 to 100. Then, the hierarchical analysis weight allocation algorithm assigns weights according to the following proportions: muscle fiber-related features 30%, fat content-related features 25%, pH value-related features 15%, shear force-related features 15%, and moisture content-related features 15% (judgment matrix consistency ratio < 0.1). Finally, the multi-dimensional index correlation analysis algorithm calculates a comprehensive score from 0 to 100, grading the muscle into five levels: 90–100 points (Level 1), 80–89 points (Level 2), 70–79 points (Level 3), 60–69 points (Level 4), and below 60 points (Level 5). This model is used for multi-dimensional correlation analysis and precise grading of duck pectoral muscle quality, outputting clear grade determination results. This research addresses the problems of traditional grading methods that rely on experience, use single indicators, and have large judgment biases. It provides a scientific basis for classifying the quality of duck breast muscle, meets the industry's needs for refined quality control and grading of meat products, and helps to enhance the commercial value and market competitiveness of meat products.

[0058] The poultry muscle tissue parameter management and analysis platform comprises a data classification and storage algorithm, a parameter correlation analysis algorithm (Pearson correlation coefficient calculation), and a data retrieval and matching algorithm. In implementation, data is first received from the parameter acquisition module at a rate of 10-20 Mbps. Using the data classification and storage algorithm, five independent databases, each with a capacity of over 1000 GB, are established according to parameter type. Each data entry is labeled with its acquisition time, sample number, etc., and stored in chronological order. Then, the Pearson correlation coefficient calculation algorithm is used to calculate the correlation coefficient (retaining four decimal places, ranging from -1 to 1) for 100-200 sets of sample data for any two parameters, generating a parameter correlation table. When other modules initiate a retrieval request, the data retrieval and matching algorithm retrieves and transmits the data within 1-2 seconds based on the requested parameter type and identifier code. The platform's function is to achieve systematic storage, correlation analysis, and efficient retrieval of parameter data, ensuring smooth data flow. Its significance lies in solving the problems of chaotic data storage, unclear correlation, and inefficient retrieval in traditional systems, providing data support for the collaborative work of various modules, ensuring the efficient operation of the entire measurement system, and providing the possibility for subsequent data traceability, analysis and optimization, thus promoting the development of duck breast muscle measurement technology towards systematization and intelligence.

[0059] like Figure 2As shown, a method for measuring breast muscle in broiler ducks includes the following steps: First, the sample fixing unit in the broiler duck breast muscle parameter acquisition module adjusts the posture and fixes the position of the sample breast muscle to be tested, ensuring the sample detection surface is in a preset relative position with different detection probes. The parameter detection unit is then activated to detect the muscle fiber diameter, intramuscular fat content, pH value, shear force value, and moisture content of the sample, acquiring the original detection signals for different parameters and transmitting them to the data conversion unit. Second, the data conversion unit filters the received original detection signals to eliminate noise signals caused by environmental interference, converts the filtered analog signals into digital signals, quantizes the digital signals to generate specific parameter values, and transmits the encrypted parameter values ​​to the poultry muscle muscle parameter management and analysis platform via the data transmission unit. Third, the parameter receiving unit of the poultry muscle muscle parameter management and analysis platform verifies the format of the transmitted parameter values, transmitting the verified parameter values ​​to the data classification and storage unit, where they are stored according to parameter type. The first step involves storing and adding a unique identifier code. The parameter correlation analysis unit calculates the correlation degree of the stored parameter data to generate a parameter correlation table. The second step involves the multimodal breast muscle image fusion segmenter module receiving multi-source images of duck breast muscles, performing pixel-level registration and feature layer fusion processing to generate a global segmentation image. The data retrieval unit of the poultry muscle muscle parameter management and analysis platform retrieves the corresponding parameter data and transmits it to the model according to the request of the muscle muscle three-dimensional feature principal component fusion model module. The third step involves the muscle muscle three-dimensional feature principal component fusion model module extracting features and performing principal component dimensionality reduction processing on the received parameter data to generate a muscle muscle three-dimensional feature fusion vector, which is then transmitted to the heterogeneous breast muscle multi-index grading model module. This grading model receives the segmented image features and feature vectors and performs multi-dimensional index correlation analysis. The fourth step involves the heterogeneous breast muscle multi-index grading model module determining the grade of duck breast muscles based on the analysis results. The parameter result output module receives the grade results and the original parameter data, integrates the data, converts the format, and outputs the final duck breast muscle measurement results.

[0060] A system and method for measuring breast muscle in broiler ducks are disclosed. The multimodal breast muscle image fusion segmenter can fuse multiple images of broiler duck breast muscle, breaking the information limitations of a single image type. At the same time, the broiler duck breast muscle parameter acquisition module can comprehensively collect key parameters such as muscle fiber diameter and intramuscular fat content. The data acquired by both are classified, stored, mapped, and easily accessed through a poultry muscle muscle parameter management and analysis platform. This completely solves the problems of fragmented multi-source data and inability to process them in conjunction with existing technologies, allowing image features and muscle muscle parameters to form a complete data chain. This provides comprehensive and closely related data support for subsequent analysis, ensuring that the measurement results can cover multi-dimensional information of broiler duck breast muscle and avoiding incomplete measurements due to missing or isolated data.

[0061] In terms of analysis, grading, and process efficiency, the system and method also demonstrate significant advantages, specifically addressing the shortcomings of existing technologies: the principal component fusion model of three-dimensional muscle features can perform in-depth processing of the collected parameter data, extract core features, and optimize data dimensions; the heterogeneous breast muscle multi-index grading model, based on this, combines multimodal image features to conduct collaborative analysis, and by scientifically setting index weights and grading thresholds, solves the problems of poor adaptability and grading result deviation in existing analysis models, making the grading of duck breast muscle quality more in line with actual conditions; at the same time, the entire process from data collection, processing, analysis to result output is systematically operated through the orderly connection of each module, replacing the traditional mode that relies on single equipment and manual operation, greatly improving measurement efficiency and accuracy, meeting the industry's needs for refined quality control of duck breast muscle, and helping large-scale breeding industries optimize production processes.

[0062] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," "link," and "fix" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0063] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A system for measuring breast muscle in meat ducks, characterized in that, include: The multimodal pectoral muscle image fusion segmenter module receives visible light images, near-infrared images, and ultrasound tomographic images of duck breast muscles, and generates a full-domain segmentation image of duck breast muscles through multi-source image pixel-level registration and feature layer fusion processing. The duck breast muscle parameter acquisition module collects parameters such as muscle fiber diameter, intramuscular fat content, pH value, shear force value, and moisture content of duck breast muscle samples in real time, and transmits the collected data to the poultry muscle muscle parameter management and analysis platform. The poultry muscle muscle parameter management and analysis platform classifies, stores, and maps the received parameter data, and then transmits the image data to the multimodal breast muscle image fusion segmenter module and the parameter data to the muscle muscle three-dimensional feature principal component fusion model module. The muscle muscle three-dimensional feature principal component fusion model module performs feature extraction and principal component dimensionality reduction processing on the received parameter data. The heterogeneous pectoral muscle multi-index grading model module receives the segmented image features output by the multimodal pectoral muscle image fusion segmenter module and the feature vector output by the principal component fusion model module of the three-dimensional features of the muscle tissue, and performs multi-dimensional index correlation analysis and grade determination processing. The parameter result output module receives the judgment results output by the heterogeneous pectoral muscle multi-index grading model module and the raw parameter data stored in the poultry muscle muscle parameter management and analysis platform. After data integration and format conversion, it outputs the measurement results of duck pectoral muscle.

2. The system for measuring breast muscle in ducks according to claim 1, characterized in that, The multimodal pectoral muscle image fusion and segmentation module uses the following formula for image fusion and segmentation: ； in, The pixel values ​​of the merged image of the duck breast muscle. Image pixel coordinates, The fusion weights for visible light images, near-infrared images, and ultrasound tomography images are respectively... These are the pixel values ​​of a visible light image. These are the pixel values ​​of the near-infrared image. These are the pixel values ​​of the ultrasound tomography image. These are the feature enhancement functions for the three types of images; the three-dimensional feature principal component fusion model module for muscle tissue uses the following formula for feature fusion: ,in, This is a fusion vector of three-dimensional features of the muscle tissue. The number of parameters for the breast muscle of meat ducks. For the first Principal component weights of various parameters For the first Types of parameters Principal component analysis results, For the first A three-dimensional feature mapping matrix of various parameters.

3. The system for measuring breast muscle in ducks according to claim 1, characterized in that, The principal component fusion model module for the three-dimensional features of the sarcoplasm uses the following formula for principal component fusion: ,in, Principal component fusion value, The number of samples for a single parameter. For the first The parameter values ​​of each sample, For the first The mean of the parameter category of each sample. For the first The principal component contribution rate of each sample; the heterogeneous pectoral muscle multi-index grading model module uses the following formula for grading: Grade Among them, Grade represents the grading results of the duck breast muscle. For the number of indicators, For the first The characteristic values ​​of each indicator For the first The weight of each indicator, The total number of levels, This is the floor function.

4. The system for measuring breast muscle in meat ducks according to claim 1, characterized in that, The heterogeneous pectoral muscle multi-index grading model module uses the following formula for multi-index analysis: ,in, The scoring is based on a combination of multiple indicators. The number of indicators used in the scoring. For the first The scoring coefficients for each indicator For the first The actual measured value of each indicator For the first The minimum value of each indicator. For the first The maximum value of each indicator; the multimodal pectoral muscle image fusion segmenter module uses the following formula for segmentation: ,in, For the segmentation results, 1 represents the pectoral muscle region, and 0 represents the non-pectoral muscle region. The segmentation threshold is... To merge image pixel values.

5. The system for measuring breast muscle in meat ducks according to claim 1, characterized in that, The poultry muscle tissue parameter management and analysis platform uses the following formula for parameter correlation: Where Corr is the correlation coefficient between the two parameters. The total number of samples, For the first The first parameter value of each sample The mean of the first parameter, For the first The second parameter value of each sample, The mean of the second parameter is used; the three-dimensional feature principal component fusion model module of the sarcoplasm generates feature vectors using the following formula: ,in, For feature vectors, The number of feature dimensions. For the first Weights of each dimension For the first Feature values ​​in each dimension.

6. The system for measuring breast muscle in ducks according to claim 1, characterized in that, The poultry muscle tissue parameter management and analysis platform uses the following formula for data storage mapping: ,in, For parameters The storage address mapping value, This represents the number of sub-parameters after parameter decomposition. For the first The mapping coefficients of each sub-parameter, For hash functions, For the first Sub-parameters; the heterogeneous pectoral muscle multi-index grading model module uses the following formula to determine the rank: Rank Where Rank is the level determination value. For the number of indicators, For the first The weighting of each indicator, For the first The measured values ​​of each indicator, For the first The lower limit of each indicator, For the first The upper limit of each indicator.

7. The system for measuring breast muscle in ducks according to claim 1, characterized in that, The heterogeneous pectoral muscle multi-index grading model module includes an index feature extraction unit, a multi-dimensional weight allocation unit, a grading threshold calculation unit, and a grade output unit. The index feature extraction unit receives the segmented image output by the multimodal pectoral muscle image fusion segmenter module and the feature vector output by the muscle tissue three-dimensional feature principal component fusion model module. It extracts edge and texture features from the segmented image, performs dimensional feature parsing on the feature vectors, and integrates the extracted image features and the parsed vector features into a multi-index feature set. The multi-dimensional weight allocation unit receives the multi-index feature set output by the index feature extraction unit and, based on the influence of different features in the quality determination of duck pectoral muscle, uses hierarchical analysis to... The method calculates the weight coefficients corresponding to different features and generates a weight allocation matrix. The grading threshold calculation unit receives the weight allocation matrix output by the multi-dimensional weight allocation unit and the historical parameter data transmitted by the poultry muscle muscle parameter management and analysis platform. It performs statistical analysis on the historical parameter data and determines the feature threshold range corresponding to different levels by combining the weight allocation matrix, and generates a grading threshold table. The level output unit receives the grading threshold table output by the grading threshold calculation unit and the multi-indicator feature set output by the indicator feature extraction unit. It compares the different feature values ​​in the multi-indicator feature set with the corresponding thresholds in the grading threshold table, determines the final level of the duck breast muscle based on the comparison results, and transmits it to the parameter result output module.

8. The system for measuring breast muscle in ducks according to claim 1, characterized in that, The poultry muscle parameter management and analysis platform includes a parameter receiving unit, a data classification and storage unit, a parameter correlation analysis unit, and a data retrieval unit. The parameter receiving unit receives parameter data such as muscle fiber diameter, intramuscular fat content, pH value, shear force value, and moisture content transmitted by the duck breast muscle parameter acquisition module. It performs format verification on the parameter data, removes data with incorrect formats, and transmits the verified data to the data classification and storage unit. The data classification and storage unit receives the verified data output by the parameter receiving unit, establishes different data storage directories according to parameter types, sorts the different parameter data according to the acquisition time order, and stores them in the corresponding directories. At the same time, a unique identifier code is added to each data entry. The parameter correlation analysis unit receives the stored data output by the data classification and storage unit, calculates the Pearson correlation coefficient between different parameters, analyzes the degree of correlation between different parameters, and generates a parameter correlation table. The data retrieval unit receives data retrieval requests sent by the multimodal pectoral muscle image fusion segmenter module, the muscle tissue three-dimensional feature principal component fusion model module, and the heterogeneous pectoral muscle multi-index grading model module. According to the parameter type and identifier code in the request, it retrieves the corresponding parameter data from the data classification and storage unit and transmits the retrieved data to the corresponding module according to the request format.

9. A system for measuring breast muscle in meat ducks according to claim 1, characterized in that, The duck breast muscle parameter acquisition module includes a sample fixation unit, a parameter detection unit, a data conversion unit, and a data transmission unit. The sample fixation unit positions the duck breast muscle sample to be tested by adjusting the sample's posture using an adjustable clamp to maintain a preset distance and angle between the sample's detection surface and the detection probe of the parameter detection unit, ensuring the sample position does not shift during the detection process. The parameter detection unit uses a muscle fiber measuring instrument, a near-infrared spectrometer, a pH meter, a shear force meter, and a moisture meter to detect the sample's muscle fiber diameter, intramuscular fat content, pH value, shear force value, and moisture content, respectively, acquiring raw detection signals for different parameters. The data conversion unit receives the raw detection signals output by the parameter detection unit, filters the signals to remove noise interference, converts the filtered analog signals into digital signals, and quantizes the digital signals to generate parameter values. The data transmission unit receives the parameter values ​​output by the data conversion unit and transmits them to the poultry muscle parameter management and analysis platform via wired communication, while simultaneously encrypting the transmitted data to prevent tampering or leakage during transmission.

10. A method for measuring breast muscle in meat ducks, characterized in that, The process includes the following steps: First, the sample fixing unit in the duck breast muscle parameter acquisition module adjusts the posture and fixes the position of the duck breast muscle sample to be tested, ensuring the sample detection surface is in a preset relative position with different detection probes. The parameter detection unit then detects the muscle fiber diameter, intramuscular fat content, pH value, shear force value, and moisture content of the sample, acquiring the raw detection signals for different parameters and transmitting them to the data conversion unit. Second, the data conversion unit filters the received raw detection signals to eliminate noise signals caused by environmental interference, converts the filtered analog signals into digital signals, quantizes the digital signals to generate specific parameter values, and transmits the encrypted parameter values ​​to the poultry muscle muscle parameter management and analysis platform via the data transmission unit. Third, the parameter receiving unit of the poultry muscle muscle parameter management and analysis platform verifies the format of the transmitted parameter values. Parameter values ​​that pass verification are transmitted to the data classification and storage unit, where they are stored according to parameter type and a unique identifier is added. The encoding and parameter correlation analysis unit calculates the correlation degree of the stored parameter data to generate a parameter correlation table. The fourth step involves the multimodal breast muscle image fusion segmenter module receiving multi-source images of duck breast muscles, performing pixel-level registration and feature layer fusion processing to generate a global segmentation image. The data retrieval unit of the poultry muscle muscle parameter management and analysis platform retrieves the corresponding parameter data and transmits it to the model according to the request from the muscle muscle three-dimensional feature principal component fusion model module. The fifth step involves the muscle muscle three-dimensional feature principal component fusion model module extracting features and performing principal component dimensionality reduction processing on the received parameter data to generate a muscle muscle three-dimensional feature fusion vector, which is then transmitted to the heterogeneous breast muscle multi-index grading model module. This grading model receives the segmented image features and feature vectors and performs multi-dimensional index correlation analysis. The sixth step involves the heterogeneous breast muscle multi-index grading model module determining the grade of the duck breast muscle based on the analysis results. The parameter result output module receives the grade results and the original parameter data, integrates the data, converts the format, and outputs the final duck breast muscle measurement results.

Images