Method and system for identifying fat content of broiler chicken

By combining near-infrared spectroscopy with multiple linear regression, the problems of time-consuming, labor-intensive, and destructive detection of fat content in broilers have been solved, achieving rapid, non-destructive, and efficient fat content identification, which is suitable for online real-time monitoring in modern production lines.

CN118961644APending Publication Date: 2024-11-15JIANGSU INST OF POULTRY SCI
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411056218.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-02
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

Existing methods for detecting fat content in broiler chickens are time-consuming, labor-intensive, destructive, and costly, making it difficult to meet the needs of rapid and non-destructive testing in modern production.

Method used

By combining near-infrared spectroscopy with multiple linear regression, samples were collected from different age groups of broilers. After preprocessing, the samples were scanned using a near-infrared spectrometer to establish a calibration model and identify fat content. Noise was removed using wavelet decomposition and denoising functions to achieve non-destructive detection.

Benefits of technology

It enables rapid, non-destructive detection of fat content in broilers, improving detection efficiency, reducing costs, and supporting online real-time monitoring and automated control.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118961644A_ABST
    Figure CN118961644A_ABST
Patent Text Reader

Abstract

The invention provides a broiler fat content identification method and system, and relates to the technical field of poultry farming, and the method comprises the steps: collecting broiler samples from each part of a broiler, and obtaining the actual fat content of the broiler samples; grinding, homogenizing and dewatering the collected sample to obtain a pretreated sample; scanning the preprocessed samples by using a near infrared spectrometer to obtain spectral data of each sample; correlating the spectral data with the actual fat content by using a multiple linear regression method, and establishing a correction model; verifying the correction model by using an independent sample set which does not participate in modeling to obtain a verified identification model; and inputting spectral data of a to-be-detected broiler chicken into the recognition model to obtain a fat recognition result. According to the method, destructive treatment on the sample is not needed, the integrity and the original state of the sample are kept, and the detection efficiency of the fat content of the broiler chicken can be further improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of poultry breeding, and in particular to a method and system for identifying the fat content of broiler chickens. Background Art

[0002] In modern broiler farming and food processing industries, the fat content of broilers is a key quality indicator. Fat content not only affects the nutritional value and taste of broilers, but is also directly related to the market value of the product and the health needs of consumers. Therefore, it is particularly important to accurately and quickly detect the fat content of broilers.

[0003] At present, the detection of broiler fat content mainly relies on chemical analysis methods, such as Soxhlet Extraction and Gravimetric Method. Although these traditional methods can provide relatively accurate fat content data, they have the following disadvantages:

[0004] 1. Time-consuming and labor-intensive: Traditional chemical analysis methods usually require complex sample pretreatment and long operation processes, which are difficult to meet the needs of rapid detection in modern production.

[0005] 2. Destructive testing: Most of these methods require destructive sampling and processing, which makes the samples irreversible and cannot be applied to online testing and continuous production.

[0006] 3. High cost: Chemical analysis requires the use of a large amount of chemical reagents and special equipment, resulting in high testing costs and is not conducive to large-scale application. Summary of the invention

[0007] In order to overcome the deficiencies of the prior art, the object of the present invention is to provide a method and system for identifying the fat content of broilers.

[0008] To achieve the above object, the present invention provides the following solutions:

[0009] A method for identifying the fat content of broiler chickens, comprising:

[0010] Collecting broiler samples from various parts of broiler chickens of different age groups, and obtaining actual fat content of the broiler samples;

[0011] Grinding, homogenizing and dehydrating the collected samples to obtain pre-treated samples;

[0012] Scanning the pre-treated samples using a near infrared spectrometer to obtain spectral data of each sample;

[0013] Using a multiple linear regression method, the spectral data is correlated with the actual fat content to establish a calibration model;

[0014] Using an independent sample set that does not participate in modeling to verify the correction model, to obtain a verified recognition model;

[0015] The spectral data of the broiler to be tested is input into the recognition model to obtain the fat recognition result.

[0016] Preferably, the wavelength range of the spectral data is 900-2500 nm.

[0017] Preferably, after scanning the pre-processed samples using a near infrared spectrometer to obtain spectral data of each sample, the method further includes:

[0018] Shifting the spectral data of each sample to obtain shifted spectral data;

[0019] Performing wavelet decomposition on the shifted spectral data to obtain a plurality of wavelet coefficients;

[0020] Determine the filtering threshold according to the decomposition scale and length of the signal;

[0021] Constructing a denoising function according to the filtering threshold;

[0022] The denoising function is used to remove noise from the shifted spectral data to obtain denoised spectral data.

[0023] Preferably, the formula of the filtering threshold is: Among them, λ is the filtering threshold, j is the decomposition scale of the spectral data after translation, and d j is the decomposition of wavelet coefficients with scale j, N represents the length of the shifted spectral data, and median represents the median operation.

[0024] Preferably, using the denoising function to remove the noise of the shifted spectral data to obtain denoised spectral data includes:

[0025] The denoising function is used to remove the corresponding wavelet coefficients to obtain smoothed spectral data; wherein the denoising function is: Among them, X represents the wavelet coefficient, n represents the adjustment coefficient;

[0026] Performing inverse translation on the smoothed spectral data to obtain inverse translated spectral data;

[0027] The average of each inverse-translated spectral data is calculated to obtain the denoised spectral data.

[0028] Preferably, the formula of the correction model is: y=Xb+e; wherein y represents the actual fat content of each sample; X is a feature matrix; the size of X is N×M; wherein N is the number of samples, and M is the number of spectral wavelengths; b is a regression coefficient vector with a size of M×1; and e is a residual vector with a size of N×1.

[0029] Preferably, the evaluation parameters for validating the calibration model using an independent sample set not involved in the modeling include a root mean square error and a correlation coefficient.

[0030] A broiler fat content identification system, comprising:

[0031] A sample collection unit, used to collect broiler samples from various parts of broiler chickens of different age groups, and obtain the actual fat content of the broiler samples;

[0032] A sample pretreatment unit, used for grinding, homogenizing and dehydrating the collected samples to obtain pretreated samples;

[0033] A scanning unit, used to scan the pre-processed samples using a near infrared spectrometer to obtain spectral data of each sample;

[0034] A model building unit, used to associate the spectral data with the actual fat content using a multiple linear regression method to build a correction model;

[0035] A verification unit, used to verify the calibration model using an independent sample set that does not participate in modeling, to obtain a verified recognition model;

[0036] The recognition unit is used to input the spectral data of the broiler to be tested into the recognition model to obtain a fat recognition result.

[0037] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0038] The present invention provides a method for identifying the fat content of broilers, comprising: collecting broiler samples from various parts of broilers of different age groups, and obtaining the actual fat content of the broiler samples; grinding, homogenizing and dehydrating the collected samples to obtain pre-treated samples; scanning the pre-treated samples using a near-infrared spectrometer to obtain spectral data of each sample; using a multiple linear regression method to associate the spectral data with the actual fat content to establish a correction model; using an independent sample set that does not participate in modeling to verify the correction model to obtain a verified recognition model; inputting the spectral data of the broiler to be tested into the recognition model to obtain a fat recognition result. The present invention can maintain the integrity and original state of the samples without destructive treatment of the samples, and can further improve the detection efficiency of the fat content of broilers. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0040] Figure 1 A flow chart of a method provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] The purpose of the present invention is to provide a method for identifying the fat content of broilers, which can maintain the integrity and original state of the samples without destructive treatment of the samples, and can further improve the detection efficiency of the fat content of broilers.

[0043] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0044] Figure 1 A flow chart of a method provided by an embodiment of the present invention, such as Figure 1 As shown, the present invention provides a method for identifying the fat content of broiler chickens, comprising:

[0045] Step 100: Collecting broiler samples from various parts of broiler chickens of different age groups, and obtaining actual fat content of the broiler samples;

[0046] Step 200: Grinding, homogenizing and dehydrating the collected sample to obtain a pre-treated sample;

[0047] Step 300: Scanning the pre-processed samples using a near infrared spectrometer to obtain spectral data of each sample;

[0048] Step 400: using a multiple linear regression method, correlating the spectral data with the actual fat content to establish a calibration model;

[0049] Step 500: using an independent sample set that does not participate in modeling to verify the calibration model, to obtain a verified recognition model;

[0050] Step 600: Input the spectral data of the broiler to be tested into the recognition model to obtain a fat recognition result.

[0051] Specifically, in this embodiment, samples are collected from different broiler parts (breast meat, thigh meat, wings, etc.). The broiler samples in this embodiment represent the overall situation, including different age groups, genders and breeds.

[0052] Furthermore, in this embodiment, the collected samples are processed, such as grinding, homogenization, etc., to ensure the uniformity of the samples; and freeze-drying (or other suitable methods) is used to remove moisture from the samples.

[0053] Preferably, the wavelength range of the spectral data is 900-2500 nm.

[0054] Optionally, this embodiment uses a near infrared spectrometer to scan the pre-treated sample to obtain its spectral data. The spectral data includes absorption spectrum characteristics at a specific wavelength (such as 900-2500nm).

[0055] Preferably, after scanning the pre-processed samples using a near infrared spectrometer to obtain spectral data of each sample, the method further includes:

[0056] Shifting the spectral data of each sample to obtain shifted spectral data;

[0057] Performing wavelet decomposition on the shifted spectral data to obtain a plurality of wavelet coefficients;

[0058] Determine the filtering threshold according to the decomposition scale and length of the signal;

[0059] Constructing a denoising function according to the filtering threshold;

[0060] The denoising function is used to remove noise from the shifted spectral data to obtain denoised spectral data.

[0061] The basic idea of ​​this embodiment is that the wavelet coefficients corresponding to the spectral data signal are generally large, while the wavelet coefficients corresponding to the noise are generally small. Therefore, based on this feature, the spectral data signal can be de-noised to improve the quality of the spectral data signal.

[0062] Preferably, the formula of the filtering threshold is: Among them, λ is the filtering threshold, j is the decomposition scale of the spectral data after translation, and d j is the decomposition of wavelet coefficients with scale j, N represents the length of the shifted spectral data, and median represents the median operation.

[0063] Specifically, this embodiment considers that in the wavelet threshold denoising process, the quality of threshold selection directly affects the denoising effect. Since the wavelet coefficient amplitude of noise has the characteristic of decreasing with the increase of decomposition scale, the present invention introduces the decomposition scale to adaptively adjust the filtering threshold.

[0064] Preferably, using the denoising function to remove the noise of the shifted spectral data to obtain denoised spectral data includes:

[0065] The denoising function is used to remove the corresponding wavelet coefficients to obtain smoothed spectral data; wherein the denoising function is: Among them, X represents the wavelet coefficient, n represents the adjustment coefficient;

[0066] Performing inverse translation on the smoothed spectral data to obtain inverse translated spectral data;

[0067] The average of each inverse-translated spectral data is calculated to obtain the denoised spectral data.

[0068] Furthermore, this embodiment uses wavelet coefficients to measure the attenuation deviation, so that the attenuation of wavelet coefficients with larger absolute values ​​decreases as their absolute values ​​increase, thereby effectively avoiding the loss of high-frequency information and improving the signal-to-noise ratio of the signal.

[0069] Preferably, the formula of the correction model is: y=Xb+e; wherein y represents the actual fat content of each sample; X is a feature matrix; the size of X is N×M; wherein N is the number of samples, and M is the number of spectral wavelengths; b is a regression coefficient vector with a size of M×1; and e is a residual vector with a size of N×1.

[0070] Specifically, this embodiment uses a chemical analysis method (such as Soxhlet extraction) to measure the actual fat content of the sample as reference data for the model, and uses multiple linear regression (MLR), partial least squares regression (PLS), or other appropriate statistical methods to correlate the spectral data with the actual fat content to establish a calibration model.

[0071] Furthermore, b=(X T X) -1 X T y; however, for PLS regression, the optimal regression coefficient vector b is usually obtained through iterative calculation rather than direct calculation.

[0072] Preferably, the evaluation parameters for validating the calibration model using an independent sample set not involved in the modeling include a root mean square error and a correlation coefficient.

[0073] Optionally, the steps to establish a PLS correction model are as follows:

[0074] 1. Standardize or normalize the data to reduce the impact of different ranges on the model.

[0075] 2. Use the PLS method to decompose X into a component matrix, including a load matrix P and a score matrix T. Similarly, y is decomposed into a load vector q and a score vector U.

[0076] 3. Select a new score vector to maximize the projection of X and y. Update P, T, U and q. And calculate the regression coefficient b and the residual.

[0077] 4. Select the optimal number of components k according to the cross-validation method (such as K-fold cross-validation) to ensure that the model is not overfitting or underfitting.

[0078] Furthermore, the steps of adjusting model parameters include:

[0079] 1. Determine the number of components k: Use cross validation to select the appropriate number of components k. Under different numbers of components, calculate the prediction error of the model (such as root mean square error RMSE) and select the number of components with the minimum error.

[0080] 2. Try different data preprocessing methods such as standardization, normalization, first-order derivatives, or second-order derivatives to get the best model performance.

[0081] 3. Choose an appropriate wavelength range to avoid wavelength information redundancy or noise. And use methods such as principal component analysis (PCA) or greedy algorithms for feature selection.

[0082] 4. Cross-validation and evaluation: Use K-fold cross-validation or leave-one-out cross-validation to evaluate the model performance under different parameter combinations. Select the parameter combination that minimizes the cross-validation error.

[0083] 5. Model Validation: The final model is evaluated using an independent validation set to ensure predictive power and robustness on real data.

[0084] Corresponding to the above method, this embodiment further provides a system for identifying the fat content of broiler chickens, comprising:

[0085] A sample collection unit, used to collect broiler samples from various parts of broiler chickens of different age groups, and obtain the actual fat content of the broiler samples;

[0086] A sample pretreatment unit, used for grinding, homogenizing and dehydrating the collected samples to obtain pretreated samples;

[0087] A scanning unit, used to scan the pre-processed samples using a near infrared spectrometer to obtain spectral data of each sample;

[0088] A model building unit, used to associate the spectral data with the actual fat content using a multiple linear regression method to build a correction model;

[0089] A verification unit, used to verify the calibration model using an independent sample set that does not participate in modeling, to obtain a verified recognition model;

[0090] The recognition unit is used to input the spectral data of the broiler to be tested into the recognition model to obtain a fat recognition result.

[0091] The beneficial effects of the present invention are as follows:

[0092] (1) The NIR technology in this embodiment can complete a detection within a few seconds, greatly improving the detection efficiency.

[0093] (2) The present invention does not require destructive treatment of the sample, thus maintaining the integrity and original state of the sample.

[0094] (3) The automated equipment and reagent-free operation of the present invention significantly reduce the detection cost.

[0095] (4) The system of the present invention can be integrated into a production line to achieve online real-time monitoring and automatic control, thereby improving the level of intelligent production.

[0096] In summary, the present invention provides an efficient, convenient and economical method and system for identifying the fat content of broilers by combining near-infrared spectroscopy technology with an intelligent system, which has broad application prospects and market value.

[0097] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0098] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for identifying the fat content of broiler chickens, characterized in that: include: Collecting broiler samples from various parts of broiler chickens of different age groups, and obtaining actual fat content of the broiler samples; Grinding, homogenizing and dehydrating the collected samples to obtain pre-treated samples; Scanning the pre-treated samples using a near infrared spectrometer to obtain spectral data of each sample; Using a multiple linear regression method, the spectral data is correlated with the actual fat content to establish a calibration model; Using an independent sample set that does not participate in modeling to verify the correction model, to obtain a verified recognition model; The spectral data of the broiler to be tested is input into the recognition model to obtain the fat recognition result.

2. The method for identifying the fat content of broiler chickens according to claim 1, characterized in that: The wavelength range of the spectral data is 900-2500nm.

3. The method for identifying the fat content of broiler chickens according to claim 1, characterized in that: After scanning the pre-processed samples using a near infrared spectrometer to obtain spectral data of each sample, the method further includes: Shifting the spectral data of each sample to obtain shifted spectral data; Performing wavelet decomposition on the shifted spectral data to obtain a plurality of wavelet coefficients; Determine the filtering threshold according to the decomposition scale and length of the signal; Constructing a denoising function according to the filtering threshold; The denoising function is used to remove noise from the shifted spectral data to obtain denoised spectral data.

4. The method for identifying the fat content of broiler chickens according to claim 3, characterized in that: The formula for the filtering threshold is: Among them, λ is the filtering threshold, j is the decomposition scale of the spectral data after translation, and d j is the decomposition of wavelet coefficients with scale j, N represents the length of the shifted spectral data, and median represents the median operation.

5. The method for identifying the fat content of broiler chickens according to claim 4, characterized in that: Using the denoising function to remove the noise of the shifted spectral data to obtain denoised spectral data includes: The denoising function is used to remove the corresponding wavelet coefficients to obtain smoothed spectral data; wherein the denoising function is: Among them, X represents the wavelet coefficient, n represents the adjustment coefficient; Performing inverse translation on the smoothed spectral data to obtain inverse translated spectral data; The average of each inverse-translated spectral data is calculated to obtain the denoised spectral data.

6. The method for identifying the fat content of broiler chickens according to claim 1, characterized in that: The formula of the correction model is: y=Xb+e; wherein y represents the actual fat content of each sample; X is a feature matrix; the size of X is N×M; wherein N is the number of samples, and M is the number of spectral wavelengths; b is a regression coefficient vector with a size of M×1; and e is a residual vector with a size of N×1.

7. The method for identifying the fat content of broiler chickens according to claim 1, characterized in that: Evaluation parameters for validating the calibration model using an independent sample set that did not participate in the modeling included root mean square error and correlation coefficient.

8. A system for identifying fat content of broiler chickens, characterized in that: include: A sample collection unit, used to collect broiler samples from various parts of broiler chickens of different age groups, and obtain the actual fat content of the broiler samples; A sample pretreatment unit, used for grinding, homogenizing and dehydrating the collected samples to obtain pretreated samples; A scanning unit, used to scan the pre-processed samples using a near infrared spectrometer to obtain spectral data of each sample; A model building unit, used to associate the spectral data with the actual fat content using a multiple linear regression method to build a correction model; A verification unit, used to verify the calibration model using an independent sample set that does not participate in modeling, to obtain a verified recognition model; The recognition unit is used to input the spectral data of the broiler to be tested into the recognition model to obtain a fat recognition result.

Citation Information

Cited By

  • Rapid detection method for fatty acid content of Beijing duck based on near infrared spectrum technology

    CN119643501A