CN105044023A - Method for quickly detecting content of benzoyl peroxide in flour in nondestructive mode and application thereof - Google Patents
Method for quickly detecting content of benzoyl peroxide in flour in nondestructive mode and application thereof Download PDFInfo
- Publication number
- CN105044023A CN105044023A CN201510477873.6A CN201510477873A CN105044023A CN 105044023 A CN105044023 A CN 105044023A CN 201510477873 A CN201510477873 A CN 201510477873A CN 105044023 A CN105044023 A CN 105044023A
- Authority
- CN
- China
- Prior art keywords
- sample
- benzoyl peroxide
- flour
- utilize
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses a method for quickly detecting the content of benzoyl peroxide in flour in a nondestructive mode and an application thereof, and belongs to the technical field of food safety quality detection. According to the method, a near-infrared spectroscopy is used for scanning flour samples, so that a sample near infrared spectroscopy curve is obtained; an initial detection model is established after the obtained spectroscopy curve is processed; the detection model is used for detecting the content of the benzoyl peroxide in the flour samples to be detected after the initial detection model is verified and calibrated. The method has the advantages of being high in detecting speed, low in cost, convenient to operate, environmentally friendly and suitable for being applied and popularized.
Description
Technical field
The present invention relates to method and the application of benzoyl peroxide (BPO) content in a kind of Fast nondestructive evaluation flour, in particular to a kind of method based on BPO content in near-infrared spectrum technique Fast nondestructive evaluation flour, belong to food security technical field of quality detection.
Background technology
Flour, as one of Main Foods raw material in the world, can be used for making the Flour product such as boiled dumpling, noodles, steamed bun.Flour product is made as master with boiling, various famous with color, and this whiteness to flour requires higher, and most consumers is also judge flour quality quality according to color and luster perception simultaneously.Wheat flour and its products is pure whiter, finer and smoother, and consumer is favor more, thus Flour whiteness extremely wheat processing enterprise attention.Owing to containing xenthophylls in wheat, the flour color that major part has just been ground out is partially yellow.Conjugated double bond in xenthophylls molecule is a chromophore, has the performance of extinction, can be slowly oxidized in atmosphere, and color can disappear gradually, and this needs a very long storage period, and at least want the several months long, this is commercially infeasible.A lot of flour mill caters to consumer demand in order to sales volume, and use bleaching agent for flour to increase the whiteness of flour widely, BPO is exactly a kind of often by the bleaching agent for flour used.BPO has very strong oxidisability, and it faster by the xenthophylls oxidation in flour, can make the yellow in flour disappear, thus play whitening effect.But add BPO too much and can destroy some intrinsic nutritional labelings in flour, as carrotene, vitamin A, Cobastab, vitamin E etc., long-term edible this kind of flour, the shortage of vitamin can be caused, thus cause some disease, as angular stomatitis, neuritis, serious even can cause the infringement of central nervous system cumulative bad, liver failure etc.To this, China has corresponding laws and regulations to retrain, and the original addition as BPO in GB2760-2007 regulation flour must not more than 60mgKg
-1.In March, 2011, the Ministry of Public Health of China issues up-to-date bulletin: forbade to add whitening agent from 1 day May in 2011 in flour.But fierceness market competition and chase the factors such as more high profit ordering about under, the small-sized flour enterprises of China Partial still at use BPO as bleaching agent for flour.
Near infrared spectrum (NearInfraredSpectroscopy, NIR) analytical technology is a swift and violent new and high technology of development, utilize the optical characteristics of chemical substance near infrared spectrum district, one or more chemical composition contents in certain sample of Fast Measurement and the physics determination techniques of characteristic.It utilizes laboratory sample to a kind of method of the absorption of light, scattering, the characteristic determination sample interior composition such as reflection and transmission.
Use traditional analytical approach to measure the multiple character of a sample or concentration data needs multiple analytical instrument, at substantial human and material resources and time, cost is high, inefficiency, far can not adapt to the demand of modern industry.Compared with conventional analytical techniques, near-infrared spectral analysis technology can in tens seconds even several seconds, by means of only the simple measurement of a near infrared spectrum to sample, just several even tens kinds of data of an energy Simultaneously test sample, and sample consumption is little, without destroying, pollution-free, have efficient, fast, the feature of the low and environmental protection of cost.
Near-infrared spectral analysis technology is as a kind of " green " Dynamic Non-Destruction Measurement, the advantages such as it has fast, harmless, pollution-free, in multiple fields such as agricultural, food industry, petrochemical complex and pharmaceutical engineering, near infrared detection has become the very effective method for quick of one that alternative chemical analysis detects, and be widely used in quantitative, the qualitative analysis of crops and secondary product thereof, play an important role.
Summary of the invention
In order to solve, prior art sense cycle is long, testing cost high-level efficiency is low, complex operation problem, the invention provides the method for Determination of benzoyl peroxide in wheat in a kind of Fast nondestructive evaluation flour, the technical scheme taked is as follows:
The object of the present invention is to provide the method for Determination of benzoyl peroxide in wheat in a kind of Fast nondestructive evaluation flour, the method is after setting up detection model, utilize near infrared spectrometer to scan flour sample to be measured, obtain sample near infrared light spectral curve, the gained curve of spectrum is processed, finally utilizes detection model to detect the content of benzoyl peroxide in flour sample to be measured.
The step of described method is as follows:
1) utilize near infrared spectrometer to scan flour sample, obtain sample near infrared light spectral curve;
2) utilize the Determination of benzoyl peroxide in wheat in high performance liquid chromatograph mensuration flour sample, set up Determination of benzoyl peroxide in wheat database;
3) to step 1) the near infrared light spectral curve of gained processes, obtain preprocessed data, using gained preprocessed data as X-axis data, simultaneously using step 2) data corresponding in the benzoyl peroxide database of gained as Y-axis data, and utilize SPXY (samplesetpartitioningbasedonjointx-ydistance) algorithm that sample is divided into calibration set and forecast set;
4) step 3 is utilized) gained calibration set data integrating step 2) set up Determination of benzoyl peroxide in wheat database, set up Preliminary detection model, recycling forecast set data are verified Preliminary detection model and correct, and obtain detection model;
5) step 4 is utilized) detection model of gained measures Determination of benzoyl peroxide in wheat in flour to be measured.
Preferably, step 1) the described near infrared spectrometer that utilizes scans flour sample, and be strike off after sample is loaded sample cup, scanning samples 5 times, sweep limit is 400-2498nm, and resolution is 2nm.
Preferably, step 2) utilize high performance liquid chromatograph to measure Determination of benzoyl peroxide in wheat in flour sample, mensuration process is, take 5g flour sample, be placed in 50mL color-comparison tube, after dripping 10.0mL methyl alcohol, 5min is left standstill after stirring 1min, at the potassium iodide aqueous solution 5.0mL of dropping 50%, 10min is left standstill after abundant stirring 1min, finally add water to 50.0mL, stirring utilizes the membrane filtration supernatant liquid of 0.22 μm after leaving standstill, get supernatant liquor after filtration to join in high performance liquid chromatograph and measure, during mensuration, mobile phase used is volume ratio is methyl alcohol: the mixed liquor of water=10:90, flow velocity is 1.0mL/min, sample size is 10.0 μ L, determined wavelength is 230nm.
Preferably, step 3) described gained near infrared light spectral curve to be processed, be utilize lever value and student residual error T to check rejecting abnormalities value, spectroscopic data is optimized in the conversion of recycling standard normal variable.
More preferably, the described standard normal variable that utilizes converts optimization spectroscopic data, utilizes principal component analysis (PCA) to carry out dimension-reduction treatment to spectroscopic data simultaneously.
Preferably, step 4) detection model, it is radial basis function (RBF) neural network model, the structure of RBF neural model is the curve of spectrum data to obtain after pre-service is input vector, with the corresponding Determination of benzoyl peroxide in wheat of high performance liquid chromatography actual measurement for desired output, square error is 0, SPREAD value is 1000, builds using Gaussian function as hidden layer basis function.
Preferably, step 5) the described detection model that utilizes measures Determination of benzoyl peroxide in wheat in flour to be measured, after utilizing near infrared spectrometer to scan testing sample, obtain preprocessed data after the gained curve of spectrum is processed, then utilize the Determination of benzoyl peroxide in wheat in detection model acquisition sample according to preprocessed data.
The concrete steps of described method are as follows:
1) strike off after flour sample being loaded sample cup, utilize near infrared spectrometer scanning samples 5 times, sweep limit is 400-2498nm, and resolution is 2nm, obtains sample near infrared light spectral curve;
2) 5g flour sample is taken, be placed in 50mL color-comparison tube, after dripping 10.0mL methyl alcohol, 5min is left standstill after stirring 1min, at the potassium iodide aqueous solution 5.0mL of dropping 50%, 10min is left standstill after abundant stirring 1min, finally add water to 50.0mL, stirring utilizes the membrane filtration supernatant liquid of 0.22 μm after leaving standstill, get supernatant liquor after filtration to join in high performance liquid chromatograph and measure, during mensuration, mobile phase used is volume ratio is methyl alcohol: the mixed liquor of water=10:90, flow velocity is 1.0mL/min, sample size is 10.0 μ L, determined wavelength is 230nm, detect the Determination of benzoyl peroxide in wheat terminated in rear acquisition flour sample, and set up benzoyl peroxide database,
3) utilize lever value and student residual error T to check and reject step 1) exceptional value in gained near infrared light spectral curve, spectroscopic data is optimized in the conversion of recycling standard normal variable, utilize principal component analysis (PCA) to carry out dimension-reduction treatment to spectroscopic data simultaneously, obtain preprocessed data, simultaneously in conjunction with benzoyl peroxide database, utilize SPXY algorithm that sample is divided into calibration set and forecast set;
4) utilize step 3) the calibration set data of gained integrating step 2) in benzoyl peroxide database, MATLAB is utilized to set up preliminary RBF neural model, recycling step 3) in the data of forecast set, with the evaluation index that prediction related coefficient, predicted root mean square error and relation analysis error are model, preliminary RBF neural model is verified and corrected, finally obtains detection model; Described detection model is: RBF neural forecast model, be specially: utilize newrbe (P, T, SPREAD) function creation accurate radial basis function network, comprising input layer, hidden layer and output layer, due to the network just about benzoyl peroxide, only has benzoyl peroxide variable layer so export.Pretreated spectrum data matrix X is as input vector P, the Determination of benzoyl peroxide in wheat value vector Y of the sample corresponding to these spectrum is as the desired output matrix T of model, square error is defaulted as 0, through repeatedly testing, when SPREAD value gets 1000, the effectiveness comparison of model is good, and hidden layer basis function adopts Gaussian function, a great advantage of RBF is that weights W is tried to achieve by linear least square method, and this makes RBF correcting algorithm have processing speed faster.
5) utilize near infrared spectrometer to scan flour sample to be measured, according to step 2) described in method the curve of spectrum is processed, then utilize step 4 according to gained preprocessed data) detection model obtains Determination of benzoyl peroxide in wheat in testing sample.
Described either method is for detecting the content of the benzoyl peroxide in flour.
Preferably, step 1) in utilize near-infrared spectrometers to obtain the method for spectrogram to be: scan temperature during scanning and control room temperature about 25 DEG C; During dress sample, sample loads after sample cup needs rim of a cup to strike off, and makes sample surfaces equal with sample cup edge as far as possible, avoids that sample surfaces is concavo-convex inconsistently causes interference to spectral dispersion; Each sample dress two glasss, sample, each sample cup duplicate measurements 5 times, each sample obtains 10 curves of spectrum altogether, and is averaging with these 10 curves of spectrum, the corresponding averaged spectrum curve of last each sample.
The beneficial effect that the present invention obtains is as follows:
1. detection speed is fast: utilize near-infrared spectrometers to analyze a sample used time only about 0.5min, and Comparatively speaking, traditional detection method required time is tens times of this technology required time even hundreds of times.Therefore NIR technology is specially adapted to quick detection to BPO content in flour, monitoring.
2. environmental protection, easy and simple to handle: the energy Ratios visible ray of near infrared photon is also low, can not damage human body, belongs to green analytical technology.Except to except sample comminution, without any need for other pre-treatment, the essence of sample is not destroyed, and can recycle.Because sample need not process, so network or optical fiber can be utilized to carry out remote sample determination, also can the sample under rugged surroundings be monitored.And in traditional detection method, complicated chemical treatment is all passed through in the pre-treatment of sample, employ many chemical reagent, create a large amount of waste liquids, waste gas.
3. testing cost is cheap: due to the pre-treatment lacking a lot of complexity, can save the use of a large amount of chemical reagent and the input of professional.
4. testing process is easy: whole testing process only needs a near-infrared spectrometers, and carry out spectral scan to sample, process is easy.And the process of classic method phase before detection, a large amount of analysers related in testing process, all result in classic method too loaded down with trivial details.
Accompanying drawing explanation
Fig. 1 is lever value and the student residual error T distribution plan of sample;
(wherein horizontal ordinate represents the lever value of sample, and ordinate represents student's residual values of sample, each some expression sample in figure).
Fig. 2 is predicting the outcome of the RBF neural model set up of embodiment 1 and the fit correlation figure between actual value;
(wherein each round dot represents each sample in forecast set, and solid line represents matched curve, and dotted line and A=T represent diagonal line).
Fig. 3 is predicting the outcome of the BP neural network model set up of embodiment 2 and the fit correlation figure between actual value.
Fig. 4 is predicting the outcome of the PLS model set up of embodiment 3 and the fit correlation figure between actual value.
Embodiment
Below in conjunction with specific embodiment, the present invention will be further described, but the present invention is not by the restriction of embodiment.
Following examples agents useful for same, material, method and instrument, without special instruction, be this area conventional reagent, material, method and instrument, all obtain by commercial channel.
Embodiment 1
1, extract the flour on market, collect 100, sample altogether.
With near-infrared spectrometers, spectral scan is carried out to flour sample.Each sample dress two glasss, sample, each sample cup duplicate measurements 5 times, each sample obtains 10 curves of spectrum, and is averaging with these 10 curves of spectrum, the corresponding averaged spectrum curve of last each sample.Therefore, 100 curves of spectrum are finally obtained altogether.
2, taking flour sample 5g respectively inserts in 50mL color-comparison tube, drips methyl alcohol 10.0mL, leaves standstill 5min, drip 50% potassium iodide aqueous solution 5.0mL after fully stirring 1min, leaves standstill 10min, finally adds water to 50.0mL, stir and leave standstill after fully stirring 1min; Utilize 0.22 μm of membrane filtration supernatant liquid; Get in supernatant liquor injection liquid chromatography, instrument parameter is set as: chromatographic column is set to 4.6mm*250mm, and determined wavelength is 230nm, and in mobile phase, the ratio of methyl alcohol and water is 10:90 (volume fraction), flow velocity is set to 1.0mL/min, and sample size is 10.0 μ L.With liquid chromatograph, each flour sample is measured, obtain BPO content in flour respectively.
3, calculate lever value and student's residual values of each sample, draw a lever value and student residual error T distribution plan, as shown in Figure 1.In figure: analyze horizontal ordinate, sample is more turned right, and shows that the lever value of sample is larger, and namely this sample is likely abnormal sample; Analyze ordinate, offset from zero coordinate axis got over by sample, shows that student's residual error of sample is larger, and this sample may be abnormal sample.Checked by T that to reject 7 samples be altogether abnormal sample 7 data points such as () S-001 marked with square frame in Fig. 1, utilize absorption method one by one to judge whether this sample is really as abnormal sample simultaneously, these 7 samples of rejecting are recovered in calibration set one by one, set up RBF model, whether the performance of analytical model has impact.If model performance reduces, this sample is abnormal sample really, needs to reject; If model performance increases on the contrary, showing that this sample has certain contribution to model, is not abnormal sample, need again be recovered in calibration set.Finally, 4 abnormal samples are rejected altogether.
4, utilize SPXY algorithm that remaining 96 samples are divided into two set, i.e. calibration set (60) and forecast set (36).Calibration set is used for the foundation of forecast model, and forecast set is used for testing to institute's established model.
5, the spectroscopic data of standard normal variable transfer pair all samples is utilized to be optimized, to eliminate the interference caused spectral dispersion by Flour particle size difference.Utilize principal component analysis (PCA) (PCA) to carry out compression dimensionality reduction to spectroscopic data simultaneously, accelerate predetermined speed of model.
6, utilize calibration set sample to set up RBF neural model, wherein the expansion rate of RBF neural is set to 1000, utilizes " newrbe " instruction in MATLAB to carry out model emulation; Forecast set sample is utilized to carry out the performance of checking R BF neural network model, with the evaluation index that prediction related coefficient, predicted root mean square error and relation analysis error are model, the parameter of adjustment model, the matched curve finally predicted the outcome is as shown in Figure 2, wherein T is actual value, and A is predicted value.Prediction related coefficient (R) reaches 0.9937, and matched curve and diagonal line are substantially identical, and each data point is comparatively evenly distributed near matched curve.The predicted root mean square error (RMSEP) of computation model is low to moderate 15.5095 simultaneously, and relation analysis error (PRD) is up to 8.8216.Concrete model is: utilize newrbe (P, T, SPREAD) function creation accurate radial basis function network, comprising input layer, hidden layer and output layer, due to the network just about benzoyl peroxide, only has benzoyl peroxide variable layer so export.Pretreated spectrum data matrix X is as input vector P, the Determination of benzoyl peroxide in wheat value vector Y of the sample corresponding to these spectrum is as the desired output matrix T of model, square error is defaulted as 0, through repeatedly testing, when SPREAD value gets 1000, the effectiveness comparison of model is good, and hidden layer basis function adopts Gaussian function, a great advantage of RBF is that weights W is tried to achieve by linear least square method, and this makes RBF correcting algorithm have processing speed faster.
Embodiment 2
1, step 1-5 is identical with the step 1-5 in embodiment 1.
2, utilize calibration set sample to set up BP neural network model, wherein the Learning Step of BP neural network is set to 10000, learning objective is set to 0.1, learning rate is set to 0.001, utilizes " newff " and " train " instruction in MATLAB to carry out model emulation; Utilize forecast set sample to check the performance of BP neural network model, the matched curve predicted the outcome as shown in Figure 3, prediction related coefficient (R) only has 0.9786, and matched curve and diagonal line are misfitted, present certain angle, each data point skewness near matched curve, distribution is dispersed.The predicted root mean square error (RMSEP) of computation model is up to 41.1540 simultaneously, and relation analysis error (PRD) only has 2.8407.Concrete model is: utilize newff function and train function to emulate, wherein network structure comprises input layer, hidden layer and output layer, due to the network just about benzoyl peroxide, only has benzoyl peroxide variable layer so export.Pretreated spectrum data matrix X is as input vector, and the Determination of benzoyl peroxide in wheat value of the sample corresponding to these spectrum vector Y is as the desired output of model.Input layer and hidden layer all have employed " logsig " transport function, and what output layer adopted is " purelin " transport function, and training function selects " traingdx " function, and Learning Step is set to 10000, learning objective is set to 0.1, learning rate is set to 0.001.
Embodiment 3
1, step 1-5 is identical with the step 1-5 in embodiment 1.
2, utilize calibration set sample to set up offset minimum binary (PLS) model, wherein the contribution rate of accumulative total of PLS model is set to 95%, utilizes MATLAB to carry out model emulation; Utilize forecast set sample to check the performance of PLS model, as shown in Figure 4, prediction related coefficient (R) reaches 0.989, and matched curve and diagonal line are tending towards identical in the matched curve predicted the outcome, the angle presented is less, and each data point distributes more even near matched curve.The predicted root mean square error (RMSEP) of computation model is 19.8895 simultaneously, and relation analysis error (PRD) is 6.4574.
Concrete model is: using pretreated spectrum data matrix X as independent variable, the Determination of benzoyl peroxide in wheat value vector Y of the sample corresponding to these spectrum is as dependent variable, observation station is set to 1050, carries out Principle component extraction between x and y respectively, finally utilizes residual sum of squares (RSS)
confirm major component number, wherein threshold value is set to 97.5%, is extracted 41 major components.
Embodiment 4
After establishing detection model, inventor again purchases and extracts 10 kinds to market, often kind of 5 samples, amount to 50 samples, by the Determination of benzoyl peroxide in wheat in high effective liquid chromatography for measuring sample, find that there is 2 kinds totally 10 samples wherein containing benzoyl peroxide, these two kinds of samples are extracted out as laboratory sample, simultaneously from remaining 8 kinds of samples, randomly draw two kinds, a kind of directly as laboratory sample, another kind adds benzoyl peroxide wherein, and to be configured to concentration be 50mgKg
-1, utilize the real content of benzoyl peroxide in high effective liquid chromatography for measuring this kind of sample simultaneously.Finally obtain 4 kinds altogether, often kind of 5 samples, amount to 20 samples, utilize the detection model constructed by embodiment 1-3 to carry out the mensuration of BPO content.The process such as infrared diaphanoscopy and high-performance liquid chromatogram determination is all identical with method listed by embodiment 1, and final detection result is as shown in table 1.As can be seen from the table, the estimated performance of the RBF neural model that embodiment 1 is set up is best, and error is minimum.
The testing result of the different detection model of table 1
Unit: mg/kg
Although the present invention with preferred embodiment openly as above; but it is also not used to limit the present invention, any person skilled in the art, without departing from the spirit and scope of the present invention; can do various change and modification, what therefore protection scope of the present invention should define with claims is as the criterion.
Claims (10)
1. the method for Determination of benzoyl peroxide in wheat in a Fast nondestructive evaluation flour, it is characterized in that, after setting up detection model, utilize near infrared spectrometer to scan flour sample to be measured, obtain sample near infrared light spectral curve, the gained curve of spectrum is processed, finally utilizes detection model to detect the content of benzoyl peroxide in flour sample to be measured.
2. method described in claim 1, is characterized in that, step is as follows:
1) utilize near infrared spectrometer to scan flour sample, obtain sample near infrared light spectral curve;
2) utilize the Determination of benzoyl peroxide in wheat in high performance liquid chromatograph mensuration flour sample, set up Determination of benzoyl peroxide in wheat database;
3) to step 1) the near infrared light spectral curve of gained processes, obtain preprocessed data, using gained preprocessed data as X-axis data, simultaneously using step 2) data, as Y-axis data, and utilize SPXY algorithm that sample is divided into calibration set and forecast set in the benzoyl peroxide database of gained;
4) step 3 is utilized) gained calibration set data integrating step 2) the Determination of benzoyl peroxide in wheat database set up, set up Preliminary detection model, recycling forecast set data are verified Preliminary detection model and correct, and obtain detection model;
5) step 4 is utilized) detection model of gained measures Determination of benzoyl peroxide in wheat in flour to be measured.
3. method described in claim 2, is characterized in that, step 1) described utilize near infrared spectrometer scan flour sample, be strike off after sample is loaded sample cup, scanning samples 5 times, sweep limit is 400-2498nm, and resolution is 2nm.
4. method described in claim 2, it is characterized in that, step 2) utilize high performance liquid chromatograph to measure Determination of benzoyl peroxide in wheat in flour sample, mensuration process is, take 5g flour sample, be placed in 50mL color-comparison tube, after dripping 10.0mL methyl alcohol, 5min is left standstill after stirring 1min, at the potassium iodide aqueous solution 5.0mL of dropping 50%, 10min is left standstill after abundant stirring 1min, finally add water to 50.0mL, stirring utilizes the membrane filtration supernatant liquid of 0.22 μm after leaving standstill, get supernatant liquor after filtration to join in high performance liquid chromatograph and measure, during mensuration, mobile phase used is volume ratio is methyl alcohol: the mixed liquor of water=10:90, flow velocity is 1.0mL/min, sample size is 10.0 μ L, determined wavelength is 230nm.
5. method described in claim 2, is characterized in that, step 3) described gained near infrared light spectral curve to be processed, be utilize lever value and student residual error T to check rejecting abnormalities value, spectroscopic data is optimized in the conversion of recycling standard normal variable.
6. method described in claim 5, is characterized in that, the described standard normal variable that utilizes converts optimization spectroscopic data, utilizes principal component analysis (PCA) to carry out dimension-reduction treatment to spectroscopic data simultaneously.
7. method described in claim 2, it is characterized in that, step 4) detection model, it is RBF neural model, the structure of RBF neural model is the curve of spectrum data to obtain after pre-service is input vector, and with the corresponding Determination of benzoyl peroxide in wheat of high performance liquid chromatography actual measurement for desired output, square error is 0, SPREAD value is 1000, builds using Gaussian function as hidden layer basis function.
8. method described in claim 2, it is characterized in that, step 5) the described detection model that utilizes measures Determination of benzoyl peroxide in wheat in flour to be measured, after utilizing near infrared spectrometer to scan testing sample, obtain preprocessed data after the gained curve of spectrum is processed, then utilize the Determination of benzoyl peroxide in wheat in detection model acquisition sample according to preprocessed data.
9. method described in claim 2, is characterized in that, concrete steps are as follows:
1) strike off after flour sample being loaded sample cup, utilize near infrared spectrometer scanning samples 5 times, sweep limit is 400-2498nm, and resolution is 2nm, obtains sample near infrared light spectral curve;
2) 5g flour sample is taken, be placed in 50mL color-comparison tube, after dripping 10.0mL methyl alcohol, 5min is left standstill after stirring 1min, at the potassium iodide aqueous solution 5.0mL of dropping 50%, 10min is left standstill after abundant stirring 1min, finally add water to 50.0mL, stirring utilizes the membrane filtration supernatant liquid of 0.22 μm after leaving standstill, get supernatant liquor after filtration to join in high performance liquid chromatograph and measure, during mensuration, mobile phase used is volume ratio is methyl alcohol: the mixed liquor of water=10:90, flow velocity is 1.0mL/min, sample size is 10.0 μ L, determined wavelength is 230nm, detect the Determination of benzoyl peroxide in wheat terminated in rear acquisition flour sample, and set up benzoyl peroxide database,
3) utilize lever value and student residual error T to check and reject step 1) exceptional value in gained near infrared light spectral curve, spectroscopic data is optimized in the conversion of recycling standard normal variable, utilize principal component analysis (PCA) to carry out dimension-reduction treatment to spectroscopic data simultaneously, obtain preprocessed data, simultaneously in conjunction with benzoyl peroxide database, utilize SPXY algorithm that sample is divided into calibration set and forecast set;
4) utilize step 3) the calibration set data of gained integrating step 2) in benzoyl peroxide database, MATLAB is utilized to set up preliminary RBF neural model, recycling step 3) in the data of forecast set, with the evaluation index that prediction related coefficient, predicted root mean square error and relation analysis error are model, preliminary RBF neural model is verified and corrected, finally obtains detection model; Described detection model is: RBF neural forecast model, model is the curve of spectrum data to obtain after pre-service is input vector, with the corresponding Determination of benzoyl peroxide in wheat of high performance liquid chromatography actual measurement for desired output, square error is 0, SPREAD value is 1000, builds using Gaussian function as hidden layer basis function;
5) utilize near infrared spectrometer to scan flour sample to be measured, according to step 2) described in method the curve of spectrum is processed, then utilize step 4 according to gained preprocessed data) detection model obtains Determination of benzoyl peroxide in wheat in testing sample.
10. either method described in claim 1-9, is characterized in that, for detecting the content of the benzoyl peroxide in flour.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510477873.6A CN105044023A (en) | 2015-08-06 | 2015-08-06 | Method for quickly detecting content of benzoyl peroxide in flour in nondestructive mode and application thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510477873.6A CN105044023A (en) | 2015-08-06 | 2015-08-06 | Method for quickly detecting content of benzoyl peroxide in flour in nondestructive mode and application thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105044023A true CN105044023A (en) | 2015-11-11 |
Family
ID=54450741
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510477873.6A Pending CN105044023A (en) | 2015-08-06 | 2015-08-06 | Method for quickly detecting content of benzoyl peroxide in flour in nondestructive mode and application thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105044023A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107356555A (en) * | 2017-06-29 | 2017-11-17 | 深圳市赛亿科技开发有限公司 | A kind of device based near infrared spectrum detection flower characters |
| CN108872135A (en) * | 2018-08-31 | 2018-11-23 | 天津科技大学 | A method of utilizing polymalic acid content near infrared ray fermentation liquid |
| CN109342352A (en) * | 2018-12-18 | 2019-02-15 | 龙口味美思环保科技有限公司 | A kind of flour quality detection method based on hybrid analog-digital simulation annealing and genetic algorithm |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0061437A1 (en) * | 1981-03-16 | 1982-09-29 | Peter Perten | Infrared analyzer, especially for edibles such as flour |
| US20040082070A1 (en) * | 2001-04-11 | 2004-04-29 | Jones Howland D. T. | Optically similar reference samples |
| CN101210907A (en) * | 2006-12-28 | 2008-07-02 | 天津市食品研究所有限公司 | Wheat flour benzoyl peroxide highly effective liquid phase chromatography detection method |
| CN102353667A (en) * | 2011-07-05 | 2012-02-15 | 邢台市产品质量监督检验所 | Method for detecting content of benzoyl peroxide in flour |
| CN104020135A (en) * | 2014-06-18 | 2014-09-03 | 中国科学院重庆绿色智能技术研究院 | Calibration model establishing method based on near infrared spectrum |
-
2015
- 2015-08-06 CN CN201510477873.6A patent/CN105044023A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0061437A1 (en) * | 1981-03-16 | 1982-09-29 | Peter Perten | Infrared analyzer, especially for edibles such as flour |
| US20040082070A1 (en) * | 2001-04-11 | 2004-04-29 | Jones Howland D. T. | Optically similar reference samples |
| CN101210907A (en) * | 2006-12-28 | 2008-07-02 | 天津市食品研究所有限公司 | Wheat flour benzoyl peroxide highly effective liquid phase chromatography detection method |
| CN102353667A (en) * | 2011-07-05 | 2012-02-15 | 邢台市产品质量监督检验所 | Method for detecting content of benzoyl peroxide in flour |
| CN104020135A (en) * | 2014-06-18 | 2014-09-03 | 中国科学院重庆绿色智能技术研究院 | Calibration model establishing method based on near infrared spectrum |
Non-Patent Citations (3)
| Title |
|---|
| LAIJUN SUN等: "prediction of benzoyl peroxide in flour using near infrared spectroscopy technique", 《INTERNATIONAL JOURNAL OF FOOD PROPERTIES》 * |
| 刘贤赵等主编: "《地理学数学方法》", 31 August 2009, 北京：科学出版社 * |
| 张志勇: "基于近红外光谱法检测面粉中过氧化苯甲酰添加量", 《中国博士学位论文全文数据库 工程科技Ⅰ辑》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107356555A (en) * | 2017-06-29 | 2017-11-17 | 深圳市赛亿科技开发有限公司 | A kind of device based near infrared spectrum detection flower characters |
| CN108872135A (en) * | 2018-08-31 | 2018-11-23 | 天津科技大学 | A method of utilizing polymalic acid content near infrared ray fermentation liquid |
| CN109342352A (en) * | 2018-12-18 | 2019-02-15 | 龙口味美思环保科技有限公司 | A kind of flour quality detection method based on hybrid analog-digital simulation annealing and genetic algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Bázár et al. | NIR detection of honey adulteration reveals differences in water spectral pattern | |
| He et al. | Discrimination of varieties of tea using near infrared spectroscopy by principal component analysis and BP model | |
| Zhang et al. | Nondestructive measurement of soluble solids content in apple using near infrared hyperspectral imaging coupled with wavelength selection algorithm | |
| Li et al. | Simultaneous determination of six main types of lipid-soluble pigments in green tea by visible and near-infrared spectroscopy | |
| Wang et al. | Green analytical assay for the quality assessment of tea by using pocket-sized NIR spectrometer | |
| Zhu et al. | Evaluation of the non-enzymatic browning in thermally processed apple juice by front-face fluorescence spectroscopy | |
| de Oliveira et al. | An environment-friendly spot test method with digital imaging for the micro-titration of citric fruits | |
| Chen et al. | Rapid measurement of total acid content (TAC) in vinegar using near infrared spectroscopy based on efficient variables selection algorithm and nonlinear regression tools | |
| CN107064047A (en) | A kind of Fuji apple quality damage-free detection method based near infrared spectrum | |
| CN104020129A (en) | Method for discriminating fermentation quality of congou black tea based on near-infrared-spectroscopy-combined amino acid analysis technology | |
| US9830429B2 (en) | Systems and methods for multi-flux color matching | |
| Peng et al. | Monitoring of alcohol strength and titratable acidity of apple wine during fermentation using near-infrared spectroscopy | |
| CN105044023A (en) | Method for quickly detecting content of benzoyl peroxide in flour in nondestructive mode and application thereof | |
| CN103134767A (en) | Method for liquor quality identification through infrared spectrum revision | |
| Xu et al. | Research on nondestructive identification of grape varieties based on EEMD‐DWT and hyperspectral image | |
| CN106872370A (en) | Anthocyanin content method for measuring in a kind of red bayberry based on EO-1 hyperion | |
| JP2010210355A (en) | Method and apparatus for nondestructive measurement of component of vegetable etc. using near-infrared spectroscopy | |
| CN104764699A (en) | Method for determination of edible oil acid value | |
| Yun et al. | Identification of tea based on CARS‐SWR variable optimization of visible/near‐infrared spectrum | |
| da Silva Ferreira et al. | Determination of pitaya quality using portable NIR spectroscopy and innovative low-cost electronic nose | |
| de Matos et al. | Innovative methodological approach using CIELab and dye screening for chemometric classification and HPLC for the confirmation of dyes in cassava flour: A contribution to product quality control | |
| Chen et al. | Modeling of soluble solid content of PE‐packaged blueberries based on near‐infrared spectroscopy with back propagation neural network and partial least squares (BP–PLS) algorithm | |
| CN106226267A (en) | A kind of near-infrared assay method of Fructus Capsici dry color valency | |
| Davis et al. | Rapid method for total carotenoid detection in canary yellow‐fleshed watermelon | |
| CN106872396B (en) | A kind of method of difference nir instrument measurement grape pol model conversion |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20151111 |