CN102879432A - Method of using electronic nose for detecting freshness of tilapia - Google Patents

Abstract

A method of using an electronic nose for detecting freshness of tilapia includes the steps: (1) respectively placing back and belly samples of tilapia to be detected into a sample container, and standing for 20-40min at the room temperature; (2) inserting a sample injection pinhead of the electronic nose into the sample container at the normal temperature, inserting an air filter filled with activated carbon into the sample container to balance air pressure, starting the electronic nose, and enabling the electronic nose to suck gas given out by the standing samples inside; (3) detecting the gas given out by the samples by the aid of a gas sensor array inside an electronic nose gas chamber, wherein detecting time is 40-60s, and cleaning time is 3-8min; (4) setting an experimental period interval as 10-12 hours; and (5) collecting data acquired by the gas sensor array, performing data processing for the acquired data, and computing so that a freshness value of the tilapia is obtained. The method is simple, convenient, rapid and effective.

Description

A kind of method of utilizing detection by electronic nose Tilapia mossambica freshness

Technical field

The present invention relates to the Estimation of The Fish Freshness detection field, more particularly, relate to a kind of method of utilizing detection by electronic nose Tilapia mossambica freshness based on gas sensor array.

Background technology

Tilapia mossambica (tilapia) is a kind of middle-size and small-size fish, is commonly called as African crucian.Its fine and tender taste, meat flavour is delicious, liked by citizen, is one delicacies on the dining table.Tilapia mossambica is rich in protein and multiple unsaturated fatty acid, and it is " protein sources that does not need protein " for Japan person, it will be following animal protein mainly obtain one of object.The Tilapia mossambica breeding is fast, and premunition is strong, and growth is rapid, and cost is low, and output is high, is the good social connections that farmers' strives for a relatively comfortable life and gets rich, and main cultured fishes have become international

Freshness is to judge an important indicator of fish or fish product quality, directly affects the final mass of product.Now develop a series of technical method and estimated fish freshness, such as sensory evaluation method, micro-biological process, physics and chemical method etc., but the difficult satisfied accurately and rapidly requirement of above detection method, ubiquity the shortcomings such as loaded down with trivial details, the consuming time length of detection mode, need professional.

At present, in problems such as the gas sensor ubiquity baseline wander of Electronic Nose, poor repeatability, affect the accuracy of testing result, hamper the development of Electronic Nose Technology.In addition, the recognition methods of traditional mode mainly comprises principal component analysis (PCA), factorial analysis, cluster analysis, partial least square method, artificial neural network etc., but all there is limitation to a certain extent in these methods.Principal component analysis (PCA) to the several variablees of minority, is assigned to estimate the contribute information influence power of raw data with original multidimensional data dimensionality reduction by calculating the major component function call.But just have when negative when the symbol of the factor loading of major component has, the evaluation function meaning with regard to indefinite and name sharpness at the bottom of.Factorial analysis restructuring original variable information is seeking the common factor of variation, but the method adopts least square method when the calculated factor score, and the method may lose efficacy.Cluster analysis obtains comparatively difficulty of cluster result when sample size is larger.The problem of regression modeling when partial least square method can effectively overcome sample size and is lower than the variable number, but exist the impact point that regression result was lost efficacy.Artificial neural network the problem that training effectiveness reduces, forecast precision descends can occur in the more situation of input sample, need repetition learning.Simultaneously, above mode identification method can be distinguished sample qualitatively, distinguishes information but substantially can't quantize sample.

Summary of the invention

Loaded down with trivial details for the process of the method that overcomes existing Tilapia mossambica freshness, cost is high, the deficiency of length consuming time, sensor baseline wander etc., the invention provides a kind of easy, utilize the method for detection by electronic nose Tilapia mossambica freshness fast and effectively.

The present invention is that the concrete technical scheme that adopts that achieves the above object is:

A kind of method of utilizing detection by electronic nose Tilapia mossambica freshness said method comprising the steps of:

(1) gets Tilapia mossambica back and belly, be housed in respectively under 3 ~ 5 ℃ of environment;

(2) sampling receptacle is put into respectively at Tilapia mossambica to be detected back and belly sample, left standstill 20 ~ 40min;

(3) under the normal temperature sample introduction syringe needle of Electronic Nose is inserted sampling receptacle, the air cleaner that activated charcoal will be housed simultaneously inserts sampling receptacle with equilibrium air pressure, and the unlocking electronic nose sucks the Electronic Nose air chamber with the gas that the sample that leaves standstill gives out;

(4) gas sensor array in the Electronic Nose air chamber detects the gas that sample gives out, and be 40 ~ 60s detection time, and scavenging period is 3 ~ 8min;

(5) experimental period is spaced apart 10 ~ 12 hours;

(6) collect the data that gas sensor array gathers, the data that gather are carried out data process, detailed process is as follows:

(6-1) the response characteristic value of extraction gas sensor array, the response characteristic value comprises initial value (Vs), stationary value (Ve), rise time (Ti), climbing speed

(6-2) non-linear stochastic is resonated algorithm solidifies in the built-in digital signal processor of Electronic Nose, and Electronic Nose is responded primitive character value input DSP to calculate output signal-to-noise ratio spectrum signature value; Computing method are as follows:

System input signal is System performance is expressed as:

Wherein a and b are the parameters of potential function, and x is the position of Brownian movement particle, and ξ (t) is white Gaussian noise, and its autocorrelation function is: E[ξ (t) ξ (0)]=2D δ (t), A is input signal strength, f ₀Be frequency modulating signal, D is noise intensity,

It is a real parameter;

Signal to noise ratio (S/N ratio) is to characterize accidental resonance characteristic parameter commonly used, with signal to noise ratio (S/N ratio) eigenwert Eig _SNRBe defined as:

${\mathrm{Eig}}_{\mathrm{SNR}}=2\left[\underset{\mathrm{\Δ\ω}\→0}{\lim }\underset{\mathrm{\Ω}-\mathrm{\Δ\ω}}{\overset{\mathrm{\Ω}+\mathrm{\Δ\ω}}{\∫}}S\left(\mathrm{\ω}\right)\mathrm{d\ω}\right]/S_N\left(\mathrm{\Ω}\right)---\left(4\right)$

S (ω) representation signal power spectrum density wherein, S _N(Ω) be the intensity of noise in the signal frequency zone, ω is frequency corresponding to power spectrum spike, and Ω is frequency corresponding to accidental resonance noise;

(6-3) according to the freshness forecast model, calculate the Tilapia mossambica grade of freshness according to signal to noise ratio (S/N ratio) spectrum signature value.

In the described step (3), the gas flow by the Electronic Nose air chamber is 800 ~ 1200ml/min.

Further, in the step (6-3), use the Electronic Nose experiment to detect the signal to noise ratio (S/N ratio) eigenwert Eig of the tilapia mossambica samples of the different freshnesss of many groups _SNR

Signal to noise ratio (S/N ratio) eigenwert match according to tilapia mossambica samples obtains Tilapia mossambica freshness forecast model:

Tq=f(Eig _SNR)

Wherein, Eig _SNRBe the signal to noise ratio (S/N ratio) eigenwert, Tq is the Tilapia mossambica grade of freshness.

Described method is further comprising the steps of:

If (6-4) Tilapia mossambica grade of freshness Tq≤T1, T1 are shelf life terminal point critical value, then be fresh; If Tilapia mossambica grade of freshness Tq＞T1, then tilapia mossambica samples has reached the shelf life terminal point.

The measuring process of described shelf life terminal point critical value T1 is as follows:

1),, the index of freezing in animality aquatic products hygienic standard GB 2733-2005 total volatile basic nitrogen bright according to country, determine the grade of freshness of different resting period tilapia mossambica samples.Wherein the freshness grade is divided into: total volatile basic nitrogen (mg/100g)≤20 is fresh; Total volatile basic nitrogen (mg/100g)＞20 is rotten;

2), according to the total volatile basic nitrogen value of the different resting period Tilapia mossambica backs of the operational measure in the mensuration file of SC T 3032-2007 total volatile basic nitrogen in fishery products and belly sample, determine the time critical values whether tilapia mossambica samples goes bad, i.e. shelf life terminal time critical value S1;

3), Tilapia mossambica freshness parameter Tq and the time critical values of tilapia mossambica samples that Electronic Nose is recorded carry out related, to determine shelf life terminal point critical value T1.

Described Tilapia mossambica freshness forecast model is:

$\mathrm{Tq}=\frac{\ln \frac{{\mathrm{Eig}}_{\mathrm{SNR}}+77.23616}{1.31724}}{0.36765}.$

In the described step (6-2), the non-linear algorithm that resonates immediately is as follows:

Stochastic resonance system comprises three factors: bistable system, and input signal and external noise source, usually in the bistable state potential well, come descriptive system characteristic by power-actuated overdamping Brownian movement of cycle particle with one:

Wherein, V (x) is non-linear symmetric potential function, and ξ (t) is white Gaussian noise, and its autocorrelation function is: E[ξ (t) ξ (0)]=2D δ (t), A is input signal strength, f ₀Be frequency modulating signal, D is noise intensity, Be a real parameter, a and b are the parameters of potential function, and x is the position of Brownian movement particle;

$V\left(x\right)=\frac{1}{4}{\mathrm{ax}}^4-\frac{1}{2}{\mathrm{bx}}^2---\left(2\right)$

Therefore formula (1) can be rewritten as:

Signal to noise ratio (S/N ratio) is to characterize accidental resonance characteristic parameter commonly used, and we are with signal to noise ratio (S/N ratio) eigenwert Eig _SNRBe defined as:

${\mathrm{Eig}}_{\mathrm{SNR}}=2\left[\underset{\mathrm{\Δ\ω}\→0}{\lim }\underset{\mathrm{\Ω}-\mathrm{\Δ\ω}}{\overset{\mathrm{\Ω}+\mathrm{\Δ\ω}}{\∫}}S\left(\mathrm{\ω}\right)\mathrm{d\ω}\right]/S_N\left(\mathrm{\Ω}\right)---\left(4\right)$

S (ω) representation signal power spectrum density wherein, S _N(Ω) be the intensity of noise in the signal frequency zone, ω is frequency corresponding to power spectrum spike, and Ω is frequency corresponding to accidental resonance noise;

Detection principle of the present invention is: bionic olfactory system (being Electronic Nose) mainly is comprised of smell sampling operation device, gas sensor array and three kinds of function elements of signal processing system.Be compared to people's olfactory system, it more has its superiority.It can avoid personal error, good reproducibility; Can also detect the gas that some noses can not detect.Tilapia mossambica is because of its high-low temperature resistant not, so denaturalization phenomenon easily occurs in storage, consumer's health produced certain impact.Electronic Nose can be monitored the smell situation of ad-hoc location continuously, in real time within the time of several hours, several days even several months, follow the tracks of its quality situation, and feedback has constantly overcome the deficiency of human olfactory to a certain extent.In recent years, the detection by electronic nose technology is widely used in by Chinese scholars in the freshness research of the food such as milk, cereal, meat.

The gas that tested tilapia mossambica samples volatilizes acts on the Electronic Nose sensor array.Before this smell was presented on a kind of sensor cover of active material, sensor converted chemistry input to electric signal, by a plurality of sensors a kind of response of smell had just been consisted of the response spectra of sensor array to this smell.It adopts Metal Oxide Gas Sensors core devices array, and is auxiliary with temperature-humidity sensor and high-accuracy data acquisition platform, realizes Tilapia mossambica Data Detection, storage and transmission.Its escaping gas composition of tilapia mossambica samples of different freshnesss is also different, follow the tracks of the variation of tilapia mossambica samples escaping gas with Electronic Nose, process the detection by electronic nose signal and extract the freshness eigenwert with non-linear stochastic resonance, realize the target that the Tilapia mossambica freshness detects.Set up Tilapia mossambica quality model, realize the Quality Detection to the Tilapia mossambica of different freshness, make its cultivation that applies to Tilapia mossambica and storage industry, better promoted.

Therefore, the present invention has following beneficial effect:

(1) detection method of the present invention adopts the Electronic Nose that is comprised of 8 class sensors that tilapia mossambica samples is detected, as long as test detects the Electronic Nose response of tilapia mossambica samples, and the signal to noise ratio (S/N ratio) eigenwert that calculates sample just can realize the purpose that the tilapia mossambica samples freshness detects, and sets up Tilapia mossambica quality model.Analytic process is simple, and amount of samples is little, has the advantages such as convenient, quick, cheap, accurate;

(2) detection method of the present invention need not tilapia mossambica samples is made muddy flesh, has simplified the detection operating process, and has reduced Tilapia mossambica and made the processing procedure of muddy flesh to the impact of testing result, and easy operating does not need very professional technician;

(3) directly by the sensor array of detection, gas can contact with sensor detection method gas of the present invention fully, and detection time is short;

(4) cost of the present invention is low, is easy to penetration and promotion, and testing result is objective, accurate, quick, good reproducibility.

Description of drawings

Fig. 1 is the process flow diagram that utilizes the method for detection by electronic nose Tilapia mossambica freshness of the present invention.

Embodiment

The present invention will be further described below in conjunction with the drawings and specific embodiments.

With reference to Fig. 1, a kind of method of utilizing detection by electronic nose Tilapia mossambica freshness, the electric nasus system of employing mainly comprises data acquisition, conditioning and transmission unit, sensor array and air chamber, and for the jet power unit three parts, and is furnished with the air cleaner that activated charcoal is housed.Data acquisition, conditioning and transmission unit core devices adopt embedded microcontroller, can realize sensor array signal collection, transmission, and pump valve is opened the functions such as control.Comprise the parts such as gas samping pump, scavenging pump, solenoid valve for jet power unit.When the Electronic Nose test macro was tested, the air chamber scavenging pump was sent the pure air that filter cleaner produces into sensor air chamber, calibrating sensors.Under the drive of gas samping pump, draw the gas that sample volatilizes in the container with the sample introduction syringe needle and enter the sensor air chamber afterwards, produce response signal after contacting with sensor, discharge by the gas outlet at last.The sensor response signal by portable computer record, preservation and processing, finally shows the grade of freshness of this sample through data acquisition unit.

A kind of method of utilizing detection by electronic nose Tilapia mossambica freshness of present embodiment, the tilapia mossambica samples of 4 ℃ of storages of detection.

(1) Tilapia mossambica belly to be detected and back sample grab sample 10g are positioned in the 50ml sample bottle, 5 parallel laboratory tests are set, each experiment repeats 3 times, leaves standstill 30min;

(2) before test sample, will by scavenging pump the pure air that filter cleaner produces be passed into the sensor air chamber under the normal temperature first, response is calibrated to sensor.Calibrate and draw the gas that the sample that left standstill 30min volatilizes with the sample introduction syringe needle of Electronic Nose at normal temperatures after complete, the absorption time is 50s, cleans 5mim, and volatilization gas information is detected by the gas sensor array in the Electronic Nose device.

(3) first sensor is sulfide sensor TGS-825 in the collection gas sensor array, the second sensor is hydrogen gas sensor TGS-821, the 3rd sensor is ammonia gas sensor TGS-826, four-sensor is alcohol, toluene, the sensor TGS-822 such as dimethylbenzene, the 5th sensor is hydrocarbon component gas (the sensor TGS-842 of C1～C8), the 6th sensor namely for methane, propane, butane sensor TGS-813, the 7th sensor is propane, butane sensor TGS-2610, the 8th sensor is the data that NOx sensor TGS-2210 gathers.

From the data obtained, extract initial value (Vs), stationary value (Ve), rise time (Ti), the climbing speed of Electronic Nose response data 4 primitive character values are input to the dsp chip that is solidified with non-linear stochastic resonance program in the electric nasus system, calculate the signal to noise ratio (S/N ratio) eigenwert Eig of output tilapia mossambica samples _SNR=-61.14835.

The signal to noise ratio (S/N ratio) eigenwert of each the freshness tilapia mossambica samples that (4) obtains according to step (3) obtains Tilapia mossambica freshness forecast model through match:

$\mathrm{Tq}=\frac{\ln \frac{{\mathrm{Eig}}_{\mathrm{SNR}}+70.21967}{4.15844}}{0.70521}---\left(1\right)$

With Eig _SNRx=-61.14835 substitution formula, calculating Tilapia mossambica freshness parameter value is Tqe=1.09.

(5) T1=1.09; Because Tq＜7, therefore tested tilapia mossambica samples is judged to be freshly, not yet reaches the shelf life terminal point.

Claims (6)

1. method of utilizing detection by electronic nose Tilapia mossambica freshness is characterized in that: said method comprising the steps of:

(1) gets Tilapia mossambica back and belly, be housed in respectively under 3 ~ 5 ℃ of environment;

(2) sampling receptacle is put into respectively at Tilapia mossambica to be detected back and belly sample, left standstill 20 ~ 40min;

(3) under the normal temperature sample introduction syringe needle of Electronic Nose is inserted sampling receptacle, the air cleaner that activated charcoal will be housed simultaneously inserts sampling receptacle with equilibrium air pressure, and the unlocking electronic nose sucks the Electronic Nose air chamber with the gas that the sample that leaves standstill gives out;

(4) gas sensor array in the Electronic Nose air chamber detects the gas that sample gives out, and be 40 ~ 60s detection time, and scavenging period is 3 ~ 8min;

(5) experimental period is spaced apart 10 ~ 12 hours;

(6) collect the data that gas sensor array gathers, the data that gather are carried out data process, detailed process is as follows:

(6-1) the response characteristic value of extraction gas sensor array, the response characteristic value comprises initial value (Vs), stationary value (Ve), rise time (Ti), climbing speed

(6-2) non-linear stochastic is resonated algorithm solidifies in the built-in digital signal processor of Electronic Nose, with Electronic Nose response characteristic value input DSP to calculate output signal-to-noise ratio spectrum signature value; Computing method are as follows:

System input signal is

System performance is expressed as:

Wherein a and b are the parameters of potential function, and x is the position of Brownian movement particle, and ξ (t) is white Gaussian noise, and its autocorrelation function is: E[ξ (t) ξ (0)]=2D δ (t), A is input signal strength, f ₀Be frequency modulating signal, D is noise intensity,

It is a real parameter;

Signal to noise ratio (S/N ratio) is to characterize accidental resonance characteristic parameter commonly used, and signal to noise ratio (S/N ratio) eigenwert EigSNR is defined as:

${\mathrm{Eig}}_{\mathrm{SNR}}=2\left[\underset{\mathrm{\Δ\ω}\→0}{\lim }\underset{\mathrm{\Ω}-\mathrm{\Δ\ω}}{\overset{\mathrm{\Ω}+\mathrm{\Δ\ω}}{\∫}}S\left(\mathrm{\ω}\right)\mathrm{d\ω}\right]/S_N\left(\mathrm{\Ω}\right)---\left(4\right)$

S (ω) representation signal power spectrum density wherein, S _N(Ω) be the intensity of noise in the signal frequency zone, ω is frequency corresponding to power spectrum spike, and Ω is frequency corresponding to accidental resonance noise;

(6-3) according to the freshness forecast model, calculate the Tilapia mossambica grade of freshness according to signal to noise ratio (S/N ratio) spectrum signature value.

2. the method for utilizing detection by electronic nose Tilapia mossambica freshness according to claim 1 is characterized in that: in the described step (4), the gas flow by the Electronic Nose air chamber is 800 ~ 1200ml/min.

3. the method for utilizing detection by electronic nose Tilapia mossambica fish freshness according to claim 1 and 2 is characterized in that, in the step (6-3), uses the Electronic Nose experiment to detect the signal to noise ratio (S/N ratio) eigenwert Eig of the tilapia mossambica samples of the different freshnesss of many groups _SNR

Signal to noise ratio (S/N ratio) eigenwert match according to tilapia mossambica samples obtains Tilapia mossambica freshness forecast model:

Tq=f(Eig _SNR)

Wherein, Eig _SNRBe the signal to noise ratio (S/N ratio) eigenwert, Tq is the Tilapia mossambica grade of freshness.

4. the method for utilizing detection by electronic nose Tilapia mossambica freshness according to claim 1 and 2, it is characterized in that: described method is further comprising the steps of:

If (6-4) Tilapia mossambica grade of freshness Tq≤T1, T1 are shelf life terminal point critical value, then be fresh; If Tilapia mossambica grade of freshness Tq＞T1, then tilapia mossambica samples has reached the shelf life terminal point.

5. the method for utilizing detection by electronic nose Tilapia mossambica freshness according to claim 4, it is characterized in that: the measuring process of described shelf life terminal point critical value T1 is as follows:

1),, the index of freezing in animality aquatic products hygienic standard GB 2733-2005 total volatile basic nitrogen bright according to country, determine the grade of freshness of different resting period tilapia mossambica samples.Wherein the freshness grade is divided into: total volatile basic nitrogen (mg/100g)≤20 is fresh; Total volatile basic nitrogen (mg/100g)＞20 is rotten;

2), according to the total volatile basic nitrogen value of the different resting period Tilapia mossambica backs of the operational measure in the mensuration file of SC T 3032-2007 total volatile basic nitrogen in fishery products and belly sample, determine the time critical values whether tilapia mossambica samples goes bad, i.e. shelf life terminal time critical value S1;

3), Tilapia mossambica freshness parameter Tq and the time critical values of tilapia mossambica samples that Electronic Nose is recorded carry out related, to determine shelf life terminal point critical value T1.

6. the method for utilizing detection by electronic nose Tilapia mossambica freshness according to claim 3, it is characterized in that: described Tilapia mossambica freshness forecast model is:

$\mathrm{Tq}=\frac{\ln \frac{{\mathrm{Eig}}_{\mathrm{SNR}}+70.21967}{4.15844}}{0.70521}.$

Images