Chicken quality nondestructive spectrum detection method based on optimal semi-unfrozen state
By employing an optimal spectral detection method for semi-thawed chicken meat, the contradiction between accuracy and meat quality damage during thawing is resolved, achieving non-destructive chicken quality assessment, improving detection accuracy, and reducing moisture loss.
Patent Information
- Application Number
- CN202510899000.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-01
- Publication Date
- 2025-10-31
AI Technical Summary
Existing chicken testing methods present a trade-off between accuracy and meat quality damage during the thawing process. Complete thawing affects testing accuracy, while not thawing is affected by the freezing effect, and inferior chicken is difficult to identify effectively.
A non-destructive spectral detection method for chicken quality under optimal semi-thawed conditions is adopted. By thawing chicken to the target semi-thawed state, hyperspectral data is acquired and preprocessed. The partial least squares regression algorithm is used to establish the relationship between spectrum and category, and the model performance is optimized by combining comprehensive indicators.
It enables accurate identification of chicken quality without damaging the meat. Through spectral detection in a semi-thawed state, it provides a controllable and reproducible detection solution, improving detection accuracy and reducing moisture loss.
Smart Images
Figure CN120870002A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of chicken meat detection technology, and in particular to a non-destructive spectroscopic detection method for chicken meat quality based on the optimal semi-thawed state. Background Technology
[0002] Chicken breast is considered an important nutritional resource, characterized by its high protein content and low fat level. In 2023, China's per capita chicken consumption reached 12.4 kg. Given China's large population, such a huge consumption volume highlights the importance of chicken quality assessment and grading.
[0003] In recent years, substandard chicken has posed a significant challenge to the entire poultry industry. Examples include woody-breast (WB) and pale, soft, and exudative (PSE) chicken. PSE chicken, as the name suggests, is characterized by its pale color, soft texture, and tendency to exude moisture. WB chicken is abnormally firm or stiff from head to tail and often has other defects such as white striping (WS), sticky liquid on the surface, and / or bleeding. These substandard chickens not only affect appearance but also the internal quality of the chicken breast. More importantly, they cannot be considered fully valuable meat and are unsuitable for direct sale to consumers.
[0004] Currently, chicken meat testing commonly employs spectral analysis, and the accuracy of chicken meat testing is related to the degree of thawing. While complete thawing can improve the spectral response, it also causes irreversible meat damage such as cell rupture and juice loss, reducing the economic value of the meat. On the other hand, complete non-thawing can avoid damage to the chicken meat caused by testing, but its accuracy is also affected by the freezing effect. Summary of the Invention
[0005] This invention discloses a non-destructive spectral detection method for chicken quality based on the optimal semi-thawed state, the specific method of which is as follows:
[0006] Thaw the frozen chicken to the target semi-thawed state;
[0007] Acquire hyperspectral data of the target chicken in a semi-thawed state;
[0008] Preprocess the hyperspectral data;
[0009] The preprocessed hyperspectral data is input into the prediction model to obtain the chicken quality classification.
[0010] Furthermore, chicken quality is classified into three categories: normal meat, WB meat, and PSE meat.
[0011] Furthermore, the formula for calculating the degree of thawing of chicken is as follows:
[0012]
[0013] In the formula, M0 is the total frozen mass of chicken, M f For whole-grain frozen chicken quality, M t This refers to the current quality of the chicken.
[0014] Furthermore, the target is in a semi-thawed state, and the specific method for obtaining this information is as follows:
[0015] Several chicken samples of different quality categories were uniformly frozen to a fully frozen state;
[0016] Frozen chicken samples were thawed at different set times, and the degree of thawing of each chicken sample after the first thawing was calculated.
[0017] Obtain hyperspectral data of chicken samples at different thawed stages;
[0018] Preprocess the hyperspectral data;
[0019] The preprocessed hyperspectral data is input into the prediction model to obtain the detection score for each chicken sample;
[0020] Chicken samples at different thawed stages were refreezed and then completely thawed a second time. The moisture loss rate of each chicken sample was then obtained.
[0021]
[0022] In the formula, M r The mass after thawing, refreezing, and then completely thawing again;
[0023] The prediction accuracy and poorness of each piece of chicken are weighted, and the comprehensive index of each piece of chicken is calculated using the following formula:
[0024] S i =α.A i +(1-α).(1-W loss,i )
[0025] Among them, A i Let W be the model classification accuracy corresponding to the i-th thawing state. loss,i Let α be the water loss rate corresponding to the i-th thawing state, and α be an empirical adjustment factor.
[0026] The first thawed state of the chicken with the best comprehensive index is selected as the target semi-thawed state.
[0027] Furthermore, the key wavelengths for acquiring hyperspectral data were selected using a regression coefficient method; near-infrared hyperspectral analyzers and halogen lamp push-broom multispectral imaging systems were used to scan the reflectance spectra of the hyperspectral data.
[0028] Furthermore, the hyperspectral data is preprocessed using the following methods:
[0029] Principal component analysis was used for denoising, followed by Savitzky-Golay smoothing, and then inverse PCA for restoration.
[0030] Furthermore, the prediction model employs partial least squares regression to establish quantitative and qualitative relationships between spectra and categories; model performance is evaluated using the correlation coefficient of the calibration set, the root mean square error of the calibration set estimate, the correlation coefficient of the independent prediction set, and the root mean square error of the independent prediction set estimate.
[0031] Due to the adoption of the above technical solutions, the present invention has the following beneficial effects:
[0032] 1. This invention proposes a semi-thawed state detection scheme for the first time, which transforms the ambiguous state of partial thawing into a quantifiable parameter Dt, achieving controllability and reproducibility; and utilizes weight changes and temperature-controlled thawing time to achieve non-destructive evaluation.
[0033] 2. This invention proposes a comprehensive index and introduces multi-index evaluation, and for the first time applies a joint optimization strategy in the spectral detection of meat.
[0034] 3. Independent spectral prediction models were established for different degrees of thawing, and the performance trend of the models with Dt was evaluated to form a complete detection performance curve.
[0035] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description
[0036] The accompanying drawings of this invention are described below.
[0037] Figure 1 This is a schematic diagram of the overall process of the present invention.
[0038] Figure 2 Average spectral curves of different types of chicken.
[0039] Figure 3 Accuracy and moisture loss rate curves at different thawing degrees.
[0040] Figure 4 Comprehensive index curves under different degrees of thawing.
[0041] Figure 5 The model confusion matrix under the optimal Dt30min. Detailed Implementation
[0042] The present invention will be further described below with reference to the accompanying drawings and embodiments.
[0043] A non-destructive spectroscopic method for detecting chicken quality based on the optimal semi-thawed state, such as... Figure 1 As shown, the specific steps are as follows:
[0044] S1. Thaw the frozen chicken until it reaches the target semi-thawed state.
[0045] The formula for calculating the degree of thawing of chicken is as follows:
[0046]
[0047] In the formula, M0 is the total frozen mass of chicken, M f For whole-grain frozen chicken quality, M t This refers to the current quality of the chicken.
[0048] S2. Obtain hyperspectral data of the target partially thawed chicken meat. The specific steps are as follows:
[0049] S21. After the frozen chicken samples are thawed for different set times, calculate the degree of thawing of each chicken sample after the first thawing.
[0050] S22. Obtain hyperspectral data of chicken samples at different thawed levels.
[0051] S23. Preprocess the hyperspectral data.
[0052] S24. Input the preprocessed hyperspectral data into the prediction model to obtain the detection score for each chicken sample.
[0053] S25. Refreeze chicken samples at different thawed levels, then thaw them completely a second time, and obtain the moisture loss rate for each chicken sample:
[0054]
[0055] In the formula, M r The quality is the product after thawing, refreezing, and then completely thawing again.
[0056] S26. Weight the prediction accuracy and poorness of each chicken sample, and calculate the comprehensive index for each chicken sample. The specific formula is as follows:
[0057] S i =α.A i +(1-α).(1-W loss,i )
[0058] Among them, A iLet W be the model classification accuracy corresponding to the i-th thawing state. loss,i Let α be the water loss rate corresponding to the i-th thawing state, and α be an empirical adjustment factor.
[0059] S27. Select the first thawing state of the chicken with the best comprehensive index as the target semi-thawed state.
[0060] In step S2, the key wavelength for acquiring hyperspectral data is selected using a regression coefficient method. Hyperspectral data is acquired using a near-infrared hyperspectral analyzer and a push-broom multispectral imaging system with a halogen lamp source for reflectance spectral scanning. Spectral range: 380-1000 nm; resolution: 5 nm; entire sample area is selected.
[0061] S3. Preprocess the hyperspectral data.
[0062] This step uses principal component analysis for denoising, followed by Savitzky-Golay smoothing, and then inverse PCA for restoration.
[0063] S4. Input the preprocessed hyperspectral data into the prediction model to obtain the chicken quality classification.
[0064] The prediction model employs partial least squares regression to establish quantitative and qualitative relationships between spectra and categories. Model performance is evaluated using the correlation coefficient of the calibration set, the root mean square error of the calibration set estimate, the correlation coefficient of the independent prediction set, and the root mean square error of the independent prediction set estimate.
[0065] In this step, the chicken quality classification includes: normal meat, WB meat, and PSE meat.
[0066] Simulation experiments such as Figures 2 to 5 As shown:
[0067] Table 1 summarizes the quantity distribution of Normal, WB, and PSE meat samples under different thawing times. For each thawing time (10-50 minutes), there were 18 samples in each category, totaling 54 samples per thawing time; a total of 90 samples in each category and 270 samples in all categories were recorded, showing the correspondence between thawing time and the number of samples in different categories.
[0068] Table 1. Number of sample categories at different thawing times
[0069]
[0070]
[0071] Table 2. Comprehensive index results based on different thawing times.
[0072]
[0073] Table 2 focuses on different thawing times (10-50 minutes). By setting weights (accuracy weight 0.7, moisture loss rate weight 0.3), a comprehensive index is derived to evaluate performance at different thawing times. At 20 minutes, the comprehensive index is highest at 0.898, with a good model accuracy of 0.88 and a relatively low moisture loss rate of 0.06, indicating the best overall performance. As thawing time increases, model accuracy fluctuates, reaching 0.90 at 50 minutes, but the moisture loss rate increases significantly, reaching 0.28 at 50 minutes. The comprehensive index shows a trend of first increasing and then decreasing. Around 20 minutes is a relatively ideal thawing time, ensuring model performance while minimizing moisture loss.
[0074] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.
Claims
1. A non-destructive spectroscopic method for detecting chicken quality based on the optimal semi-thawed state, characterized in that, The specific method is as follows: Thaw the frozen chicken to the target semi-thawed state; Acquire hyperspectral data of the target chicken in a semi-thawed state; Preprocess the hyperspectral data; The preprocessed hyperspectral data is input into the prediction model to obtain the chicken quality classification.
2. The non-destructive spectroscopic detection method for chicken quality based on the optimal semi-thawed state as described in claim 1, characterized in that, Chicken quality is classified into three categories: normal meat, WB meat, and PSE meat.
3. The non-destructive spectroscopic detection method for chicken quality based on the optimal semi-thawed state as described in claim 1, characterized in that, The formula for calculating the degree of thawing of chicken is as follows: In the formula, M0 is the total frozen mass of the chicken, M f For whole-grain frozen chicken quality, M t This refers to the current quality of the chicken.
4. The non-destructive spectroscopic detection method for chicken quality based on the optimal semi-thawed state as described in claim 3, characterized in that, The target is in a partially thawed state. The specific method for obtaining this information is as follows: Several chicken samples of different quality categories were uniformly frozen to a fully frozen state; Frozen chicken samples were thawed at different set times, and the degree of thawing of each chicken sample after the first thawing was calculated. Obtain hyperspectral data of chicken samples at different thawed stages; Preprocess the hyperspectral data; The preprocessed hyperspectral data is input into the prediction model to obtain the detection score for each chicken sample; Chicken samples at different thawed stages were refreezed and then completely thawed a second time. The moisture loss rate of each chicken sample was then obtained. In the formula, M r The mass after thawing, refreezing, and then completely thawing again; The prediction accuracy and poorness of each piece of chicken are weighted, and the comprehensive index of each piece of chicken is calculated using the following formula: S i =a.A i +(1-a).(1-W loss,i ) Among them, A i Let W be the model classification accuracy corresponding to the i-th thawing state. loss,i Let α be the water loss rate corresponding to the i-th thawing state, and α be an empirical adjustment factor. The first thawed state of the chicken with the best comprehensive index is selected as the target semi-thawed state.
5. The non-destructive spectroscopic detection method for chicken quality based on the optimal semi-thawed state as described in claim 1 or 4, characterized in that, The key wavelengths for acquiring hyperspectral data were selected using the regression coefficient method; near-infrared hyperspectral analyzers and halogen lamp push-broom multispectral imaging systems were used to scan the reflectance spectra of the hyperspectral data.
6. The non-destructive spectroscopic detection method for chicken quality based on the optimal semi-thawed state as described in claim 1 or 4, characterized in that, The hyperspectral data preprocessing method is as follows: Principal component analysis was used for denoising, followed by Savitzky-Golay smoothing, and then inverse PCA for restoration.
7. The non-destructive spectroscopic detection method for chicken quality based on the optimal semi-thawed state as described in claim 1, characterized in that, The prediction model uses partial least squares regression to establish quantitative and qualitative relationships between spectra and categories; Model performance was evaluated using the correlation coefficient of the calibration set, the root mean square error of the calibration set estimate, the correlation coefficient of the independent prediction set, and the root mean square error of the independent prediction set estimate.