CN103674851A - Meat quality detection method - Google Patents

Abstract

The invention discloses a rapid and nondestructive meat quality detection method. The method comprises the following steps: acquiring reflection spectrums of various different microscopic structures of a meat sample to be detected by utilizing a microscopic spectrum system, and combining the spectroscopic data into a synthetic spectrum; selecting a characteristic wavelength combination which can reflect a predictive index in the synthetic spectrum by utilizing a stepwise regression method, establishing a multiple linear regression prediction model by using the characteristic wavelengths, and judging the meat quality index by using the prediction model. According to the method, the meat quality parameters can be rapidly and nondestructively detected, so that the detection means for quality of the meats in China is in line with that of the developed countries.

Description

A kind of measurement techniques for quality detection of meat detection method

Technical field

The invention belongs to quality of agricultural and poultry products detection field, being used for raiseeing meat detects, the non-destruction fast detecting of the microspectrum detection method that relates to a kind of measurement techniques for quality detection of meat, more specifically, relate to the method for spectra collection, Spectral beam combining, feature extraction and the Q factor detection of Meat Surface microstructure.

Technical background

China is production, the consumption big country of various meat products, and annual production, consumption figure are also in continuous growth.But various quality problems also emerge in an endless stream, the problem meat such as clenbuterol hydrochloride, the meat of dying of illness, water-injected meat are also perplexing consumers in general.The present invention proposes a kind of method that fast detecting measurement techniques for quality detection of meat based on microspectrum technology detects.

At present, the main appraisal procedure of the Quality Detection that Meat Industry is used has: 1) hedonic scoring system, by the personnel through professional training, sample is carried out to grade assessment, and this method is subject to the impact of the subjective factor of evaluating member; 2) equipment rating method, as used boxshear apparatus to measure the tender degree value of beef, detects the moisture of beef with infrared drying equipment, these methods detect more consuming time, and sample is had to destructiveness, are not suitable for online detection; 3) spectrum detection technique, as a kind of emerging technology, spectrum detection technique is widely used in every profession and trade, and spectrum detection technique comprises high spectrum, microspectrum, fluorescence spectrum, multispectral etc.

Spectral technique utilizes light to realize the detection to sample parameters at absorption, reflection and the scattering signatures of sample inside.The absorption of light is relevant with the chemical composition of sample, and the reflection of light and scattering are mainly determined by the architectural characteristic of sample, so spectral technique can be for the Q factor of working sample.

The present invention is the measurement techniques for quality detection of meat detection method based on microspectrum technology, detection method compared with other based on spectral technique, this method is each micro-composition tissue impact on its Q factor of meat respectively, and the spectrum of each structural constituent is synthesized to predict the Q factor of sample, to improve accuracy of detection and stability, there is quick nondestructive and detect characteristic, all departments and manufacturing enterprise are expanded sensing range, carry out grading and be of great immediate significance.

Summary of the invention

The object of this invention is to provide a kind of microspectrum detection method detecting for measurement techniques for quality detection of meat, the quick nondestructive of realizing measurement techniques for quality detection of meat parameter detects.

For achieving the above object, the invention provides a kind of method for quick of measurement techniques for quality detection of meat parameter, comprise the following steps:

S1, microspectrum data acquisition: utilize microspectrograph to obtain the microspectrum data I of the different tissues such as poultry meat sample surfaces muscle fiber to be measured, adipose tissue, muscle segment tissue ₀, I ₂, I ₃... etc.;

Synthesizing of S2, spectroscopic data: by the microspectrum data I of different tissues ₀, I ₂, I ₃... wait and synthesize synthetic spectrum I ₀;

Choosing of S3, characteristic wavelength: adopt stepwise regression method from synthetic spectrum I ₀in obtain the optimal wavelength that can characterize testing sample Q factor;

S4, testing sample Q factor detect: use synthetic spectrum I ₀in obtain the optimal wavelength that can characterize testing sample Q factor and set up multiple regression mathematical model, detect the Q factor of testing sample.

The spectrum I of each microstructure of the poultry meat sample to be measured wherein, obtaining in described step S1 ₀, I ₂, I ₃... waiting is resulting spectroscopic data after being proofreaied and correct by the spectrum extraction software of microspectrograph, and software is proofreaied and correct and comprised black, white reference correction.

Wherein, in described step S2, the formula of synthetic spectrum is:

$I_0=a+\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{b}_i\×I_i$

I wherein ₀for synthetic spectrum, a and b _ifor composite coefficient, i is for extracting the number of the microstructure of spectrum, and i begins=1,2 ..., n, I _ibe the microspectrum data of i kind microstructure.

Wherein, the multiple linear regression mathematical model of setting up in described step S4 is:

$F=f_0+\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{f}_j{X}_j$

Wherein, F is predicted parameter value; f ₀and f _jregression equation coefficient, j=1,2 ..., m; M is the number of preferred feature wavelength in model; X _jit is the reflectance value of selected j characteristic wavelength.

Embodiment

Below the specific embodiment of the present invention is described in further detail.Once embodiment is only for the present invention is described, but is not used for limiting the scope of the invention.

Microspectrum method for quick according to the measurement techniques for quality detection of meat of the embodiment of the present invention comprises the following steps:

S1, microspectrum data acquisition: utilize microspectrograph to obtain the microspectrum data I of the different tissues such as poultry meat sample surfaces muscle fiber to be measured, adipose tissue, muscle segment tissue ₁, I ₂, I ₃... etc., spectroscopic data I wherein ₁, I ₂, I ₃... waiting is resulting spectroscopic data after being proofreaied and correct by the spectrum extraction software of microspectrograph, and software is proofreaied and correct and comprised black, white reference correction.

Synthesizing of S2, spectroscopic data: by the microspectrum data I of different tissues ₁, I ₂, I ₃... wait and synthesize synthetic spectrum I ₀, the formula of synthetic spectrum is:

$I_0=a+\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{b}_i\×I_i$

I wherein ₀for synthetic spectrum, a and b _ifor composite coefficient, i is for extracting the number of the microstructure of spectrum, and i begins=1,2 ..., n, I _ibe the microspectrum data of i kind microstructure.

Choosing of S3, characteristic wavelength: adopt stepwise regression method from synthetic spectrum I ₀in obtain the optimal wavelength that can characterize testing sample Q factor.

S4, testing sample Q factor detect: use synthetic spectrum I ₀in obtain the optimal wavelength that can characterize testing sample Q factor and set up multiple regression mathematical model, detect the Q factor of testing sample.The multiple linear regression mathematical model of setting up is:

$F=f_0+\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{f}_j{X}_j$

Wherein, F is predicted parameter value; f ₀and f ₁regression equation coefficient, j=1,2 ..., m; M is the number of preferred feature wavelength in model; X _jit is the reflectance value of selected j characteristic wavelength.

Claims (4)

1. a breeding stock meat parameter detection method, is characterized in that comprising the following steps:

S1, microspectrum data acquisition: utilize microspectrograph to obtain the microspectrum data I of the different tissues such as poultry meat sample surfaces muscle fiber to be measured, adipose tissue, muscle segment tissue ₁, I ₂, I ₃... etc.;

Synthesizing of S2, spectroscopic data: by the microspectrum data I of different tissues ₁, I ₂, I ₃deng synthesizing synthetic spectrum I ₀;

Choosing of S3, characteristic wavelength: adopt stepwise regression method from synthetic spectrum I ₀in obtain the optimal wavelength that can characterize testing sample Q factor;

S4, testing sample Q factor detect: use synthetic spectrum I ₀in obtain the optimal wavelength that can characterize testing sample Q factor and set up multiple regression mathematical model, detect the Q factor of testing sample.

2. as the poultry meat parameter detection method under claim 1, it is characterized in that the spectrum I of each microstructure of the poultry meat sample to be measured obtaining in affiliated step S1 ₁, I ₂, I ₃deng be by the spectrum of microspectrograph, extract software and proofread and correct after resulting spectroscopic data, software proofread and correct comprise black, white with reference to correction.

3. as the poultry meat parameter detection method under claim 1, it is characterized in that, in affiliated step S2, the formula of synthetic spectrum is:

I wherein ₀for synthetic spectrum, a and b _ifor composite coefficient, i is for extracting the number of the microstructure of spectrum, and i begins=1,2 ..., n, I _ibe the microspectrum data of i kind microstructure.

4. as the poultry meat parameter detection method under claim 1, it is characterized in that, the multiple linear regression mathematical model of setting up in affiliated step S4 is:

Wherein, F is predicted parameter value; f ₀and f ₁regression equation coefficient, j=1,2 ..., m; M is the number of preferred feature wavelength in model; X _jit is the reflectance value of selected j characteristic wavelength.