A snowflake beef processing method and device, electronic equipment and medium

Abstract

The application provides a snowflake beef processing method and device, electronic equipment and medium, comprising: when it is judged that the beef image data is the image data of the target snowflake beef through the pre-trained snowflake beef detection model, obtaining the target image data and the target spectrum data of the target snowflake beef obtained after preprocessing; extracting the color feature from the target image data and the wave band feature from the target spectrum data; fusing the color feature and the wave band feature to obtain the fusion feature of the target snowflake beef; inputting the fusion feature of the target snowflake beef into the trained grade prediction model to determine the grade of the target snowflake beef, thereby quickly and automatically detecting the snowflake beef and automatically determining the grade of the snowflake beef.

Description

Technical Field

[0001] This application relates to the field of data processing, and more specifically, to a method, apparatus, electronic device, and medium for processing Wagyu beef. Background Technology

[0002] Yunling cattle are one of the four major beef cattle breeds in my country with completely independent intellectual property rights in Yunnan Province since the founding of the People's Republic of China. They are rich in high-grade Wagyu beef, with meat quality comparable to Kobe beef. However, there is still a lack of an effective automatic grading method for Wagyu beef.

[0003] In recent years, deep learning has been used in image recognition of livestock products such as chicken and mutton. However, deep learning technology still faces challenges in addressing target detection and automatic grading of Yunling beef products, including weak data foundations, insufficient theoretical research, and an incomplete algorithmic methodology. Therefore, overcoming the technical bottlenecks in target detection and automatic grading of Yunling beef is crucial for improving the automation level of Yunling beef processing and manufacturing. Summary of the Invention

[0004] In view of this, the purpose of this application is to provide a method, apparatus, electronic device and medium for processing Wagyu beef, which can quickly and automatically detect Wagyu beef and automatically determine the grade of Wagyu beef.

[0005] This application provides a method for processing Wagyu beef, including the following steps:

[0006] When the beef image data is determined to be the target marbled beef image data by the pre-trained marbled beef detection model, the target image data and target spectral data of the target marbled beef after preprocessing are obtained.

[0007] Color features are extracted from the target image data, and band features are extracted from the target spectral data;

[0008] By fusing the color features and band features, the fused features of the target snowflake beef are obtained;

[0009] The fusion features of the target marbled beef are input into a trained grade prediction model to determine the grade of the target marbled beef.

[0010] In some embodiments, the method for processing marbled beef further includes, before acquiring the preprocessed target image data and target spectral data of the target marbled beef:

[0011] Acquire beef image data;

[0012] The beef image data is input into the trained marbled beef detection model to extract the marble pattern and color features from the beef image data;

[0013] The trained snowflake beef detection model determines whether the snowflake beef region in the beef image data meets the preset conditions based on the marble pattern features and color features.

[0014] If so, then the beef image data is determined to be the image data of the target marbled beef.

[0015] In some embodiments, the method for processing marbled beef includes obtaining target image data and target spectral data of the target marbled beef after preprocessing, including:

[0016] Acquire the image data and spectral data of the target marbled beef;

[0017] The image data of the target marbled beef is preprocessed to obtain target image data;

[0018] The target spectral data of the target marbled beef is preprocessed to obtain the target spectral data.

[0019] In some embodiments, the method for processing marbled beef includes extracting color features from the target image data, including:

[0020] The brightness values ​​of each pixel in the red component, the green component, and the blue component of the target image data are extracted respectively as color features.

[0021] In some embodiments, the method for processing marbled beef includes extracting band features from the target spectral data, including:

[0022] The target spectral data is input into the trained PLS model to determine the preset number of bands with the highest weight values ​​in the spectral data as band features.

[0023] In some embodiments, the marbled beef processing method described herein constructs a marbled beef sample set using the following method:

[0024] When the sample image data of beef is determined to be sample image data of marbled beef, the location of the marbled beef and the grade label of the marbled beef are marked in the sample image data of marbled beef to obtain the marked sample image data.

[0025] Determine the sample spectral data associated with the labeled sample image data;

[0026] A marbled beef sample set is constructed, which includes marbled beef sample data of multiple grades. Each marbled beef sample data includes: the number of sample images labeled with the location and grade of the marbled beef, and sample spectral data associated with the sample image data.

[0027] In some embodiments, the method for processing marbled beef further includes:

[0028] By training the snowflake beef detection model and the grade prediction model using the constructed snowflake beef sample set, the trained snowflake beef detection model and the trained grade prediction model are obtained.

[0029] In some embodiments, a snowflake beef processing apparatus is also provided, comprising:

[0030] The acquisition module is used to acquire the target image data and target spectral data of the target snowflake beef after preprocessing when the beef image data is determined to be the target snowflake beef image data by the pre-trained snowflake beef detection model.

[0031] An extraction module is used to extract color features from the target image data and band features from the target spectral data;

[0032] A fusion module is used to fuse the color features and band features to obtain the fused features of the target snowflake beef;

[0033] The determination module is used to input the fusion features of the target marbled beef into the trained grade prediction model to determine the grade of the target marbled beef.

[0034] In some embodiments, an electronic device is also provided, including: a processor, a memory, and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the memory via the bus, and when the machine-readable instructions are executed by the processor, the steps of the snowflake beef processing method are performed.

[0035] In some embodiments, a computer-readable storage medium is also provided, on which a computer program is stored, which, when executed by a processor, performs the steps of the described snowflake beef processing method.

[0036] This application provides an embodiment of a method, apparatus, electronic device, and medium for processing marbled beef. First, it detects whether the image data of the beef is marbled beef. If so, it extracts the color features and spectral band features of the image data of marbled beef and uses a trained grade prediction model to automatically determine the grade of the marbled beef. The whole process is highly automated, non-detectable, non-destructive, and non-contact, and the grade assessment is accurate and objective. Attached Figure Description

[0037] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 A flowchart of a method for processing Wagyu beef is shown;

[0039] Figure 2 A flowchart of the method for determining whether beef image data is marbled beef image data according to an embodiment of this application is shown;

[0040] Figure 3 This invention illustrates a flowchart of a method for obtaining preprocessed target image data and target spectral data of the target marbled beef, as described in an embodiment of this application.

[0041] Figure 4 A flowchart illustrating the method for constructing a snowflake beef sample set according to an embodiment of this application is shown;

[0042] Figure 5 A schematic diagram of the structure of a snowflake beef processing device according to an embodiment of this application is shown;

[0043] Figure 6 A schematic diagram of the structure of an electronic device according to an embodiment of this application is shown. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the accompanying drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0045] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0046] It should be noted that the term "comprising" will be used in the embodiments of this application to indicate the presence of the features declared thereafter, but does not exclude the addition of other features.

[0047] Yunling cattle are one of the four major beef cattle breeds in my country with completely independent intellectual property rights in Yunnan Province since the founding of the People's Republic of China. They are rich in high-grade Wagyu beef, with meat quality comparable to Kobe beef. However, there is still a lack of an effective automatic grading method for Wagyu beef.

[0048] In recent years, deep learning has been used in image recognition of livestock products such as chicken and mutton. However, deep learning technology still faces challenges in addressing target detection and automatic grading of Yunling beef products, including weak data foundations, insufficient theoretical research, and an incomplete algorithmic methodology. Therefore, overcoming the technical bottlenecks in target detection and automatic grading of Yunling beef is crucial for improving the automation level of Yunling beef processing and manufacturing.

[0049] Based on this, this application proposes a method for processing marbled beef. First, it detects whether the image data of the beef is marbled beef. If so, it extracts the color features of the image data of marbled beef and the band features of the spectral data. Then, it uses a trained grade prediction model to automatically determine the grade of the marbled beef. The whole process is highly automated, non-detection, non-destructive, and non-contact, and the grade assessment is accurate and objective.

[0050] Based on this, such as Figure 1 As shown, this application provides a method for processing Wagyu beef, including the following steps S101-S104:

[0051] S101. When the beef image data is determined to be the target snowflake beef image data by the pre-trained snowflake beef detection model, the target image data and target spectral data of the target snowflake beef after preprocessing are obtained.

[0052] S102. Extract color features from the target image data and extract band features from the target spectral data;

[0053] S103. The color features and band features are fused to obtain the fused features of the target snowflake beef;

[0054] S104. Input the fusion features of the target marbled beef into the trained grade prediction model to determine the grade of the target marbled beef.

[0055] Before step S101, that is, before obtaining the target image data and target spectral data of the target marbled beef after preprocessing, it is necessary to determine whether the beef image data is marbled beef image data.

[0056] Specifically, such as Figure 2 As shown, determining whether beef image data is marbled beef image data includes the following steps S201-S204.

[0057] S201. Obtain beef image data;

[0058] S202. Input the beef image data into the trained snowflake beef detection model to extract the marble pattern features and color features from the beef image data;

[0059] S203. The trained snowflake beef detection model determines whether the snowflake beef region in the beef image data meets the preset conditions based on the marble pattern features and color features.

[0060] S204. If so, then determine that the beef image data is the image data of the target marbled beef.

[0061] The beef image data refers to beef images captured by a high-definition camera.

[0062] The target spectral data of the target marbled beef includes spectral data collected by near-infrared equipment and spectral data collected by hyperspectral equipment.

[0063] The snowflake beef detection model is a target detection model. The target of the target detection model is snowflake beef. If the target detection model detects snowflake beef, the classification result is 1; otherwise, the classification result is 0.

[0064] The target detection model uses the sliding window method and transfer learning to determine the area where the marbled beef is located by extracting the marble pattern and color features from the beef image data, and then selects the area where the marbled beef is located. If the area where the marbled beef is located meets a preset condition, such as the proportion of the area where the marbled beef is located exceeds a preset ratio, then the beef image data is determined to be the image data of the target marbled beef.

[0065] It should be noted that, for all types of beef, spectral data of the beef is also collected simultaneously when acquiring beef image data.

[0066] When a pre-trained marbled beef detection model determines that the beef image data is the target marbled beef image data, the spectral data associated with the target marbled beef image data is obtained.

[0067] like Figure 3 As shown, the acquisition of the target image data and target spectral data of the target marbled beef after preprocessing includes the following steps S301-S303:

[0068] S301. Acquire the image data and spectral data of the target marbled beef;

[0069] S302. Preprocess the image data of the target marbled beef to obtain target image data;

[0070] S303. Preprocess the target spectral data of the target marbled beef to obtain target spectral data.

[0071] The raw data obtained from measuring the sample contains valuable information about the sample, but also non-experimental data such as noise and dark current. To obtain accurate spectral data and effectively improve the signal-to-noise ratio, preprocessing of the raw spectrum is necessary. Common preprocessing methods include detrending correction, multivariate scattering correction, derivative, max-min normalization, smoothing and denoising, mean centering, and variable standardization. In this embodiment, the method for preprocessing the target spectral data of the target marbled beef includes at least one of the following: trend correction, multivariate scattering correction, derivative, max-min normalization, smoothing and denoising, mean centering, and variable standardization.

[0072] Commonly used image data preprocessing methods include image grayscale processing, image smoothing processing, and image morphological processing. In this embodiment, the method for preprocessing the target spectral data of the target marbled beef includes at least one of the following: image grayscale processing, image smoothing processing, and image morphological processing.

[0073] In step S202, the beef image data is input into the trained marbled beef detection model to extract the marble pattern and color features from the beef image data.

[0074] Specifically, extracting band features from the target spectral data includes:

[0075] The target spectral data is input into the trained PLS model to determine the preset number of bands with the highest weight values ​​in the spectral data as band features.

[0076] Given that directly using the entire spectral data would result in excessive data volume and other drawbacks, we analyze and reduce the dimensionality of the full-band spectral data to extract a few representative bands. These characteristic bands contain a large amount of useful spectral information, have minimal redundancy, and are less susceptible to noise. Using these extracted representative characteristic bands to build and train the grade prediction model makes the model more stable and improves classification accuracy.

[0077] In this embodiment, a trained PLS model is used for feature band selection. The specific methods for feature band selection using the PLS model mainly include the PLS pruning algorithm and the PLS weighted regression coefficient method. The purpose of the PLS weighted regression coefficient method is to select the spectral band with the highest weight. The principle of this method is to model the processed spectrum using the least squares method. After modeling, the weighted regression coefficients for each band are obtained. These weighted regression coefficients are then arranged, and the band with the highest weight value is selected. The band with the highest weight value represents the band that has the greatest impact on the spectral data. This process is repeated cyclically, selecting a preset number of bands with the highest weight values—that is, the most useful bands—as band features, while removing bands with strong interference.

[0078] Extracting color features from the target image data includes:

[0079] The brightness values ​​of each pixel in the red component, the green component, and the blue component of the target image data are extracted respectively as color features.

[0080] In color feature extraction from image data: R represents red, G represents green, and B represents blue. These three color components exist in matrix form, and when combined, they can display various colors, such as cyan, purple, and pink. Within the defined range of the spectrum, the wavelength of red is defined as 700nm, the wavelength of green as 546.1nm, and the wavelength of blue as 435.8nm. According to the basic principle of the three primary colors, any color can be represented by the RGB color equation:

[0081]

[0082] In the formula, α, β, and γ are called the three color coefficients, and G represents the color after the three primary colors are processed in a certain proportion. The color under any spectral wavelength can be digitally quantized and represented by the numerical values ​​of the three primary colors: red (R), green (G), and blue (B).

[0083] In digital image research, color feature information should be considered, especially when using this information to detect meat quality content. This application's embodiments extract the brightness values ​​of each pixel in the red component, the green component, and the blue component of the target image data as color features.

[0084] In step S103, the color features and band features are fused to obtain the fused features of the target snowflake beef.

[0085] In this embodiment of the application, a fusion model is constructed based on the partial least squares (PLSR) method. The fusion model is used to fuse the color features and band features to obtain the fused features of the target snowflake beef.

[0086] Partial least squares (PLS) is used to find the fundamental relationship between two matrices (X and Y), i.e., a latent variable approach that models the covariance structure in these two spaces. A PLS model attempts to find multidimensional directions in the X space that explain the largest variance in the Y space. PLS regression is particularly suitable when the prediction matrix has more variables than the observations, and when there is multicollinearity in the values ​​of X, by projecting the predictors and observed variables into a new space to find a linear regression model.

[0087] In step S104, the grade prediction model is constructed based on partial least squares.

[0088] Partial least squares (PLS) combines the advantages of principal component analysis (PCA) and linear discriminant analysis (LDA). The principle of PLS ​​is to analyze the relationships between variables, build a prediction model based on a calibration set, and then use this model to test the data in the prediction set. This method can reliably establish a meat quality grading model. Methods like LLS analyze one matrix at a time, which can lead to large variance. LLS, however, analyzes two matrices simultaneously, ensuring the accuracy of the grading prediction model.

[0089] Currently, there is no standard dataset for training deep learning networks using Wagyu beef, and due to cost limitations, the dataset inevitably has a small sample size. Therefore, such as Figure 4 As shown, the embodiments of this application construct a Wagyu beef sample set using the following method:

[0090] S401. When the sample image data of beef is determined to be sample image data of marbled beef, the location of the marbled beef and the grade label of the marbled beef are marked in the sample image data of marbled beef to obtain the marked sample image data.

[0091] S402. Determine the sample spectral data associated with the labeled sample image data;

[0092] S403. Construct a snowflake beef sample set, which includes snowflake beef sample data of multiple grades. Each snowflake beef sample data includes: the number of sample images labeled with the location and grade of the snowflake beef, and sample spectral data associated with the sample image data.

[0093] Before step S401, obtain no less than 2000 samples of Yunling beef.

[0094] In steps S401 and S402, for each beef sample, it is first determined whether the beef sample contains marbled beef by expert scoring and fuzzy evaluation. If so, a grade is given according to the colorimetric card method and a grade label is affixed. Subsequently, image data of the determined sample is acquired by a high-definition camera, and spectral data of the sample is acquired by near-infrared and hyperspectral equipment.

[0095] In step S403, a mapping table is established between the grade labels determined by experts and the image data and spectral data obtained by non-destructive testing methods. Through data cleaning, normalization and standardization, the "Yunling Beef Automatic Grading Standard Dataset", namely the snowflake beef sample set, is formed. The snowflake beef sample set includes snowflake beef sample data of multiple grades. Each snowflake beef sample data includes: the number of sample images labeled with the location and grade label of the snowflake beef, and the sample spectral data associated with the sample image data.

[0096] For example, the Wagyu beef sample set should include no less than 1,000 Wagyu beef sample data, and according to the international standard A1-A5 grade of Wagyu beef, there should be no less than 250 sample data for each grade.

[0097] The method for processing marbled beef described in this application embodiment further includes:

[0098] By training the snowflake beef detection model and the grade prediction model using the constructed snowflake beef sample set, the trained snowflake beef detection model and the trained grade prediction model are obtained.

[0099] Similarly, the trained Wagyu beef detection model and PLS model were both obtained by training on the constructed Wagyu beef sample set.

[0100] like Figure 5 As shown in the illustration, this application also provides a snowflake beef processing device, comprising:

[0101] The acquisition module 501 is used to acquire the target image data and target spectral data of the target snowflake beef after preprocessing when the beef image data is determined to be the target snowflake beef image data by the pre-trained snowflake beef detection model.

[0102] Extraction module 502 is used to extract color features from the target image data and extract band features from the target spectral data;

[0103] The fusion module 503 is used to fuse the color features and band features to obtain the fused features of the target snowflake beef;

[0104] The determination module 504 is used to input the fusion features of the target marbled beef into the trained grade prediction model to determine the grade of the target marbled beef.

[0105] This application proposes a snowflake beef processing device. First, it detects whether the image data of the beef is snowflake beef. If so, it extracts the color features and spectral band features of the image data of the snowflake beef and uses a trained grade prediction model to automatically determine the grade of the snowflake beef. The whole process is highly automated, non-detection, non-destructive, and non-contact, and the grade assessment is accurate and objective.

[0106] like Figure 6 As shown in the figure, this application embodiment also provides an electronic device 600, including: a processor 602, a memory 601 and a bus. The memory 601 stores machine-readable instructions that can be executed by the processor 602. When the electronic device 600 is running, the processor 602 communicates with the memory 601 through the bus. When the machine-readable instructions are executed by the processor 602, the steps of the snowflake beef processing method are performed.

[0107] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the described snowflake beef processing method.

[0108] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and devices described above can be referred to the corresponding processes in the method embodiments, and will not be repeated here. In the several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection can be through some communication interfaces; the indirect coupling or communication connection of devices or modules can be electrical, mechanical, or other forms.

[0109] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0110] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0111] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a platform server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0112] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method of processing snowflake beef, characterized by, The method comprises the following steps: When it is judged by the pre-trained snowflake beef detection model that the beef image data is the image data of the target snowflake beef, the target image data and the target spectrum data obtained after preprocessing of the target snowflake beef are acquired; Color features are extracted from the target image data, and waveband features are extracted from the target spectrum data; The color features and the waveband features are fused to obtain the fusion features of the target snowflake beef; The fusion features of the target snowflake beef are input into the trained grade prediction model to determine the grade of the target snowflake beef. Before acquiring the target image data and the target spectrum data obtained after preprocessing of the target snowflake beef, the method further comprises: Acquiring beef image data; The marble pattern features and the color features in the beef image data are extracted by inputting the beef image data into the trained snowflake beef detection model; The trained snowflake beef detection model judges whether the snowflake beef region in the beef image data meets the preset condition according to the marble pattern features and the color features; If yes, the beef image data is judged as the image data of the target snowflake beef; The waveband features are extracted from the target spectrum data, comprising: The target spectrum data is input into the trained PLS model to determine the preset number of wavebands with the highest weight values in the spectrum data as the waveband features; The snowflake beef sample set is constructed by the following method: When it is judged that the sample image data of beef is the sample image data of snowflake beef, the position of the snowflake beef and the grade label of the snowflake beef are labeled in the sample image data of the snowflake beef to obtain labeled sample image data; the sample spectrum data associated with the labeled sample image data is determined; The snowflake beef sample set is constructed, and the snowflake beef sample set comprises snowflake beef sample data of multiple grades, and each snowflake beef sample data comprises: sample image data labeled with the position of the snowflake beef and the grade label of the snowflake beef, and sample spectrum data associated with the sample image data.

2. The snowflake beef processing method according to claim 1, characterized by, Acquiring the target image data and the target spectrum data obtained after preprocessing of the target snowflake beef comprises: Acquiring image data and spectrum data of the target snowflake beef; The image data of the target snowflake beef is preprocessed to acquire target image data; The target spectrum data of the target snowflake beef is preprocessed to acquire target spectrum data.

3. The snowflake beef processing method of claim 1, wherein, The color features are extracted from the target image data, comprising: The brightness value of each pixel point of the target image data on the red component, the brightness value of each pixel point of the target image data on the green component, and the brightness value of each pixel point of the target image data on the blue component are extracted as color features.

4. The snowflake beef processing method of claim 1, wherein, The method further comprises: The snowflake beef detection model and the grade prediction model are trained by the constructed snowflake beef sample set to obtain the trained snowflake beef detection model and the trained grade prediction model.

5. A snowflake beef processing apparatus, characterized by, Comprise: The acquisition module is configured to acquire target image data and target spectral data of the target marbling beef after preprocessing when the marbling beef detection model that is pre-trained determines that the beef image data is image data of the target marbling beef. The extraction module is configured to extract color features from the target image data and extract band features from the target spectral data. The fusion module is configured to fuse the color features and the band features to obtain fusion features of the target marbling beef. The determination module is configured to input the fusion features of the target marbling beef into the trained grade prediction model to determine the grade of the target marbling beef. Before the target image data and the target spectral data of the target marbling beef after preprocessing are acquired, the method further includes: acquiring beef image data; inputting the beef image data into the trained marbling beef detection model to extract marbling features and color features in the beef image data; The trained marbling beef detection model determines whether a marbling beef region in the beef image data meets a preset condition according to the marbling features and the color features. If yes, the beef image data is determined as image data of the target marbling beef. The band features are extracted from the target spectral data, including: inputting the target spectral data into a trained PLS model to determine a preset number of bands with the highest weight values in the spectral data as the band features; The marbling beef sample set is constructed by the following method: When the sample image data of beef is determined as sample image data of marbling beef, the position of the marbling beef and a marbling beef grade label are marked in the sample image data of the marbling beef to obtain labeled sample image data; and sample spectral data associated with the labeled sample image data is determined. The marbling beef sample set includes marbling beef sample data of multiple grades, and each marbling beef sample data includes sample image data labeled with the position of the marbling beef and a marbling beef grade label and sample spectral data associated with the sample image data.

6. An electronic device, comprising: The processor, the memory, and the bus, the memory stores machine readable instructions executable by the processor, when the electronic device runs, the processor and the memory communicate through the bus, the machine readable instructions are executed by the processor to execute the steps of the marbling beef processing method in any one of claims 1 to 4. The computer readable storage medium stores a computer program, and the computer program is executed by the processor to execute the steps of the marbling beef processing method in any one of claims 1 to 4.

7. A computer readable storage medium characterized in that, ​

Images