CN1657907A - Agricultural products, food near-infrared spectral specragion selection method - Google Patents
Agricultural products, food near-infrared spectral specragion selection method Download PDFInfo
- Publication number
- CN1657907A CN1657907A CN2005100385289A CN200510038528A CN1657907A CN 1657907 A CN1657907 A CN 1657907A CN 2005100385289 A CN2005100385289 A CN 2005100385289A CN 200510038528 A CN200510038528 A CN 200510038528A CN 1657907 A CN1657907 A CN 1657907A
- Authority
- CN
- China
- Prior art keywords
- interval
- spectrum
- pls
- rmsecv
- rmsep
- 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
Images
Landscapes
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention relates to a method which is used for near-infrared spectral analysing the quality of the product and food, selects the width of the appropriate characteristic spectrum spectrum area, for the near-infrared spectrum divides the obtained entire near-infrared spectrum into certain sectors, then separately processing the PLS to each sector; Obtains the best PLS model of each sector through comparison orthogonal examination root-mean-square error RMSECV compares each sector best PLS model of RMSEVC and the RMSEP selection characteristic spectrum area sector. Finally carries on the PLS analysis establishment model for the characteristic wave length sector. The advantages are easy obtaining the width and the best characteristic sector and high forecast ability and precision.
Description
Technical field
The present invention relates to utilize the method for near-infrared spectrum analysis quality of agricultural product food, refer in particular to agricultural product, food near-infrared spectral system of selection based on interval partial least square.
Background technology
Development along with near-infrared spectrum technique and stoechiometric process, near-infrared spectrum technique is applied in quality of agricultural product, the food analysis more and more widely, near-infrared spectral analysis technology with it at a high speed, accurately, and characteristics such as measuring-signal digitizing and analytic process greenization make it become with fastest developing speed, the most noticeable spectral analysis technique since the nineties in last century.
The application foundation theory of near-infrared spectrum analysis is a Lambert-Beer's law.In fact the suitable condition of Lambert-Beer's law is pure element or compound, and the near infrared spectrum of the such multi-component material of agricultural product, food to absorb mainly be that the frequency multiplication of molecule absorbs and sum of fundamental frequencies absorbs, the information that has comprised very abundant multi-component material in the absorption spectrum, because the stochastic error that multi-component phase mutual interference is closed in the spectral detection all can have a strong impact on the analyzing and testing precision, therefore use Lambert-Beer's law can produce very big error simply, be difficult to obtain desirable effect.So choose the method for information in the near infrared spectrum district is the one of the main reasons of restriction near-infrared spectral analysis technology always.
The overall tendency of the analysis of the near infrared spectrum of most of agricultural product, food is milder, and crest and trough do not have violent fluctuating.The spectroscopic data that single wavelength obtains down is difficult to obtain analytical model accurately, and the analysis of modern near infrared spectrum data is all carried out under multi-wavelength.Graphic interpretation and expertise are chosen crest, trough and component characteristics wavelength, and to set up model almost be impossible.Existing polynary alignment technique such as principal component regression (principalcomponentre-gression, be called for short PCR) or partial least square method (partialleastsquares, be called for short PLS) to agricultural product, when food near-infrared spectrum data are set up spectral prediction model, need to determine the characteristic wavelength spectrum district of specific components, reduce modeling and predict operation time, and cancelling noise pollutes excessive spectrum district etc., all will face the problem of selecting suitable light spectral.
Summary of the invention
For overcoming the deficiency of above-mentioned technology, the purpose of this invention is to provide a kind of agricultural product Region Selection Method of Near Infrared Spectrum based on interval partial least square.
Described agricultural product Region Selection Method of Near Infrared Spectrum based on interval partial least square comprises following processing:
To the near infrared spectrum after the denoising, choose the width of suitable feature light spectral, the whole near infrared spectrum of gained is divided between several region; Then PLS being carried out in each interval respectively handles; By comparing the best PLS model that orthogonal test root-mean-square error RMSECV and forecast set root-mean-square error RMSEP (Root Mean Square Error of Cross Validation/Prediction) obtain each interval; Same by relatively the RMSECV and the RMSEP selected characteristic of the PLS model of the best in each interval are composed trivial.At last the PLS analysis is carried out in selected characteristic wavelength interval and set up model.
Near infrared spectrum is meant by original spectrum is carried out suitable pre-service and reduces or eliminates the factor that various spectrum exert an influence to spectrum after the described denoising, the spectrum after the purification, and this spectrum comprises a calibration set and a forecast set.
The choosing method of described suitable characteristics spectrum spectrum sector width has random choice method, halving method and moving window method.
Described random choice method is rule of thumb to get n point on whole spectrum at random, and spectrum is divided into n+1 interval, and the spectrum point number in each is interval not necessarily equates.
Described halving method is that whole spectrum is divided into n interval, and the spectroscopic data point number in each is interval equates.
Described moving window method may further comprise the steps:
(1): choose an initial window width, be generally the width of 30～50 spectroscopic datas;
(2): on the spectrum axle with step-length be a spectroscopic data point move this window, intercept the spectroscopic data of window width at every turn;
(3): the data in each intercepting window are carried out PLS handle, and the orthogonal test root-mean-square error RMSECV and the forecast set root-mean-square error RMSEP of the best PLS model of each window preserved.
(4): allow window width increase by 10 spectroscopic datas then, repeat (2), (3) step, obviously increase and stop when maximum with window up to RMSEP;
(5) relatively more all RMSECV and RMSEP preserves when window width and spectroscopic data that RMSECV and RMSEP reach hour simultaneously, and the window width of this moment is final spectrum spectrum sector width, and the spectroscopic data interval of this moment is the best features interval.
Trivial of described selected characteristic spectrum can be more than one, when the PLS model accuracy of being set up when best features recited above interval is not high enough, can choose a plurality of characteristic intervals.
Choosing of described a plurality of characteristic intervals may further comprise the steps:
(1) on the both sides in best features interval, the window of composing sector width with above-mentioned final spectrum is divided into whole spectrum between several region;
(2) each interval interior data is carried out the PLS modeling, calculate the RMSECV and the RMSEP of each model.
(3) the RMSECV value being pressed in those intervals sorts from small to large.
(4) get in (3) in interval two intervals altogether in first interval (interval of RMSECV and RMSEP mean value minimum just) after the ordering and best features data and carry out the PLS modeling, calculate the RMSECV and the RMSEP of model at this moment simultaneously;
(5) get in (3) preceding two intervals after the ordering and best features interval altogether three each and every one interval in data carry out the PLS modeling, calculate the RMSECV and the RMSEP of model at this moment simultaneously;
(6) continue to increase the interval, up to the RMSECV and the RMSEP minimum of the PLS model of being set up, and till the related coefficient maximum, then participating in the interval of modeling this moment is final characteristic interval.
Because the present invention adopts above technical scheme, obtains following effect:
Solved with near infrared spectrum by the moving window method and to have carried out quality of agricultural product characteristic spectrum area width size issue when detecting modeling, can obtain the width and the best characteristic interval of characteristic light spectral easily.Solved the problem that many characteristic intervals are selected by crosscheck mean square deviation and prediction mean square deviation minimum.Can reduce modeling operation time by choosing of characteristic interval, the spectrum district that cancelling noise is excessive, it is higher to make the quality of agricultural product of final foundation detect the predictive ability and the precision of near infrared light spectrum model.
Description of drawings
Fig. 1 processing procedure flow diagram
Fig. 2 is through pretreated 124 apple near infrared spectrum data such as denoising, centralizations;
The whole apple spectrum of Fig. 3 is divided into 40 intervals, the RMSECV of each interval best PLS model (the italic numeral is the number of variable in the PLS model), and thick black line is an apple near infrared light spectral shape among the figure, dotted line is the RMSECV that whole spectrum participates in the PLS modeling;
Fig. 4 chooses the spectroscopic data in the 2nd, 3,5,8,9,10,12,13,22 intervals and sets up the iPLS model;
When Fig. 5 moving window width is 72 spectral widths apple sugar content is carried out the near infrared modeling, change situation when RMSCV moves with window, thick line is typical apple near infrared spectrum among the figure, dotted line is the whole spectrum RMSCV when participating in modeling;
Minimum RMSECV of Fig. 6 and RMSECP increase situation of change with window width;
The whole beer spectrum of Fig. 7 is divided into 20 intervals, the RMSECV of each interval best PLS model;
Fig. 8 chooses the PLS modeling situation that the beer data are carried out in the 10th interval
Embodiment
Embodiment describes in conjunction with following two embodiments.
Embodiment 1: interval partial least square is selected apple sugar content near-infrared analysis light spectral.
Fig. 1 is the synoptic diagram of processing procedure, Fig. 2 is through pretreated 124 apple near infrared spectrum data such as denoising, centralizations, spectral range is 4279～9843cm-1, every spectrum comprises 2886 data points, as forecast set, 46 apple spectrum are as forecast set with the spectroscopic data of 78 apples wherein.Choosing suitable feature spectrum spectrum sector width by the moving window method is 72 spectrum points, and whole spectrum is divided into 40 intervals.PLS is carried out in each interval handles, each interval best PLS model as shown in Figure 3, choosing wherein, the spectrum in the 2nd, 3,5,8,9,10,12,13,22 intervals carries out the PLS modeling, the result of gained such as Fig. 4.Wherein spectral width is that 72 spectrum points are chosen by the following method:
Choose the width that the home window width is 32 spectroscopic datas;
On the spectrum axle with step-length be a spectroscopic data point move this window, intercept 32 spectroscopic datas at every turn;
Data in each intercepting window are carried out PLS handle, and the orthogonal test lowest mean square root error RMSECV and the forecast set root-mean-square error RMSEP of the best PLS model of each window preserved;
Allow window width increase by 10 spectroscopic datas then, be the width of 42 spectroscopic datas, repeat (2), (3) step, stop up to 142 spectrum point width of window width;
Fig. 5 is carrying out the modeling of apple sugar content near infrared, when the moving window width is 72 spectral widths, and change situation when RMSCV moves with window, thick line is an apple near infrared spectrum among the figure, dotted line is the whole spectrum RMSCV when participating in modeling.
When Fig. 6 is the ascending variation of moving window width, each minimum RMSECV, RMSEP situation of change that obtains.RMSECV reduces along with the increase of window width among the figure, reduce along with the increase of window width when RMSEP begins, but RMSEP becomes big when continue increasing with rear hatch.RMSEP reaches minimum during 72 spectrum points of window width as seen from Figure 6.Therefore the best window width is 72 spectrum points.
Interval selection realizes by step once:
Choose the 12nd interval of RMSECV minimum, the PLS model related coefficient of setting up on its data has only 0.7, RMSCV=0.82, and obviously model is good inadequately;
On the basis in the 12nd interval, add the 2nd interval, on the data in two intervals, set up the PLS model, but the result still is bad, continue to increase intervally, the characteristic interval that final characteristic interval is determined is: 2,3,5,8,9,10,12,13,22 data in 9 intervals altogether.The related coefficient of the PLS model of setting up reaches 0.8958, RMSECV=0.5892
To 60 beer test near infrared spectrum samples, spectral range is 400～2250nm, and every spectrum comprises 926 data points, and as forecast set, 20 beer spectrum are as forecast set with the spectroscopic data of 40 beer wherein.Choosing suitable feature spectrum spectrum sector width by the moving window method is 41 spectrum points, and whole spectrum is divided into 20 intervals.PLS is carried out in each interval to be handled, each interval best PLS model as shown in Figure 7, choose the data in the 10th interval and carry out the PLS modeling, the result who obtains as shown in Figure 8, this moment, related coefficient reached 0.9981, RMSECV=0.151, and precision is fine, therefore the finally definite interval of iPLS is the 10th interval, and spectral range is: 1240～1330nm.
Claims (6)
1. based on agricultural product, the food near-infrared spectral system of selection of interval partial least square, it is characterized in that at first the near infrared spectrum after the denoising is chosen the width of suitable feature light spectral, the whole near infrared spectrum of gained is divided between several region; Then PLS being carried out in each interval respectively handles; By comparing the best PLS model that orthogonal test root-mean-square error RMSECV and forecast set root-mean-square error RMSEP obtain each interval; Same by relatively the RMSECV and the RMSEP selected characteristic of the PLS model of the best in each interval are composed trivial; At last the PLS analysis is carried out in selected characteristic wavelength interval and set up model.
2. system of selection according to claim 1 is characterized in that the choosing method of described suitable characteristics spectrum spectrum sector width is random choice method, halving method or moving window method.
3. system of selection according to claim 1 is characterized in that described random choice method, is rule of thumb to get n point at random on whole spectrum, and spectrum is divided into n+1 interval, and the spectrum point number in each is interval not necessarily equates.
4. system of selection according to claim 1 is characterized in that described halving method, is whole spectrum is divided into n interval, and the spectroscopic data point number in each is interval equates.
5. system of selection according to claim 1 is characterized in that described moving window method may further comprise the steps:
(1) chooses an initial window width, be generally the width of 30～50 spectroscopic datas;
(2) on the spectrum axle with step-length be a spectroscopic data point move this window, intercept the spectroscopic data of window width at every turn;
(3) data in each intercepting window are carried out PLS and handle, and the orthogonal test root-mean-square error RMSECV and the forecast set root-mean-square error RMSEP of the best PLS model of each window preserved;
(4) allow window width increase by 10 spectroscopic datas then, repeat (2), (3) step, obviously increase and stop when maximum with window up to RMSEP;
(5) relatively more all RMSECV and RMSEP preserves when window width and spectroscopic data that RMSECV and RMSEP reach hour simultaneously, and the window width of this moment is final spectrum spectrum sector width, and the spectroscopic data interval of this moment is the best features interval.
6. system of selection according to claim 1 is characterized in that choosing of described a plurality of characteristic intervals may further comprise the steps:
(1) on the both sides in best features interval, the window of composing sector width with above-mentioned final spectrum is divided into whole spectrum between several region;
(2) each interval interior data is carried out the PLS modeling, calculate the RMSECV and the RMSEP of each model;
(3) the RMSECV value being pressed in those intervals sorts from small to large;
(4) get in (3) in interval two intervals altogether in first interval (interval of RMSECV and RMSEP mean value minimum just) after the ordering and best features data and carry out the PLS modeling, calculate the RMSECV and the RMSEP of model at this moment simultaneously;
(5) get in (3) preceding two intervals after the ordering and best features interval altogether three each and every one interval in data carry out the PLS modeling, calculate the RMSECV and the RMSEP of model at this moment simultaneously;
(6) continue to increase the interval, up to the RMSECV and the RMSEP minimum of the PLS model of being set up, and till the related coefficient maximum, then participating in the interval of modeling this moment is final characteristic interval.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2005100385289A CN1657907A (en) | 2005-03-23 | 2005-03-23 | Agricultural products, food near-infrared spectral specragion selection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2005100385289A CN1657907A (en) | 2005-03-23 | 2005-03-23 | Agricultural products, food near-infrared spectral specragion selection method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN1657907A true CN1657907A (en) | 2005-08-24 |
Family
ID=35007536
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2005100385289A Pending CN1657907A (en) | 2005-03-23 | 2005-03-23 | Agricultural products, food near-infrared spectral specragion selection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1657907A (en) |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101807228A (en) * | 2010-03-12 | 2010-08-18 | 江苏大学 | Method for selecting subinterval of near-infrared spectrum wavelength based on simulated annealing algorithm |
| CN101275912B (en) * | 2008-05-08 | 2010-08-25 | 中国农业大学 | Fluid food browning testing method |
| CN101832909A (en) * | 2010-03-12 | 2010-09-15 | 江苏大学 | Selection method for subintervals of near infrared spectral characteristics based on simulated annealing-genetic algorithm |
| CN101832922A (en) * | 2010-05-19 | 2010-09-15 | 中国农业大学 | Method for transferring near infrared model of organic fertilizer product |
| CN101788459B (en) * | 2010-02-08 | 2012-07-18 | 暨南大学 | Quasi-continuous spectroscopic wavelength combination method |
| CN103308463A (en) * | 2013-06-28 | 2013-09-18 | 中国农业大学 | Characteristic spectrum area selection method for near infrared spectrum |
| CN103335978A (en) * | 2013-07-02 | 2013-10-02 | 暨南大学 | Light splitting wavelength screening method based on partner wavelength |
| CN104091089A (en) * | 2014-07-28 | 2014-10-08 | 温州大学 | Infrared spectrum data PLS modeling method |
| CN104155281A (en) * | 2014-07-07 | 2014-11-19 | 江苏大学 | Method for detecting concentrations of ethanol aqueous solutions at different temperatures |
| CN104990895A (en) * | 2015-07-27 | 2015-10-21 | 浙江中烟工业有限责任公司 | Near infrared spectral signal standard normal correction method based on local area |
| CN105334186A (en) * | 2015-12-10 | 2016-02-17 | 山东大学 | Infrared spectral analysis method |
| CN105806803A (en) * | 2016-03-15 | 2016-07-27 | 潘涛 | Multi-index collaborative analysis wavelength combination and selection method thereof |
| CN105866042A (en) * | 2016-04-27 | 2016-08-17 | 南京林业大学 | Detection method for spatial distribution of biological quality indexes based on pixel index unbiased estimation method |
| CN105891147A (en) * | 2016-03-30 | 2016-08-24 | 浙江中烟工业有限责任公司 | Near infrared spectrum information extraction method based on canonical correlation coefficients |
| CN106198446A (en) * | 2016-06-22 | 2016-12-07 | 中国热带农业科学院热带作物品种资源研究所 | The method of L-Borneol content near infrared spectrum quick test Herba Blumeae Balsamiferae leaf powder |
| CN106596456A (en) * | 2016-04-01 | 2017-04-26 | 南京理工大学 | Changeable moving window based selection method of near infrared spectral region of solution |
| CN106644983A (en) * | 2016-12-28 | 2017-05-10 | 浙江大学 | Spectrum wavelength selection method based on PLS-VIP-ACO algorithm |
| CN106918567A (en) * | 2017-03-27 | 2017-07-04 | 中南大学 | A kind of method and apparatus for measuring trace metal ion concentration |
| CN112816435A (en) * | 2021-02-01 | 2021-05-18 | 广东省农业科学院蔬菜研究所 | Method for detecting pumpkin sugar content by utilizing near infrared spectrum |
| CN114324233A (en) * | 2021-11-16 | 2022-04-12 | 贵州省生物技术研究所(贵州省生物技术重点实验室、贵州省马铃薯研究所、贵州省食品加工研究所) | Near-infrared nondestructive online quality detection method and system for nutritional ingredients of agricultural products |
-
2005
- 2005-03-23 CN CN2005100385289A patent/CN1657907A/en active Pending
Cited By (31)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101275912B (en) * | 2008-05-08 | 2010-08-25 | 中国农业大学 | Fluid food browning testing method |
| CN101788459B (en) * | 2010-02-08 | 2012-07-18 | 暨南大学 | Quasi-continuous spectroscopic wavelength combination method |
| WO2011109922A1 (en) * | 2010-03-12 | 2011-09-15 | 江苏大学 | Near-infrared spectrum characteristic subinterval selection method based on simulated annealing-genetic algorithm |
| CN101807228A (en) * | 2010-03-12 | 2010-08-18 | 江苏大学 | Method for selecting subinterval of near-infrared spectrum wavelength based on simulated annealing algorithm |
| CN101832909B (en) * | 2010-03-12 | 2012-01-18 | 江苏大学 | Selection method for subintervals of near infrared spectral characteristics based on simulated annealing-genetic algorithm |
| CN101807228B (en) * | 2010-03-12 | 2012-07-04 | 江苏大学 | Method for selecting subinterval of near-infrared spectrum wavelength based on simulated annealing algorithm |
| CN101832909A (en) * | 2010-03-12 | 2010-09-15 | 江苏大学 | Selection method for subintervals of near infrared spectral characteristics based on simulated annealing-genetic algorithm |
| CN101832922A (en) * | 2010-05-19 | 2010-09-15 | 中国农业大学 | Method for transferring near infrared model of organic fertilizer product |
| CN101832922B (en) * | 2010-05-19 | 2012-04-18 | 中国农业大学 | Method for transferring near infrared model of organic fertilizer product |
| CN103308463A (en) * | 2013-06-28 | 2013-09-18 | 中国农业大学 | Characteristic spectrum area selection method for near infrared spectrum |
| CN103335978A (en) * | 2013-07-02 | 2013-10-02 | 暨南大学 | Light splitting wavelength screening method based on partner wavelength |
| CN104155281A (en) * | 2014-07-07 | 2014-11-19 | 江苏大学 | Method for detecting concentrations of ethanol aqueous solutions at different temperatures |
| CN104091089A (en) * | 2014-07-28 | 2014-10-08 | 温州大学 | Infrared spectrum data PLS modeling method |
| CN104091089B (en) * | 2014-07-28 | 2016-04-27 | 温州大学 | A kind of ir data PLS modeling method |
| CN104990895A (en) * | 2015-07-27 | 2015-10-21 | 浙江中烟工业有限责任公司 | Near infrared spectral signal standard normal correction method based on local area |
| CN104990895B (en) * | 2015-07-27 | 2017-10-31 | 浙江中烟工业有限责任公司 | A kind of near infrared spectrum signal standards normal state bearing calibration based on regional area |
| CN105334186A (en) * | 2015-12-10 | 2016-02-17 | 山东大学 | Infrared spectral analysis method |
| CN105806803B (en) * | 2016-03-15 | 2019-07-02 | 潘涛 | A kind of multi objective Cooperative Analysis wavelength combination and its selection method |
| CN105806803A (en) * | 2016-03-15 | 2016-07-27 | 潘涛 | Multi-index collaborative analysis wavelength combination and selection method thereof |
| CN105891147A (en) * | 2016-03-30 | 2016-08-24 | 浙江中烟工业有限责任公司 | Near infrared spectrum information extraction method based on canonical correlation coefficients |
| CN106596456A (en) * | 2016-04-01 | 2017-04-26 | 南京理工大学 | Changeable moving window based selection method of near infrared spectral region of solution |
| CN105866042B (en) * | 2016-04-27 | 2018-06-01 | 南京林业大学 | Based on pixel index unbiased esti-mator method biology index of quality spatial distribution detection method |
| CN105866042A (en) * | 2016-04-27 | 2016-08-17 | 南京林业大学 | Detection method for spatial distribution of biological quality indexes based on pixel index unbiased estimation method |
| CN106198446A (en) * | 2016-06-22 | 2016-12-07 | 中国热带农业科学院热带作物品种资源研究所 | The method of L-Borneol content near infrared spectrum quick test Herba Blumeae Balsamiferae leaf powder |
| CN106644983A (en) * | 2016-12-28 | 2017-05-10 | 浙江大学 | Spectrum wavelength selection method based on PLS-VIP-ACO algorithm |
| CN106644983B (en) * | 2016-12-28 | 2019-12-31 | 浙江大学 | Spectral wavelength selection method based on PLS-VIP-ACO algorithm |
| CN106918567A (en) * | 2017-03-27 | 2017-07-04 | 中南大学 | A kind of method and apparatus for measuring trace metal ion concentration |
| CN106918567B (en) * | 2017-03-27 | 2019-05-28 | 中南大学 | A kind of method and apparatus measuring trace metal ion concentration |
| CN112816435A (en) * | 2021-02-01 | 2021-05-18 | 广东省农业科学院蔬菜研究所 | Method for detecting pumpkin sugar content by utilizing near infrared spectrum |
| WO2022160662A1 (en) * | 2021-02-01 | 2022-08-04 | 广东省农业科学院蔬菜研究所 | Method for measuring content of sugar in pumpkins by means of near infrared spectrum instrument |
| CN114324233A (en) * | 2021-11-16 | 2022-04-12 | 贵州省生物技术研究所(贵州省生物技术重点实验室、贵州省马铃薯研究所、贵州省食品加工研究所) | Near-infrared nondestructive online quality detection method and system for nutritional ingredients of agricultural products |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1657907A (en) | Agricultural products, food near-infrared spectral specragion selection method | |
| US20100297291A1 (en) | Visible/near-infrared spectrum analyzing method and grape fermenting method | |
| CN101915744B (en) | Near infrared spectrum nondestructive testing method and device for material component content | |
| CN107300536B (en) | Hyperspectrum-based mango impact damage soluble solid content prediction method | |
| CN108918446B (en) | Ultra-low concentration sulfur dioxide ultraviolet difference feature extraction algorithm | |
| CN103983617A (en) | Improved laser probe quantitative analysis method based on wavelet transform | |
| CN108169165B (en) | Maltose mixture quantitative analysis method based on terahertz spectrum and image information fusion | |
| CN106596456A (en) | Changeable moving window based selection method of near infrared spectral region of solution | |
| CN101788459B (en) | Quasi-continuous spectroscopic wavelength combination method | |
| CN1648638A (en) | Multiple component gas concentration quantitative analyzer based on supporting vector machine | |
| CN116559110A (en) | Self-adaptive near infrared spectrum transformation method based on correlation and Gaussian curve fitting | |
| CN109883990A (en) | A kind of medicinal fungi near-infrared spectral analytical method | |
| CN111766210B (en) | Near-shore complex seawater nitrate nitrogen multispectral measurement method | |
| Wei et al. | Confocal Raman microspectroscopy combined with spectral screening algorithms for quantitative analysis of starch in rice | |
| CN114062306B (en) | Near infrared spectrum data segmentation preprocessing method | |
| JP2002340792A (en) | Method for measuring density of polymer material | |
| CN113049526B (en) | Corn seed moisture content determination method based on terahertz attenuated total reflection | |
| CN115146219A (en) | Raman spectrum spike removing method and system based on derivative spectrum and cubic smoothing | |
| CN115598162A (en) | Soil heavy metal content detection method based on stacked model | |
| CN108226090B (en) | Method for constructing component content detection model | |
| CN112229817A (en) | Method for establishing soda saline-alkali soil heavy metal quantitative inversion model | |
| Luo et al. | Prediction of the Harvest Time of Cabernet Sauvignon Grapes Using Near-Infrared Spectroscopy | |
| Uddin et al. | Development of method for rapid quantification of glucose, fructose and sucrose in mango juice by chemometric techniques in de-noised FTIR spectroscopic data | |
| CN115290628B (en) | Quality analysis method of gasoline | |
| CN117589710A (en) | Method for determining components of gasoline unconventional additive |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |