CN105181643B - A kind of near infrared detection method of rice quality and application - Google Patents
A kind of near infrared detection method of rice quality and application Download PDFInfo
- Publication number
- CN105181643B CN105181643B CN201510657245.6A CN201510657245A CN105181643B CN 105181643 B CN105181643 B CN 105181643B CN 201510657245 A CN201510657245 A CN 201510657245A CN 105181643 B CN105181643 B CN 105181643B
- Authority
- CN
- China
- Prior art keywords
- rice
- designated
- sample
- content
- during
- 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.)
- Active
Links
Abstract
The invention belongs to agricultural product composition detection technical field, and in particular to a kind of near infrared detection method of rice quality and application.It is related to rice physico-chemical analysis technical field.It is a feature of the present invention that：Testing sample is subjected to infrared diaphanoscopy, obtain the original wave spectrogram of light absorption value based on wavelength, then original wave spectrogram is subjected to denoising Processing and pretreatment, characteristic wavelength is screened from the wave spectrum after processing, characteristic wavelength is substituted into the forecast model of sample quality index again, obtain the near infrared detection value of sample quality.Described rice includes paddy, brown rice and polished rice, and detection quality includes protein, fat, moisture, total reducing sugar, ash content, and the broken rice rate of rice, husking yield, chaff powder rate and head rice rate.Described detection apply including：(1) quality of model prediction rice is used；(2) amendment of model.The present invention has outstanding advantages of test speed is fast, to sample nondestructive.
Description
Technical field
The invention belongs to Analyzing The Quality of Agricultural Products technical field, and in particular to a kind of near infrared detection method of rice quality
With application.
Background technology
Rice is the main cereal crops in China, containing protein, fat, moisture, total reducing sugar, ash grade a variety of nutrition into
Point, the detection of these compositions has the shortcomings that complex operation, time-consuming and laborious.Near infrared light is wavelength between visible region with
Near infrared spectrum is defined as 780-2526nm area by the electromagnetic wave between infrared region, U.S. material detection association (ASTM)
Domain.Near infrared spectrum makes molecular vibration from ground state to being produced during high energy order transition mainly due to the anharmonicity of molecular vibration,
The sum of fundamental frequencies and frequency multiplication for recording hydric group (C-H, N-H, O-H) vibration absorb.Quality in rice contains abundant hydric group,
There is stronger response near infrared spectrum section.
Near-infrared spectrum analysis has been applied to the content of detection paddy amylose, protein in paddy, rice, fat,
The content of moisture and total reducing sugar, abroad, Delwiche S R【1】Deng the amylose that paddy is established using near-infrared spectrum technique
Calibration model, coefficient correlation reaches 0.95, achieves preferable effect；Himmerlsbach[2]Deng utilizing near infrared spectrum skill
Art establishes the calibration model of the protein of polished rice, and the coefficient correlation of measured value and predicted value is up to 0.992；At home, Lv Hui【3】Deng
Using near-infrared spectrum technique, the quantitative analysis that optimal Spectral range establishes rice moisture, protein and amylose is screened
Model, the coefficient of determination of forecast model are up to 0.992；Guo Yongmei【4】Deng near-infrared spectral analysis technology is based on, use is partially minimum
Square law, and screen the calibration model that optimal Spectral range establishes brown rice protein prediction, the coefficient of determination 0.9289.Application number
Numbers 2013105009402【5】The method that document discloses amylose content in a kind of quick detection rice, using
850nm-1045nm near infrared light spectrum information, near infrared correction, application number are established with PLS
201110187788.8【6】Document discloses a kind of method using near-infrared quick detection rice bran fats enzymatic activity, using
800nm-2500nm near infrared light spectrum information, model is established with principal component regression and PLS.
But the redundancy that these researchs are remained in spectrum is more, spectral signature wavelength is indefinite, forecast model report
It is few, while the problems such as broken rice rate to rice, husk rate, few near-infrared model report of chaff powder rate and head rice rate.
The content of the invention
The purpose of the present invention is the defects of being to overcome prior art, there is provided a kind of near infrared detection method of rice quality
With application.The present invention is quick, lossless and easy, to find the near infrared light spectrum signature of rice quality, by establishing forecast model
Method, provide a kind of method using near infrared detection rice quality for the detection of rice quality.
The present invention is achieved through the following technical solutions：
The present invention selects the characteristic wavelength of rice quality, and characteristic wavelength quantity is few, and rice product are established with characteristic wavelength
The method of the near-infrared model of matter, in the case of identical precision of prediction, predetermined speed of the invention is fast.The present invention establishes simultaneously
Rice broken rice rate, husk rate, chaff powder rate and head rice rate near-infrared forecast model method, to rice quality com-parison and analysis
Comprehensively.
Specifically, the near infrared detection method of rice quality provided by the invention, comprises the following steps：
A kind of near infrared detection method of rice quality and application, comprise the following steps：
Testing sample is subjected to infrared diaphanoscopy, the original wave spectrogram of light absorption value based on wavelength is obtained, then by original ripple
Spectrogram carries out denoising Processing and pretreatment, characteristic wavelength is screened from the wave spectrum after processing, then characteristic wavelength is substituted into sample product
In the forecast model of matter index, the near infrared detection value of sample quality is obtained.
Wherein：The construction method of the forecast model is as described below：
(1) collect representational Rice Samples, including paddy, brown rice, polished rice, husk, rice bran or crack rice；
(2) test chemical is carried out to collected sample, the test chemical value of gained is designated as ymj, wherein：M is m-th
Index, m=1,2,3 ..., 20；It is designated as paddy moisture as m=1, is designated as rice glutelin matter content during m=2, during m=3
It is designated as paddy fat content, is designated as during m=4 as paddy total sugar content, be designated as paddy content of ashes during m=5, is designated as paddy during m=6
Shell rate, is designated as brown rice broken rice- containing rate during m=7, is designated as broken rice rate during m=8, is designated as chaff powder rate during m=9, is designated as head milled rice during m=10
Rate, is designated as brown rice moisture content during m=11, is designated as protein content of rice during m=12, is designated as Brown rice lipid content during m=13,
It is designated as brown rice total sugar content during m=14, is designated as brown rice content of ashes during m=15, is designated as polished rice moisture, m=17 during m=16
Shi Jiwei polished rice protein contents, are designated as polished rice fat content during m=18, are designated as polished rice total sugar content during m=19, during m=20
It is designated as polished rice content of ashes；J is j-th of sample, common n sample, n >=40；
(3) infrared diaphanoscopy, light absorption value x are carried out to sampleij, wherein i i-th of wavelength of expression, wavelength i=1000nm,
1001nm, 1002nm ..., 1799nm；
(4) denoising Processing and pretreatment：With wavelet Denoising Method by xijDenoising Processing is carried out, obtains de-noising light absorption value, then
De-noising light absorption value is located in advance with one or more combination treatment methods in normalization, first derivative, second dervative again
Reason, obtain pre-processing light absorption value Aij；
(5) characteristic wavelength of near-infrared spectrum is screened：With competitive adaptive weight weight sampling (CARS) method and partially most
Small square law (PLS) method screens characteristic wavelength of near-infrared spectrum, establishes the forecast model of rice quality：zm=bm+∑amiBi, its
Middle zmFor the near infrared detection value of rice quality, BiFor AijIn j-th of sample light absorption value, bm、amiFor regression coefficient, mould is predicted
The evaluation index of type is coefficient of determination R2With calibration standard deviation RMSEC, the conspicuousness of regression coefficient is examined with T, as coefficient ami's
When conspicuousness is that t values are more than 0.05, then a is mademi=0, amiWavelength i is characterized wavelength corresponding to ≠ 0 place；
Characteristic wavelength is as follows：
Paddy moisture：1310nm, 1402nm, 1593nm, 1738nm and 1772nm；
Rice glutelin matter：1206nm, 1254nm, 1274nm, 1563nm and 1752nm；
Paddy fat：1343nm, 1369nm, 1489nm, 1574nm and 1583nm；
Paddy total reducing sugar：1086nm, 1273nm, 1279nm, 1577nm and 1643nm；
Paddy ash content：1079nm, 1181nm, 1417nm, 1426nm and 1494nm；
Brown rice moisture：1026nm, 1102nm, 1213nm, 1313nm and 1746nm；
Brown rice protein：1168nm, 1170nm, 1250nm, 1780nm and 1779nm；
Brown rice fat：1625nm, 1536nm, 1712nm, 1026nm and 1042nm；
Brown rice total reducing sugar：1008nm, 1326nm, 1377nm, 1525nm and 1599nm；
Brown rice ash content：1073nm, 1068nm, 1141nm, 1259nm and 1785nm；
Polished rice moisture：1060nm, 1274nm, 1293nm, 1328nm and 1408nm；
Polished rice protein：1254nm, 1285nm, 1516nm, 1554nm and 1717nm；
Polished rice fat：1018nm, 1536nm, 1608nm, 1625nm and 1712nm；
Polished rice total reducing sugar：1304nm, 1338nm, 1617nm, 1726nm and 1745nm；
Polished rice ash content：1452nm, 1472nm, 1481nm, 1724nm and 1759nm；
Husk rate：1127nm, 1264nm, 1446nm, 1495nm and 1597nm；
Brown rice broken rice- containing rate：1123nm, 1301nm, 1317nm, 1326nm and 1681nm；
Broken rice rate：1183nm, 1243nm, 1579nm, 1584nm and 1723nm；
Chaff powder rate：1157nm, 1602nm, 1723nm, 1728nm and 1730nm；
Head rice rate：1114nm, 1151nm, 1257nm, 1659nm and 1680nm；
Features described above wavelength allows the deviation for having ± 2nm.
(6) amendment of forecast model：Increase by 1 sample, make sample number n=n+1, chemical survey is carried out to the sample newly increased
Examination, infrared diaphanoscopy, denoising Processing and pretreatment, obtain new Aij, as the near-infrared characteristic wavelength described in step (5), obtain
The near infrared detection value z of new rice qualitymForecast model.
The present invention can be used for the granular solids cereal such as rice, wheat, corn, Semen Coicis, soybean, sesame, peanut and rice bran,
The Quality Detection of the granular material sample such as wheat flour, the described index of quality are not limited to These parameters.
The invention has the advantages that：
1st, the characteristic wavelength of rice quality is determined.The analysis method of model is established with characteristic wavelength, improves model
Precision.
2nd, the quick detection of rice quality is realized.The detection of conventional rice quality, time-consuming, and near infrared detection is only
Need the 3-4 seconds；
3rd, the Non-Destructive Testing of rice quality is realized.The detection of conventional rice quality is needed to crush, and rice is caused to damage
Evil, and near infrared detection can detect to whole grain rice；
4th, rice quality is analyzed more comprehensive.Establish the protein of rice, fat, moisture, total reducing sugar, ash content, broken rice rate,
The forecast model and method of testing of husk rate, chaff powder rate and head rice rate.
Brief description of the drawings
Fig. 1：The spectrum of rice sample after being the undressed rice sample atlas of near infrared spectra of the present invention and pre-processing
Figure.Description of reference numerals：Fig. 1 (a) figures are undressed rice sample atlas of near infrared spectra；Fig. 1 b) figure is to disappear by small echo
Rice sample atlas of near infrared spectra after making an uproar；Fig. 1 (c) figures are the atlas of near infrared spectra after normalized.
Fig. 2：It is the characteristic wavelength screening figure of the present invention.Description of reference numerals：(a) figure in Fig. 2 is selected in screening process
Go out the variation tendency of variable, with the increase of number of run, the variable number of reservation is fewer and fewer, and exponentially function from fast to slow
Successively decrease；(b) figure in Fig. 2 is the changing trend diagram of cross validation mean square deviation RMSECV during wavelength Variable Selection；In Fig. 2
(c) point in figure corresponding to " * " is RMSECV minimum points；Each line represents to become as number of run increases each wavelength in Fig. 2 (c)
Measure the variation tendency of regression coefficient.
Fig. 3：It is undressed brown rice sample atlas of near infrared spectra and brown rice moisture characteristic wavelength the selection result.Accompanying drawing mark
Remember explanation：(a) figure in Fig. 3 is undressed brown rice sample atlas of near infrared spectra；(b) figure in Fig. 3 is in screening process
The variation tendency of variable is selected, with the increase of number of run, the variable number of reservation is fewer and fewer, and exponentially letter from fast to slow
Number successively decreases；(c) figure in Fig. 3 is the changing trend diagram of cross validation mean square deviation RMSECV during wavelength Variable Selection；In Fig. 3
(d) figure in point corresponding to " * " be RMSECV minimum points, each line is represented as number of run increases each wavelength in Fig. 3 (d)
The variation tendency of variable regression coefficient.
Fig. 4：It is undressed polished rice sample atlas of near infrared spectra and polished rice protein characteristic wavelength the selection result.Accompanying drawing
Description of symbols：(a) figure in Fig. 4 is undressed polished rice sample atlas of near infrared spectra；(b) figure in Fig. 4 is screening process
In select the variation tendency of variable, with the increase of number of run, the variable number of reservation is fewer and fewer, and from fast to slow exponentially
Function successively decreases；(c) figure in Fig. 4 is the changing trend diagram of cross validation mean square deviation RMSECV during wavelength Variable Selection；Fig. 4
In (d) figure in point corresponding to " * " be RMSECV minimum points, each line is represented as number of run increases each ripple in Fig. 4 (d)
The variation tendency of long variable regression coefficient.
Embodiment
Test material and method
The test raw material or material relevant information used in the embodiment of the present invention is shown in Table 1.
1 46 rice variety numberings of table and title
Near infrared spectra collection
Whole grain paddy, brown rice and polished rice sample are put into sample disc respectively, and filled, is compacted with sample disk cover, is used
SupNIR-2720 near infrared spectrometers carry out spectra collection, treat that instrument preheats 30min, spectrum is carried out after performance test and reference
Scanning.It is 15-25 DEG C to scan temperature.
Sweep parameter is as follows：Instrument bandwidth：1nm, sweep spacing：1nm, scanning times：3 times, metering system：Diffusing reflection,
Wave-length coverage：1000-1799nm, spectroscopic data points：800.
Sample quality detection method
(1) water content detection
According to GB 5009.3-2010《National food safety standard：The measure of moisture in food》In direct drying method come
The content of moisture in detection paddy, brown rice and polished rice.
(2) protein detection
According to GB 5009.5-2010《National food safety standard：The measure of Protein in Food》In Kjeldahl's method
To detect the content of protein in paddy, brown rice and polished rice.
(3) fat detection
According to GB/T 5511-2008《Crude fat content determines in grain and oil detection grain》In soxhlet extraction methods detection rice
Fatty content in paddy, brown rice and polished rice.
(4) total sugar detection
According to GB/T 5009.7-2008《The measure of reduced sugar in food》In direct titrimetric method detection paddy, brown rice and
The content of total reducing sugar in polished rice.
(5) ash content detects
According to GB 5009.4-2010《National food safety standard：The measure of ash content in food》To detect paddy, brown rice
With the content of ash content in polished rice.
(6) husk rate detects
Paddy quality to be processed is designated as m1, the husk obtained when paddy is processed with hulling machine is collected, and weigh, matter
Amount is designated as m2, unit g.Rice hulls rate (husk content) specific formula for calculation is：
(7) broken rice rate detects
Collect the brown rice obtained when paddy removes husk with hulling machine and broken brown rice, quality are designated as m respectively3And m4, unit
For g.Brown rice broken rice- containing rate (broken brown rice yield) specific formula for calculation is：
Broken rice rate (broken rice yield) is according to GB/T 5503-2009《Inspection of grain and oil is cracked rice method of inspection》In meter
Calculate formula to calculate, specific formula for calculation is：
In formula, ycFor broken rice rate, m5For quality of white rice, m6For quality of being cracked rice in polished rice, unit g.
(8) chaff powder rate detects
Brown rice quality to be processed is designated as m7, the chaff powder obtained when brown rice machines with husk rice is collected, and weigh, matter
Amount is designated as m8, unit g.Chaff powder rate (rice bran powder yield) specific formula for calculation is：
(9) head rice rate detects
Whole White Rice Percentage Dynamic (head rice yield) is according to GB/T 21719-2008《Whole White Rice Percentage Dynamic method of inspection》
In calculation formula calculate, specific formula for calculation is：
In formula, yeFor head rice rate, m9For paddy quality, m8For head milled rice quality, unit is gram g.
(10) biased sample chemical composition calculates
The test value of the main nutrient composition of 46 paddy, brown rice and polished rice kind determines one by one according to state's calibration method
Afterwards, the calculation formula of the main nutrient composition of biased sample is:
yj=(Aj+Bj)/2
Y in formulajFor the calculated value of the main nutrient composition of j-th of biased sample, AjAnd BjRespectively form biased sample
Chemical score measured by two kind national standard methods.
Model-evaluation index
With coefficient of determination R2, calibration standard deviation RMSEC, verification standard deviation RMSEP carry out the calibration effect of evaluation model and pre-
Survey ability.
In formula, n, m are respectively that calibration set and checking collect sample number, yiFor the master detected with National Standard Method of i-th of sample
Want nutritional ingredient value, paddy i=90, brown rice i=66, polished rice i=66,For the prediction of the main nutrient composition of i-th of sample
Value,For the average value of sample main nutrient composition.R2Closer to 1, illustrate that regression effect is notable, RMSEC and RMSEP more connect
0 is bordering on, illustrates that model has good stability and predictive ability.
Embodiment 1：The foundation of the near-infrared model of rice glutelin matter
(1) typical Rice Samples are gathered extensively：Rice Samples are provided by Hubei Province Huanggang City farm institute, 46 sample product
Kind is shown in Table 1.By sample segment two-by-two according to 1:1、1:2 ratio is well mixed, and obtains 44 new samples, totally 90 samples,
That is n=90.Using national standard method【8】The content of protein is determined, the fundamental statistics of rice sample protein content are shown in Table 2.
The fundamental statistics (%) of the rice glutelin matter content of table 2
(2) spectra collection being carried out with SupNIR-2720 near infrared spectrometers, the measurement range of spectrum is 1000-1799nm,
Spectrum step-length is 1nm, and to reduce error, each Sample Scan 3 times, before scanning, instrument preheating 30min, scan data is with extinction
Degree form stores, and scanning temperature is 25 DEG C of room temperature.Paddy seed is placed in sample disc, and is filled, is compacted, and each sample is swept one by one
Retouch, original spectrum is shown in accompanying drawing 1 (a).
(3) near infrared spectrum pre-processes：Paddy near infrared spectrum first uses wavelet noise【13】Processing, using " db2 " small echo,
Maximum decomposition scale is 3 progress de-noisings.Wavelet noise influences to see embodiment 4, and the spectrum after wavelet noise is shown in accompanying drawing 1 (b).Use again
Normalized【14】, variance 1, average 0.Normalization creep function precision influences to see embodiment 5, obtained de-noising and pretreatment
Spectrum light absorption value Aij afterwards is shown in accompanying drawing 1 (c), from accompanying drawing 1 (c), before the curve of spectrum after wavelet noise is compared with de-noising
Curve smoothing.
(4) characteristic wavelength of near-infrared spectrum is screened：The adaptive weight weight sampling method (CARS) of competition【7】In variables choice
During, the wavelength that regression coefficient absolute value is big in PLS model is retained by adaptively sampled technology every time and become
Amount, the small regression coefficient of absolute value is defaulted as 0, Monte Carlo sampling number is 200 times, sees Fig. 2, picks 31 characteristic waves
Long, significance analysis the results are shown in Table 3.
3 rice glutelin matter of table, 31 characteristic wavelengths and its conspicuousness
(5) multiple linear regression analysis method optimization characteristic wavelength.Independent variable, the egg of paddy are used as by the use of the characteristic wavelength selected
White matter content carries out multiple linear regression analysis as dependent variable with PLS[19], with the 31 of step (4) screening
Individual characteristic wavelength establishes the forecast model of rice glutelin matter, makes the not high (t of conspicuousness>0.05) system is returned corresponding to characteristic wavelength
Number is 0, and preferred feature wavelength finally gives 24 characteristic wavelengths, and regression coefficient and conspicuousness are shown in Table 4.
4 rice glutelin matter of table, 24 characteristic wavelengths and its conspicuousness
The data of table 4 are substituted into following formula and obtain the forecast model of rice glutelin matter content：
z2=b2+∑a2iBi (1)
In formula, z2For rice glutelin matter content, b2For regression constant item, now b2=67.310, a2iFor each characteristic wavelength
Regression coefficient, BiThe light absorption value for being characterized wavelength is characterized wavelength by de-noising and pretreated numerical value, i.
(6) near infrared detection of protein content
No. 3 sample protein matter contents are detected with formula (1), sample protein matter chemical detection value is 8.75%, near red
Outer detected value z2For 8.3700%, the coefficient of determination R of model2For 0.8916, standard deviation 0.2149, illustrate institute's established model prediction
Precision is higher.
Embodiment 2：Paddy total reducing sugar, fat, ash content, brown rice moisture and polished rice protein near-infrared model foundation
(1) remove husk with hulling machine and obtain the brown rice of 46 kinds, brown rice is removed into chaff powder with rice mill, obtains 46
The polished rice of kind.Part brown rice sample presses 1:1 ratio mixes two-by-two, obtains 20 biased samples, altogether 66 brown rice samples,
Part polished rice sample presses 1:1 ratio mixes two-by-two, obtains 20 biased samples, altogether 66 polished rice samples.
(2) with state's calibration method measure paddy total reducing sugar【11】, fat【9】, ash content【12】, brown rice moisture【10】With polished rice albumen
Matter【8】.Wherein brown rice moisture and polished rice protein test result fundamental statistics is shown in 5 tables, table 6.
The fundamental statistics (%) of the brown rice moisture content of table 5
The fundamental statistics (%) of the polished rice protein content of table 6
(3) near infrared spectrum scanning is carried out to sample according to the method for the step of embodiment 1 (2), obtains original spectrum；
(4) near infrared spectrum de-noising and pretreatment：The de-noising of paddy near infrared spectrum and pretreatment with embodiment 1, difference
It is in brown rice near infrared spectrum first uses wavelet noise, then uses normalized；Polished rice near infrared spectrum, first with smooth de-noising,
Normalized is used again.【18】
(5) characteristic wavelength of 5 indexs of the present embodiment is determined respectively using the method for the step of embodiment 1 (4)-(5), takes
The forecast model that 5 characteristic wavelengths of conspicuousness highest are established is shown in formula (2)-(6).
Paddy fat：
z3=3.65+7321.89B1343-5457.37B1369-8170.05B1489-5348.06B1574
-5875.20B1583 (2)
Paddy total reducing sugar：
z4=-5598.11+1674.04B1086+1661.46B1273-1795.13B1279-2264.71B1577
+2155.48B1643 (3)
Paddy ash content：
z5=-58.78+146.48B1079+145.52B1181+147.00B1417-170.55B1426
-137.90B1494 (4)
Brown rice moisture：
z11=-1860.68-235.90B1026+210.77B1102-262.47B1213-296.55B1313
-210.27B1746 (5)
Polished rice protein：
z17=891.84+265.19B1254–221.13B1285+220.95B1516–224.46B1554
-281.71B1717 (6)
(6) near infrared detection of index
Characteristic wavelength is substituted into corresponding forecast model, calculates the near infrared detection value z of each indexm, result of calculation
And standard deviation is shown in Table 7.As shown in Table 7, the higher (coefficient of determination R of model of the precision of near infrared detection20.74 is all higher than, standard
1.89) difference is respectively less than, the deviation of conventional method measured value and near-infrared measuring value is small, has preferable practical value.
The model calculation of table 7 and precision of prediction
Embodiment 3：Broken rice rate, husk rate, the head rice rate of paddy, the foundation of the near-infrared model of the chaff powder rate of brown rice and
Detection
(1) spectral scan of paddy, brown rice and polished rice, spectrum de-noising, pretreatment, method is the same as embodiment 1.
(2) screening of characteristic wavelength, method is the same as embodiment 1.
(3) broken rice rate of paddy, husk rate, head rice rate, the foundation of the near-infrared model of the chaff powder rate of brown rice, method are same
Embodiment 1.
Wherein, the forecast model for building paddy husk rate is as follows：
Husk rate：
z6=242.98+237.70B1127-334.69B1264-229.12B1446+254.35B1495
-251.58B1597
Broken rice rate：
z8=4405.08+5499.36B1183-7169.48B1243-4562.46B1579+5392B1584
-5103.13B1723
Chaff powder rate：
z9=-569.08-483.37B115-539.07B1602-406.66B1723+413.39B1728
-333.22B1730
Head rice rate：
z10=-10065-2434.67B1114+2215.91B1151-2450.11B1257+2500.71B1659
-2382.3B1680
(4) near infrared detection of index
Characteristic wavelength is substituted into corresponding forecast model, calculates the near infrared detection value z of each indexm, result of calculation
And standard deviation is shown in Table 8.As shown in Table 8, the higher (coefficient of determination R of model of precision using above-mentioned model prediction2It is all higher than 0.8,
3.2) standard deviation is respectively less than, the deviation of conventional method measured value and near-infrared measuring value is small, has preferable practical value.
The model calculation of table 8 and precision of prediction
*：The specific steps of conventional method detection are shown in " sample detection methods (6)-(9) "
Embodiment 4：Influence of the de-noising to model accuracy
By taking paddy fat index as an example, noise-eliminating method is described further (other indexs disappear the same the present embodiment of drying method,
The present invention is not exhaustive one by one, and those skilled in the art can repeat the effect of the present invention using the method for the present invention).
The near infrared spectrum scanning of paddy, method is with embodiment 1, respectively with small echo and SG smoothly to paddy near infrared light
Spectrum carries out de-noising, obtains the light absorption value after de-noising, substitutes into formula (2) and near-infrared predicted value, near-infrared predicted value is calculated
9 are the results are shown in Table with measured value error analysis.Table 9 understands, the coefficient of determination R of model during de-noising smooth with SG2(0.9647) highest,
Standard deviation (0.2182) is minimum, is convenient near infrared spectrum noise-eliminating method.
Influence of the different noise-eliminating method of table 9 to model accuracy
Embodiment 5：Pre-process the influence to model accuracy
By taking rice glutelin matter index as an example, it is described further that (preprocess method of other indexs is same to preprocess method
The present embodiment, the present invention is not exhaustive one by one, and those skilled in the art can repeat the effect of the present invention using the method for the present invention
Fruit).
The near infrared spectrum scanning of paddy, method carry out de-noising to spectrum with small echo, use return respectively with embodiment 1
One changes【14】, first derivative【15】, second dervative【16】The pretreatment after wavelet noise is carried out to paddy near infrared spectrum, obtains pre- place
Light absorption value after reason, substitute into formula (1) and near-infrared predicted value is calculated, near-infrared predicted value and measured value error analysis result
It is shown in Table 10.
Influence of the different preprocess method of table 10 to model accuracy
As shown in Table 10, the coefficient of determination R of the forecast model of normalized2(0.9480) highest, standard deviation
(0.1470) it is minimum, it is that the near infrared spectrum of convenient rice glutelin matter carries out preprocess method.
Embodiment 6：Influence of the wavelength to model accuracy
By taking rice glutelin matter index as an example, characteristic wavelength is selected to CARS methods and is described further (the spy of other indexs
Same the present embodiment of detection method of influence of the wavelength to model accuracy is levied, the present invention is not exhaustive one by one, those skilled in the art
The effect of the present invention can be repeated using the method for the present invention).
(1) the near infrared spectra collection method of paddy carries out de-noising with embodiment 1 with db2 small echos to spectrum, with returning
One changes【18】The pretreatment after wavelet noise is carried out to paddy near infrared spectrum.
(2) screening of characteristic wavelength, with Monte Carlo sampling number during CARS Variable Selections be respectively 0 time (do not screen,
Using all-wave length, totally 800), 100 times, 150 times, 200 times, the spy that other method determines with embodiment 1, different sampling numbers
Sign number of wavelengths is shown in Table 11.
(3) forecast model of the rice glutelin matter content of characteristic wavelength quantity is established using the method for embodiment 1, is calculated
The near infrared detection value of protein content.Influence of the characteristic wavelength to model accuracy is shown in Table 11.As shown in Table 11, sampling number and
Characteristic wavelength quantity has an impact to model accuracy, and the forecast model that 24 characteristic wavelengths obtained with 200 samplings are established is determined
Determine coefficient (0.9480) maximum, standard deviation (0.1470) is minimum, precision highest.
Influence of the characteristic wavelength of table 11 to model accuracy
Embodiment 7：The amendment of model
By taking rice glutelin matter as an example, the amendment to model is described in further detail the (inspection of the amendment of other index models
The same the present embodiment of survey method, the present invention is not exhaustive one by one, and those skilled in the art can repeat this using the method for the present invention
The effect of invention).
(1) increase a new paddy sample, variety name be osmanthus towards 13, then n=91.
(2) method for using embodiment 1, test chemical, infrared diaphanoscopy, denoising Processing and pre- place are carried out to the sample
Reason, near-infrared characteristic wavelength is screened, obtains the near infrared detection value z of new rice qualitymForecast model.Rice glutelin matter
Forecast model (only listing the forecast model that the higher characteristic wavelength of 5 conspicuousnesses is formed) is as follows：
z2=145.69-204.71B1206+374.48B1254-213.53B1274+224.41B1563
-2.6.69B1752
The coefficient of determination R of model2For 0.9651, standard deviation RMSEC is 0.1284, than embodiment 1 model accuracy
Improve.No. 3 rice samples are detected with the model, protein content (z2) it is 8.42%, before relatively correcting (8.37%)
Precision increases.
Embodiment 8：The near infrared detection of rice quality
The index (index that embodiment 1-5 is not detected) of 90 Rice Samples to embodiment 1 is detected respectively, tied
Fruit is shown in Table 12.
(1) near infrared spectrum scanning, de-noising and pretreatment, method is the same as example 1.Brown rice and polished rice sample infrared diaphanoscopy are former
Beginning spectrum is shown in Fig. 3 (a) and Fig. 4 (a) respectively.
(2) characteristic wavelength is screened, method is the same as embodiment 1.Brown rice moisture, polished rice protein characteristic wavelength the selection result are shown in point
Fig. 3 (b)-(d) and Fig. 4 (a) (b)-(d) are not seen.
(3) the near-infrared inspection of the forecast model calculating rice quality of rice quality index is established using the method for embodiment 1
Measured value, it the results are shown in Table 12.
As shown in Table 12, the coefficient of determination R of forecast model20.7 is all higher than, standard deviation is respectively less than 1.8, illustrates model prediction
Precision is high, and the deviation of conventional method measured value and near-infrared measuring value is small, has preferable practical value.
The rice quality near infrared detection result of table 12
Bibliography
[1]Delwiche S R,Bean M M.Apparent amylose content of milledrice by
near-infrared reflectance spectroscopytry[J].Cereal Chemistry,1995,72(2):182-
187；
[2]Himmelsbach H S,Barton F E,McClung A M,et al.Protein and apparent
amylase content of milled rice by NIR-FT/Raman spectroscopy.Cereal Chem,2011,
78(4):488-492；
[3] Lv Hui etc., the rice quality analysis based on near-infrared spectrum technique differentiate [J] food industry science and technology with species,
2012,03:322-325；
[4] correlation analysis of Guo Yongmei etc., protein content of rice and Mineral Elements Content and the foundation of NIRS models
[J] plant genetic resources journals, 2013,01:173-178；
[5] it is straight in number of patent application 201310500940.2, near infrared spectrum and visible light analysis instrument quick detection rice
Method [P] .2014.02.12 of chain content of starch；
[6] a kind of sides that rice bran fats enzymatic activity is quickly determined using near-infrared of number of patent application 201110187788.8..
Method [P] .2011.12.14；
[7] such as Zhang Huaxiu near infrared spectrums combination CARS Variable Selections method contains for protein in liquid milk with fat
Measure [J] analysis test journals of amount, 2010,29 (5):430-434.
[8] national food safety standard of Ministry of Health of the People's Republic of China .GB 5009.5-2010:Protein in Food
Measure [S] Beijing:China Standards Press, 2010；
[9] People's Republic of China's national quality supervision and inspection quarantine epidemic disease general bureau, Standardization Administration of China
.GB/T crude fat content measure [S] Beijing in 5511-2008 grain and oil detection grains：China Standards Press, 2009；
[10] Ministry of Health of the People's Republic of China, the national food safety standard of GB5009.3-2010:Moisture in food
Determine [S] Beijing:China Standards Press, 2010；
[11] Standardization Administration of China of Ministry of Health of the People's Republic of China .GB/T 5009.7-2008 food
Measure [S] Beijing of middle reduced sugar：China Standards Press, 2009；
[12] Ministry of Health of the People's Republic of China, GB5009.4-2010 national food safety standard：The survey of ash content in food
Fixed [S] Beijing:China Standards Press, 2010；
[13] poem equality, wavelet noise and its application [J] west in Protein Content in Wheat near-infrared spectrum analysis are wished
Southern agriculture university's journal (natural science edition), 2003 (6):522-525；
[14] Lu Wanzhen etc., NIR spectra analysis [M] Beijing Sinopecs publishing house .2001,10-
12；
[15] small superfine, near-infrared spectral analysis technology and its application [M] Beijing electronics works in modern agriculture are opened
Industry publishing house .2012,105；
[16] Buddhist nun treasure etc., the effect of preprocessing procedures and its development [J] Pharmaceutical Analysis are miscellaneous in near-infrared spectrum analysis
Will, 2008.05:824-829.
[17]R W Kennard,L A Stone.Computer Aided Design of
Experiments.Technometrics,1969,11:137-148；
[18] Yan Yanlu is edited, and near-infrared spectrum analysis basis is with applying [M], Beijing：China Light Industry Press,
2005.01 98-112；
[19] Xiao Zhihong, Yu Jialin (author), multivariate statistics and SAS application Wuhan：Publishing house of Wuhan University .2013.
Claims (2)
1. a kind of near infrared detection method of rice quality, it is characterised in that comprise the following steps：
Testing sample is subjected to infrared diaphanoscopy, the original wave spectrogram of light absorption value based on wavelength is obtained, then by original wave spectrogram
Denoising Processing and pretreatment are carried out, characteristic wavelength is screened from the wave spectrum after processing, then characteristic wavelength substitution sample quality is referred to
In target forecast model, the near infrared detection value of sample quality is obtained；
Wherein：
The construction method of forecast model is as follows：
(1) collect representational Rice Samples, including paddy, brown rice, polished rice, husk, rice bran or crack rice；
(2) test chemical is carried out to collected sample, the test chemical value of gained is designated as ymj, wherein：M is m-th of index,
M=1,2,3 ..., 20；It is designated as paddy moisture as m=1, is designated as rice glutelin matter content during m=2, is designated as rice during m=3
Paddy fat content, it is designated as during m=4 as paddy total sugar content, is designated as paddy content of ashes during m=5, is designated as husk rate, m during m=6
It is designated as brown rice broken rice- containing rate when=7, is designated as broken rice rate during m=8, be designated as chaff powder rate during m=9, is designated as head rice rate, m=during m=10
It is designated as brown rice moisture content when 11, is designated as protein content of rice during m=12, is designated as Brown rice lipid content, m=14 during m=13
Shi Jiwei brown rice total sugar contents, are designated as brown rice content of ashes during m=15, are designated as polished rice moisture during m=16, remember during m=17
It is designated as polished rice fat content for polished rice protein content, during m=18, is designated as polished rice total sugar content during m=19, is designated as during m=20
Polished rice content of ashes；J is j-th of sample, common n sample, n >=40；
(3) infrared diaphanoscopy, light absorption value x are carried out to sampleij, wherein i i-th of wavelength of expression, wavelength i=1000nm,
1001nm, 1002nm ..., 1799nm；
(4) denoising Processing and pretreatment：With wavelet Denoising Method by xijDenoising Processing is carried out, de-noising light absorption value is obtained, then uses again
One or more combination treatment methods in normalization, first derivative, second dervative pre-process to de-noising light absorption value, obtain
To pretreatment light absorption value Aij；
(5) characteristic wavelength of near-infrared spectrum is screened：With competitive adaptive weight weight sampling (CARS) method and an inclined most young waiter in a wineshop or an inn
Multiplication (PLS) method screens characteristic wavelength of near-infrared spectrum, establishes the forecast model of rice quality：zm=bm+∑amiBi, wherein zm
For the near infrared detection value of rice quality, BiFor AijIn j-th of sample light absorption value, bm、amiFor regression coefficient, forecast model
Evaluation index is coefficient of determination R2With calibration standard deviation RMSEC, the conspicuousness of regression coefficient is examined with T, as coefficient amiIt is notable
Property be t values be more than 0.05 when, then make ami=0, amiWavelength i is characterized wavelength corresponding to ≠ 0 place；
Characteristic wavelength is as follows：
Paddy moisture：1310nm, 1402nm, 1593nm, 1738nm and 1772nm；
Rice glutelin matter：1206nm, 1254nm, 1274nm, 1563nm and 1752nm；
Paddy fat：1343nm, 1369nm, 1489nm, 1574nm and 1583nm；
Paddy total reducing sugar：1086nm, 1273nm, 1279nm, 1577nm and 1643nm；
Paddy ash content：1079nm, 1181nm, 1417nm, 1426nm and 1494nm；
Brown rice moisture：1026nm, 1102nm, 1213nm, 1313nm and 1746nm；
Brown rice protein：1168nm, 1170nm, 1250nm, 1780nm and 1779nm；
Brown rice fat：1625nm, 1536nm, 1712nm, 1026nm and 1042nm；
Brown rice total reducing sugar：1008nm, 1326nm, 1377nm, 1525nm and 1599nm；
Brown rice ash content：1073nm, 1068nm, 1141nm, 1259nm and 1785nm；
Polished rice moisture：1060nm, 1274nm, 1293nm, 1328nm and 1408nm；
Polished rice protein：1254nm, 1285nm, 1516nm, 1554nm and 1717nm；
Polished rice fat：1018nm, 1536nm, 1608nm, 1625nm and 1712nm；
Polished rice total reducing sugar：1304nm, 1338nm, 1617nm, 1726nm and 1745nm；
Polished rice ash content：1452nm, 1472nm, 1481nm, 1724nm and 1759nm；
Husk rate：1127nm, 1264nm, 1446nm, 1495nm and 1597nm；
Brown rice broken rice- containing rate：1123nm, 1301nm, 1317nm, 1326nm and 1681nm；
Broken rice rate：1183nm, 1243nm, 1579nm, 1584nm and 1723nm；
Chaff powder rate：1157nm, 1602nm, 1723nm, 1728nm and 1730nm；
Head rice rate：1114nm, 1151nm, 1257nm, 1659nm and 1680nm；
Features described above wavelength allows the deviation for having ± 2nm；
(6) amendment of forecast model：Increase by 1 sample, make sample number n=n+1, the sample that newly increases is carried out test chemical,
Infrared diaphanoscopy, denoising Processing and pretreatment, obtain new Aij, by the near-infrared characteristic wavelength described in step (5), obtain new
Rice quality near infrared detection value zmForecast model.
2. a kind of near infrared detection method of rice quality described in claim 1 rice, wheat, corn, Semen Coicis, soybean,
Sesame, peanut pellets shape solid cereal and rice bran, wheat flour granular material sample Quality Detection in application.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510657245.6A CN105181643B (en) | 2015-10-12 | 2015-10-12 | A kind of near infrared detection method of rice quality and application |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510657245.6A CN105181643B (en) | 2015-10-12 | 2015-10-12 | A kind of near infrared detection method of rice quality and application |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105181643A CN105181643A (en) | 2015-12-23 |
| CN105181643B true CN105181643B (en) | 2018-03-02 |
Family
ID=54903876
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510657245.6A Active CN105181643B (en) | 2015-10-12 | 2015-10-12 | A kind of near infrared detection method of rice quality and application |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105181643B (en) |
Families Citing this family (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105606559A (en) * | 2015-12-31 | 2016-05-25 | 深圳市芭田生态工程股份有限公司 | Method for spectrometric determination of vegetable protein in agricultural product |
| CN106706553A (en) * | 2016-03-17 | 2017-05-24 | 西北农林科技大学 | Method for quick and non-destructive determination of content of amylase in corn single grains |
| CN105973836A (en) * | 2016-05-06 | 2016-09-28 | 安徽贝通智能科技有限公司 | Method for on-line detection of grain water content in automatic sampling |
| CN106442397B (en) * | 2016-09-05 | 2019-03-19 | 中国科学院合肥物质科学研究院 | A kind of paddy near-infrared spectroscopy optimization method deducted based on spectrum ratio |
| CN106770016B (en) * | 2017-01-16 | 2023-04-07 | 中国科学院合肥物质科学研究院 | NIR (near infrared) transmission spectrum measurement method for protein quantitative analysis of single rice seeds |
| CN106970041B (en) * | 2017-04-12 | 2020-04-07 | 江苏省农业科学院 | Near-infrared determination method for insoluble glutelin macro-polymer content of wheat flour |
| CN107219827B (en) * | 2017-05-02 | 2020-06-05 | 华中农业大学 | Rice food full-industry chain information system, establishment method and application |
| CN107764767A (en) * | 2017-09-20 | 2018-03-06 | 农业部南京农业机械化研究所 | The method of the corn protein and moisture of detection united reaper in real time |
| CN108169168A (en) * | 2017-12-19 | 2018-06-15 | 信阳师范学院 | Test and analyze rice grain protein content mathematical model and construction method and application |
| CN108982408A (en) * | 2018-07-20 | 2018-12-11 | 哈尔滨工业大学 | A method of organic rice and non-organic rice are distinguished using near-infrared spectrum technique |
| CN109374548A (en) * | 2018-11-14 | 2019-02-22 | 深圳职业技术学院 | A method of quickly measuring nutritional ingredient in rice using near-infrared |
| CN109839358B (en) * | 2019-01-22 | 2021-08-10 | 北京农业质量标准与检测技术研究中心 | Agricultural product quality analysis method and device |
| CN109975237A (en) * | 2019-03-04 | 2019-07-05 | 浙江大学 | The quickly method of measurement rice milled rice flour phytic acid |
| CN109946250A (en) * | 2019-03-27 | 2019-06-28 | 浙江大学 | The quickly method of measurement rice total phosphorus content |
| CN110044842A (en) * | 2019-05-22 | 2019-07-23 | 江南大学 | A kind of method of the odd sub- seed moisture content of quick measurement |
| TWI707140B (en) * | 2019-10-02 | 2020-10-11 | 國立中興大學 | Device for detecting moisture content of cereals |
| CN111337450A (en) * | 2020-03-16 | 2020-06-26 | 中国科学院合肥物质科学研究院 | Rapid nondestructive determination method and device for rice processing precision grade |
| CN112461784A (en) * | 2020-11-16 | 2021-03-09 | 山东大学 | Near infrared spectrum-based drug cocrystal detection method and application thereof |
| CN112485216B (en) * | 2020-11-20 | 2022-02-01 | 华中农业大学 | Multi-source information fusion Thailand jasmine rice adulteration identification method |
| CN112630180B (en) * | 2020-12-14 | 2022-02-08 | 华南农业大学 | Ultraviolet/visible light absorption spectrum model for detecting concentration of organophosphorus pesticide in water body |
| CN113484270A (en) * | 2021-06-04 | 2021-10-08 | 中国科学院合肥物质科学研究院 | Construction and detection method of single-grain rice fat content quantitative analysis model |
| CN114018859B (en) * | 2021-10-13 | 2023-03-17 | 中国水稻研究所 | Method for rapidly and synchronously measuring apparent amylose, amylose and amylopectin contents of rice flour |
| CN114136918B (en) * | 2021-11-29 | 2023-11-14 | 中国科学院合肥物质科学研究院 | Near infrared-based rice taste quality evaluation method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101655454A (en) * | 2009-09-15 | 2010-02-24 | 北京市农林科学院 | Rapid determination method for evaluation of storage quality of grain |
| CN102279166A (en) * | 2011-07-06 | 2011-12-14 | 长沙理工大学 | Method for rapidly measuring activity of lipase in rice bran by utilizing near-infrared |
| CN103487384A (en) * | 2013-10-15 | 2014-01-01 | 无锡艾科瑞思产品设计与研究有限公司 | Near-infrared-spectrum-based method for rapidly detecting components of bagged rice |
| CN103575689A (en) * | 2013-10-11 | 2014-02-12 | 西北农林科技大学 | Method for rapidly detecting amylose content in rice by near infrared spectrum and visible light analyzer |
| CN104215584A (en) * | 2014-08-29 | 2014-12-17 | 华南理工大学 | Hyper-spectral image technology-based detection method for distinguishing rice growing areas |
-
2015
- 2015-10-12 CN CN201510657245.6A patent/CN105181643B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101655454A (en) * | 2009-09-15 | 2010-02-24 | 北京市农林科学院 | Rapid determination method for evaluation of storage quality of grain |
| CN102279166A (en) * | 2011-07-06 | 2011-12-14 | 长沙理工大学 | Method for rapidly measuring activity of lipase in rice bran by utilizing near-infrared |
| CN103575689A (en) * | 2013-10-11 | 2014-02-12 | 西北农林科技大学 | Method for rapidly detecting amylose content in rice by near infrared spectrum and visible light analyzer |
| CN103487384A (en) * | 2013-10-15 | 2014-01-01 | 无锡艾科瑞思产品设计与研究有限公司 | Near-infrared-spectrum-based method for rapidly detecting components of bagged rice |
| CN104215584A (en) * | 2014-08-29 | 2014-12-17 | 华南理工大学 | Hyper-spectral image technology-based detection method for distinguishing rice growing areas |
Non-Patent Citations (5)
| Title |
|---|
| Near-infrared Spectroscopy of Cereals;Phil Williams;《Agricultural Applications》;20061231;全文 * |
| 大米蛋白质含量近红外光谱检测模型研究;刘明博等;《中国农学通报》;20121231;第29卷(第12期);全文 * |
| 近红外光谱分析技术在大米检测中的应用;齐龙等;《农机化研究》;20110731(第7期);全文 * |
| 近红外光谱快速测定稻谷水分含量的研究;范维燕等;《粮油食品科技》;20081231;第16卷(第5期);全文 * |
| 近红外光谱技术在稻米特性检测中的应用;吴维政等;《亚热带植物科学》;20141231;第43卷(第2期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105181643A (en) | 2015-12-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105181643B (en) | A kind of near infrared detection method of rice quality and application | |
| Su et al. | Non-destructive and rapid evaluation of staple foods quality by using spectroscopic techniques: a review | |
| Zhang et al. | Determination of water content in corn stover silage using near-infrared spectroscopy | |
| Williams et al. | Comparison of commercial near infrared transmittance and reflectance instruments for analysis of whole grains and seeds | |
| Zhao et al. | Near infrared reflectance spectroscopy for determination of the geographical origin of wheat | |
| CN105181642B (en) | A kind of near infrared detection method of peanut quality and application | |
| Chen et al. | Determination of protein, total carbohydrates and crude fat contents of foxtail millet using effective wavelengths in NIR spectroscopy | |
| Revilla et al. | Predicting the physicochemical properties and geographical ORIGIN of lentils using near infrared spectroscopy | |
| US20210247318A1 (en) | Non-Destructive Detection of Egg Freshness Based on Raman Spectroscopy | |
| Hadiwijaya et al. | Multi-product calibration model for soluble solids and water content quantification in Cucurbitaceae family, using visible/near-infrared spectroscopy | |
| Arazuri et al. | Rheological parameters determination using Near Infrared technology in whole wheat grain | |
| Malvandi et al. | Non-destructive measurement and real-time monitoring of apple hardness during ultrasonic contact drying via portable NIR spectroscopy and machine learning | |
| Liu et al. | Determination of total protein and wet gluten in wheat flour by Fourier transform infrared photoacoustic spectroscopy with multivariate analysis | |
| Yang et al. | Quantifying soluble sugar in super sweet corn using near-infrared spectroscopy combined with chemometrics | |
| Wang et al. | Development of near‐infrared online grading device for long jujube | |
| Jiang et al. | Rapid determination of acidity index of peanuts by near-infrared spectroscopy technology: Comparing the performance of different near-infrared spectral models | |
| John et al. | Germplasm variability-assisted near infrared reflectance spectroscopy chemometrics to develop multi-trait robust prediction models in rice | |
| Liu et al. | Prediction of talc content in wheat flour based on a near-infrared spectroscopy technique | |
| Padhi et al. | Development and optimization of NIRS prediction models for simultaneous multi-trait assessment in diverse cowpea germplasm | |
| CN110231302A (en) | A kind of method of the odd sub- seed crude fat content of quick measurement | |
| Fan et al. | Rapid and Nondestructive Determination of origin, volatile oil, sanshoamides and crack rate in the ‘Sichuan Pepper’Based on a Novel Portable Near Infrared Spectrometer | |
| Qian et al. | Identification of Baha'sib mung beans based on Fourier transform near infrared spectroscopy and partial least squares | |
| Archibald et al. | Determination of total dietary fiber of intact cereal food products by near-infrared reflectance | |
| Budiastra et al. | Determination of Chemical Content of Porang Flour (Amorphophallus muelleri blume) by Near Infrared Spectroscopy | |
| Phanomsophon et al. | Primary assessment of macronutrients in durian (CV Monthong) leaves using near infrared spectroscopy with wavelength selection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |