CN106404748A - Multispectral combined laser induced breakdown spectroscopy cereal crop producing area identification method - Google Patents
Multispectral combined laser induced breakdown spectroscopy cereal crop producing area identification method Download PDFInfo
- Publication number
- CN106404748A CN106404748A CN201610802039.4A CN201610802039A CN106404748A CN 106404748 A CN106404748 A CN 106404748A CN 201610802039 A CN201610802039 A CN 201610802039A CN 106404748 A CN106404748 A CN 106404748A
- Authority
- CN
- China
- Prior art keywords
- sample
- rice
- cereal crops
- spectral line
- spectrum
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
- G01N21/71—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light thermally excited
- G01N21/718—Laser microanalysis, i.e. with formation of sample plasma
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N1/00—Sampling; Preparing specimens for investigation
- G01N1/28—Preparing specimens for investigation including physical details of (bio-)chemical methods covered elsewhere, e.g. G01N33/50, C12Q
- G01N1/286—Preparing specimens for investigation including physical details of (bio-)chemical methods covered elsewhere, e.g. G01N33/50, C12Q involving mechanical work, e.g. chopping, disintegrating, compacting, homogenising
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/02—Food
- G01N33/10—Starch-containing substances, e.g. dough
Landscapes
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Immunology (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Analytical Chemistry (AREA)
- Pathology (AREA)
- Engineering & Computer Science (AREA)
- Optics & Photonics (AREA)
- Plasma & Fusion (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Food Science & Technology (AREA)
- Medicinal Chemistry (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The method discloses a multispectral combined laser induced breakdown spectroscopy cereal crop producing area identification method. The method comprises SVM classification modeling and producing area identification. Through combination of a support vector machine algorithm and LIBS rapid material analysis, the method realizes rapid and accurate classification of rice products in different producing areas. Through use of similar large signal-to-background ratio characteristic spectral lines of the same element, difference of cereals such as rice in different producing areas is increased so that algorithm recognition accuracy is improved. Through combination of laser induced breakdown spectroscopy and SVM algorithm, fast and accurate identification is realized.
Description
Technical field
The invention belongs to field of spectral analysis technology is and in particular to a kind of identify paddy using LIBS
The method in the place of production of class crop, especially rice, ground is also referred to as traced to the source in the place of production.
Background technology
Rice is described as " first of five cereals ", is one of staple food crop of China, accounts for plant cultivation grain area
30%, because being rich in abundant carbohydrate, vitamin and mineral matter, and become the various micronutrient element of supplementary needed by human body
Basic food.However, the nutrient content of rice because kind, the place of production, growth conditions difference, there is very big difference, plus
Market on " inferior rice pretend to be high quality white rice ", the bad phenomenon such as " place of production pretend to be " continuously emerge it is therefore desirable to effective examine
Survey method carries out quality and place of production discriminating to the rice of market circulation.Traditional discrimination method mainly includes sense organ and chemistry inspection
Survey.Wherein organoleptic detection subjectivity is strong and wastes time and energy；Chemical detection needs professional that sample is carried out with the pre- place of loaded down with trivial details chemistry
Reason, the requirement of the process such as time-consuming, it is quick, green to meet, high-volume.Therefore, be badly in need of that research is a kind of quick, environmental protection, standard
True detection method is realizing the discriminating to rice.
LIBS (laser-induced breakdown spectroscopy, abbreviation LIBS) technology is
A kind of novel components analysis method.This technology can quick mass detection sample；Sampling spot size is little, destructive little；And can
Multiple element is analyzed simultaneously；The features such as detection under adverse circumstances can also be realized, has been widely used for industry, examines
The fields such as Gu, soil, water body and food inspection.
Research paper《Gas-chromatography combines Chemical Measurement and distinguishes Rice storage time and the place of production》(analysis test journal,
Volume 32,10 phases, in October, 2013 delivers) different storage time and different sources rice are analyzed respectively using gas chromatography
The volatile ingredient of sample, by PCA (principal components analysis, abbreviation PCA) with partially
Least square techniques of discriminant analysis (partial least square discriminant analysis, abbreviation PLS-DA) is to big
Rice sample is classified and discriminant analysis, obtains different storage time and different sources rice sample identification rate is respectively 96% He
100%.Research paper《Based on the research to geographical sign 5 constant virtues rice authentication technique for the Elements》(spectroscopy and spectrum
Analysis, volume 36,21 phases, in March, 2016 delivers) apply inorganic unit in inductively coupled plasma spectrometry and mass spectroscopy rice
The content of element, differentiates in conjunction with PCA, Fisher, artificial neural network is differentiated to 5 constant virtues rice and non-5 constant virtues rice, result
Show poor using PCA recognition effect, Fisher differentiates, the knowledge accuracy rate of artificial neural network respectively reaches 93.5% and
96.4%.The studies above method obtains higher precision in rice classification, but need rice is carried out SPME and
Wet is so that detection of complex.Chinese patent literature《A kind of method improving laser microprobe plastic identification precision and its dress
Put》(notification number is CN104730041A, and the day for announcing is on June 24th, 2015) is disclosed and a kind of is entered based on the LIBS technology of SVM
The method of row plastics classification.It passes through to increase the weight of 3 nonmetallic characteristic spectral lines, increases the matrix difference between different plastics
Property and discrimination is improved.Although SVM achieves good result in plastic identification, it is not directed to agricultural product
(the such as rice place of production) is applied.
From the point of view of above research, existing to rice classification detection technique mainly using chemical detection means, complicated, consumption
When it is impossible to meet the demand of commercial Application.
Content of the invention
Low for traditional rice place of production discriminating detection efficiency, detection process is complicated, and the present invention proposes a kind of induced with laser
The method that breakdown spectral technology identifies the cereal crops place of production, LIBS is combined by the method with SVM algorithm
Get up, to reaching the purpose quick and precisely identifying.
A kind of multiline combination laser induced breakdown spectroscopy cereal crops Production area recognition method that the present invention provides, its feature
It is, the method includes setting up svm classifier model and two processes of Production area recognition, specific as follows：
(1) set up svm classifier model process：
1.1. sample preparation：
The powder sample taking equal quality makes sheet sample in uniform thickness, and wherein, sample is the paddy of n different sources
Class crop；
1.2. the spectrum data gathering of sample：
LIBS spectrum data gathering, every kind of place of production cereal crops sample collection m are carried out to the cereal crops of n different sources
Individual plasma spectrometry, obtains N=n × m plasma spectrometry, wherein, the light spectrum amplitude of every kind of sample that m gathers for LIBS altogether
Number, m=1,2 ..., 1000；
1.3. selected characteristic spectral line
The plasma spectrometry of n different sources cereal crops of collection is analyzed, chooses the s in each Main elements
Individual the strongest characteristic spectral line, as Algorithm Analysis index, reads each characteristic spectral line intensity level, selects one of Main elements spy
Levy the intensity of spectral line and normalized is made to other s-1 Main elements characteristic spectrum intensity, wherein s-1 represents index to be analyzed
Number；
1.4. spectral line combination
Characteristic spectral line of the same race near s in selected Main elements the strongest characteristic spectral line is combined, final
Index to be analyzed to after q combination；
1.5.SVM disaggregated model is set up
The data of linearly inseparable is mapped in higher dimensional space by way of constructing kernel mappings function, thus in height
Dimension space is linearly distinguished, and described kernel mappings function adopts RBF：
By N group cereal crops spectroscopic data, selector packet spectrum, as training set, is used for setting up SVMs mould
Type, using remaining set spectrum as model measurement collection, for testing set up supporting vector machine model accuracy of identification；
(2) Production area recognition
First, according to the step of step 1.1 to 1.4 cereal crops to be identified are carried out sample preparation, spectrum data gathering,
Selected characteristic spectral line and carry out spectral line combination, obtains some groups of cereal crops spectroscopic datas；
Then, using the svm classifier model of the different sources cereal crops set up in step 1.5, the light to unknown sample
Modal data is classified, thus obtaining the prediction of the corresponding place of production.
The present invention adopts algorithm of support vector machine, in conjunction with the advantage of the quick species analysis of LIBS, different sources rice is entered
Row is quick, Accurate classification, and by the way of the combination of the larger and near one another identity element of signal-to-background ratio a plurality of characteristic spectral line
Strengthen the otherness between different sources cereal crops (as rice), thus improving the accuracy of identification of algorithm.Specifically, this
Bright method has the characteristics that and effect：
(1) the inventive method compared with traditional rice Production area recognition method it is not necessary to the chemical pre-treatment of complexity, this
Bright only simple compressing tablet sample preparation, carries out spectrum letter by LIBS elemental analysis technology to different sources rice
Number collection, decrease sample detection time and complicated chemical analysis processes, it is to avoid secondary pollution, improves the effect of detection
Rate.
(2) method that the inventive method passes through the larger and near one another identity element characteristic spectral line combination of signal-to-background ratio,
To increase the otherness between different sources rice, but not increase the number of situational variables, thus improving the classification of SVM classifier
Effect.
(3) the inventive method flexibility ratio is high, and the penalty parameter c of SVM slack variable of introduction and mapping kernel parameter g are permissible
Effectively improve the generalization ability of grader, and the training effectiveness of algorithm is high, parameter regulation is rapid.Can also be with other intelligence
Algorithm such as genetic algorithm (Genetic Algorithm, GA), random forest (Random Forest, RF), neutral net
Methods such as (Artificial Neural Network, ANN) combines, and improves the accuracy of classification further.
(4) it is important to note that the inventive method does not limit to the rice place of production, it is possible to achieve different sources not of the same race
The rice of class is identified.The Production area recognition of other cereal crops equally applicable.
The present invention is different from existing method, and by Main elements, nearby characteristic spectral line combines to improve svm classifier the present invention
The accuracy of model, has quick, green, the high advantage of recognition accuracy.
Brief description
The structural representation of the LIBS rice Production area recognition device that Fig. 1 provides for present example：
Wherein, 1. laser instrument；2. optical maser wavelength speculum；3. condenser lens；4. testing sample；5. signal pickup assembly；
6. optical fiber；7. spectrometer；8. firing line；9.ICCD；10. data wire；11. displacement platforms；12. computers.
Fig. 2 is the LIBS spectrogram in 200～900nm wave-length coverage for the Foshan gold globefish THAI Fragrant rice.
Fig. 3 is to combine, with a plurality of characteristic spectral line, the contrast classified using single features spectral line.
Specific embodiment：
With reference to specific embodiment, the specific embodiment of the present invention is further illustrated.Here it should be noted that
Explanation for these embodiments is used to help understand the present invention, but does not constitute limitation of the invention.Additionally, following institute
As long as in each embodiment of the present invention of description involved technical characteristic do not constitute conflict each other just can be mutual
Combination.
Due to the difference of rice growing environment, the component content directly resulting between the rice that different sources grow out is deposited
In difference, it is feasible for therefore according to the material composition in rice, different sources rice being made a distinction.Present example carries
For a kind of multiline method that combines algorithm of support vector machine auxiliary laser induced breakdown spectroscopy rice Production area recognition.
The inventive method includes setting up svm classifier model and two processes of Production area recognition, specifically includes following step：
(1) set up svm classifier model process：
1.1. sample preparation
First with pulverizer, known place of production rice is pulverized respectively, preparation uniformly powder sample, then claim
The powder sample taking equal quality makes sheet sample in uniform thickness using tablet press machine.
1.2. the spectrum data gathering of sample
In order to obtain spectral intensity and the high LIBS signal of signal to noise ratio, choose one of the n different sources rice place of production
Rice carry out collection time delay, laser energy and the defocusing amount experiment parameter of spectral signal and be optimized.With main in rice
The spectral intensity stability such as secondary element C, N, H and signal to noise ratio are judging quota, and spectral detection stability indicator passes through to calculate relatively
Standard deviation (Relative Standard Deviation, abbreviation RSD) determining, specific RSD=(standard deviation/calculating
The arithmetic mean of instantaneous value of result) * 100%, RSD value is the smaller the better；The specific calculating of signal to noise ratio is the spectral intensity values by collection
Subtracting background signal strength values, divided by noise figure, select signal to noise ratio maximum.By optimize Main elements spectrum-stable degree and
Maximum signal to noise ratio, comprehensive determination optimum experimental condition.Under optimal experiment condition, the rice of n different sources is carried out
LIBS spectrum data gathering, every kind of place of production rice sample gathers m plasma spectrometry, obtains N=n × m plasma altogether
Spectrum, wherein, n is the rice place of production number of collection；M is LIBS collection light spectrum amplitude number, m=1,2 ..., 1000.
1.3. selected characteristic spectral line
The plasma spectrometry of n different sources rice of collection is analyzed, chooses the strongest feature of s Main elements
Spectral line, as algorithm index to be analyzed, reads each characteristic spectral line intensity level with light Spectrum data processing software, selects s main amount unit
One of element characteristic spectral line intensity makees normalized to other s-1 Main elements characteristic spectrum intensity, as sample number
Represent analysis indexes number according to, wherein s, s=5,6 ..., 100, finally give s-1 non-combination line index to be analyzed.
1.4. spectral line combination
The selected neighbouring p characteristic spectral line of the same race of the strongest characteristic spectral line obtaining s Main elements is combined, after combination
S Main elements in a characteristic spectral line intensity not combining of selection to other s-1 Main elements characteristic spectrum intensity
Make normalized, obtain index to be analyzed (the index q number to be analyzed after combination and the analysis do not combined after q combination
Index s -1 number is identical).P value be by material composition and its physical characteristic determined it is however generally that, p=1,2 ..., 100.
1.5.SVM disaggregated model is set up
The mathematical procedure of training svm classifier model is as follows：If each example is (X in training seti,yi), i=1 ..., n,
Xi,yiRepresent i-th sample data and its corresponding label value, wherein X respectivelyi(xi1,xi2,…,xir) ∈ Rr, Rr represents training
Sample data set, r represents the number of property value；yi∈ { 1,2,3 ... }, 1,2,3 ... for every group of property value label value.Algorithm
Important feature is to be mapped in higher dimensional space the data of linearly inseparable by way of constructing kernel function, thus in higher-dimension
Space is linearly distinguished.The present invention selects RBF (RadialBasis Function, RBF) as kernel mappings letter
Number：
Thus obtained non-linear SVM classifier equation is：
Constraints:
Wherein,Represent the sample data of forecast set；Represent the mapping function of forecast set sample；Represent
Two norm distances；L represents the number of training set；αiFor Lagrange multiplier；B is the bias factor in equation；C is punishment ginseng
Number, g is mapping kernel parameter.
C be can be seen that by equation, g parameter is set up to model and had a major impact.The general range of choice that c, g are set：C=2e, g
=2w, e, w ∈ { -10, -9 ..., 9,10 }.
Present example carries out data processing using SVMs software, and N group rice spectroscopic data is selected in the middle part of it
Packet spectrum as training set, be used for setting up supporting vector machine model, remaining set spectrum as model measurement collection, for testing
Set up supporting vector machine model accuracy of identification.
SVMs software can be from as the SVMs software tool box (A of Lin Zhiren et al. exploitation
Library for Support Vector Machines—LIBSVM).
(2) Production area recognition
First, according to the step of step 1.1 to 1.4, sample preparation, spectrum data gathering, selection are carried out to rice to be identified
Characteristic spectral line and carry out spectral line combination, obtains some groups of rice spectroscopic datas；
Then, using the svm classifier model of the different sources rice set up in step 1.5, the spectrum number to unknown sample
According to being classified, thus obtaining the prediction of the corresponding place of production.
Example：
Embodiment 1
1. sample preparation.
The implementation case chooses the rice sample (shown in table 1) of 10 kinds of different sources, and sample specific name and the place of production are as follows：
Golden globefish THAI Fragrant rice (Foshan), Hubei Guan Miaoshan special product organic rice (Hubei Zhijiang), Bama of Guangxi orthodox school glutinous rice (Guangxi
Bar horse), disk mansion Feng Jin rice (Liaoning Panjin), 5 constant virtues rice fragrance of a flower rice (Heilungkiang 5 constant virtues), (Siping is double for northeast brown rice
The Liao Dynasty), cross people Yu Gong (Anqing), Hunan pond rice (Fauna of Taoyuan, Nw Hunan), Wan Niangong rice (Jiangxi Shangrao) and Chongming Island rice (Jiangsu
Taizhou).Place of production classification is carried out to them using the inventive method.Process for simplifying the analysis, is carried out to different sources rice point
Do not number, label value is followed successively by：1,2,3,4,5,6,7,8,9,10；The corresponding place of production is with place name abbreviation Guangdong (GD), Hubei
(HB), Guangxi (GX), Liaoning (LN), Heilungkiang (HLJ), Jilin (JL), Anhui (AH), Hunan (HN), Jiangxi (JX), Jiangsu
(JS).In view of the physical characteristic of sample, the irregularly meeting of such as sample to a certain degree affect spectral signal, gather letter in LIBS
Before number, sample is pre-processed, using pulverizer, different sources rice is pulverized respectively, preparation uniformly powder-like
Product, then weigh 15g about powder sample adopt tablet press machine to make sheet sample in uniform thickness with 25MPa pressure to be used for
LIBS spectral detection.
2. the spectrum data gathering of rice sample.
Spectra collection is carried out under air ambient, and experimental provision is as shown in Figure 1.Using Q-switch Nd：YAG pulse laser
Device 1 (Quantel Brilliant B, wavelength 532nm, pulse width 8ns, maximum repetition rate 10Hz) as excitation source,
The plasma resonance light inspiring collected by collection head 5 and transmit to spectrometer 7 (Andor Technology,
Mechelle5000, wave-length coverage 200～975nm, resolution lambda/△ λ=5000) carry out light splitting, ICCD9 (Andor
Technology, iStar DH-334T, 1024 × 1024 pixels) spectral signal that spectrometer is transmitted through coming carries out opto-electronic conversion.
Sample surfaces are controlled to do the motion of " bow " shape in X direction with Y-direction using electricity driving displacement platform 11 during spectra collection.For preventing
Air breakdown, laser focusing lens 3 (focal length 15cm) focus is located at below sample surfaces 1.27mm.For obtaining optimal spectrum
Intensity and spectrum signal-to-background ratio, pulsed laser energy is set to 40mJ, and ICCD time delay and gate-width are set to 1.5 μ s and 3 μ s.More than
Under technological parameter, the spectrum of the plasma of 10 kinds of rice of collection, every kind of 100 spectrum of sample collection, therefore 10 kinds rice are altogether
1000 spectrum of collection.For example, Fig. 2 is the LIBS spectrum in 200～900nm wave-length coverage for the Foshan gold globefish THAI Fragrant rice
Figure.
3. selected characteristic spectral line.
Due to containing abundant carbohydrate and mineral matter element in rice, therefore select 13 features of Main elements
Spectral line C-N (0,0) 388.34nm, C I 247.86nm, N I 746.83nm, O I 777.19nm, H I 656.29nm, C-C
(0,0) 516.52nm, Mg II 279.55nm, Mn I 403.08nm, Ca I 422.67nm, Si I 288.16nm, Al I
394.40nm, Na I 588.95nm and K I 766.49nm, specific characteristic spectral line is as shown in table 2.Will be special in every group of data
Levy the intensity of spectral line value to normalize divided by Ca I 422.67nm the intensity of spectral line value, finally give 1000 groups of spectroscopic datas, every group of light
Modal data comprises 12 situational variables.
4. characteristic spectral line combination
A plurality of spy by characteristic spectral line the strongest for selected every kind of element and after identity element intensity near one another normalization
Levy spectral line (C-N (0,0) 388.34nm+C-N (1,1) 387.14nm+C-N (2,2) 386.19nm+C-N (3,3) 385.47nm+C-
N (4,4) 385.09nm, N I 746.83nm+N I 744.23nm+N I 742.36nm, O I 777.19nm+O I
777.42nm, Mg II 279.55nm+Mg II 280.27nm+Mg I 285.21nm, Mn I 403.08nm+Mn I
403.31nm+Mn I 403.45nm, Al I394.40nm+Al I 396.15nm, Na I588.95nm+Na I589.59nm and
K I 766.49nm+K I 769.90nm) combination, increase the otherness between different sources rice, be similarly obtained 1000 groups of light
Modal data, every group of spectroscopic data comprises 12 situational variables.From the figure 3, it may be seen that 7,8,9 different sources spectral line normalized intensities than
Discrimination very little at C-N (0,0) 388.34nm characteristic spectral line, by C-N (0,0) 388.34nm, C-N (1,1) 387.14nm, C-
The combination of N (2,2) 386.19nm, C-N (3,3) 385.47nm and C-N (4,4) 385.09nm characteristic spectral line is effectively increased the place of production
Between otherness.The concrete difference value of different sources rice is as shown in table 3, Δ in table12Refer to the place of production 1 and the place of production 2 characteristic spectrum
Difference after line strength normalization.
5.SVM disaggregated model is set up.
Carry out mathematical modeling, the kernel of employing using LIBSVM tool box under the software environment of Matlab2010b version
Function is RBF, and the classification of suitable nonlinear data has higher stability.G parameter in kernel RBF and
It is the principal element of impact SVM algorithm performance on the penalty factor c parameter of slack variable, the central principle of algorithm is also to find out
Good mapping kernel parameter g and penalty factor c.In order to training result has more convincingness, prevent from modeling study, tested using interaction
Demonstration is optimizing factor of influence.C, g parameter of single features spectral line is respectively 222.8609 and 0.125；Assemblage characteristic spectral line
C, g parameter is respectively 588.1336 and 0.047366.The final training set knowledge obtaining single features spectral line with SVM training pattern
Rate is not 91.2%, and forecast set discrimination is 90.8%, and the training set discrimination of multiline combination is 94.2%, and forecast set identifies
Rate is 94.6%, obtains higher discrimination, specific different sources recognition result and the SVM identification knot combining with spectral line
Fruit contrasts as shown in table 4, and wherein the discrimination of Hubei Zhijiang rice brings up to 88% by 64%.Therefore pass through multiple features spectral line group
Conjunction can improve the degree of accuracy of LIBS rice Production area recognition.
6. other weighted value control methods contrast
Contrast spectral line combined method and Chinese patent literature《A kind of improve laser microprobe plastic identification precision method and its
Device》(notification number is CN104730041A, and the day for announcing is on June 24th, 2015) discloses a kind of spectral line weights adjustment and combines
The LIBS technology of SVM, by increasing nonmetallic Main elements C, N, H weighted value, result is as shown in table 4, using spectral line combination side
Formula is to the discrimination 90.8% that the discrimination 94.6% of rice is higher than spectral line weights adjustment.
Embodiment 2
1. sample preparation.
The implementation case chooses the rice sample (as table 5) of same province 6 kinds of different sources, sample specific product title and
The place of production is as follows：Seven riverheads (Heilungkiang Suihua), 5 constant virtues rice fragrance of a flower rice (Heilungkiang 5 constant virtues), baby's diatery supplement (Heilungkiang Ning'an lotus
Hua Cun), extraction woods glutinous rice (Heilungkiang Qiqihar), northeast Xiangshui County rice (Heilungkiang village of Ning'an Xiangshui County) and vegetarian diet cat Old Taylor rice
(Heilungkiang Qiqihar).Place of production classification is carried out to them using the inventive method.Process for simplifying the analysis, to different sources
Rice is numbered respectively, and label value is followed successively by：1,2,3,4,5,6；The corresponding place of production is with place name abbreviation Suihua (SH), 5 constant virtues
(WC), lotus flower (LH), Qiqihar 1 (QQHE1), Xiangshui County (XS), Qiqihar 2 (QQHE2).Also with pulverizer to difference
Place of production rice is pulverized respectively, preparation uniformly powder sample, then weigh 15g about powder sample adopt tablet press machine
Sheet sample in uniform thickness is made for LIBS spectral detection with 25MPa pressure.
2. the spectrum data gathering of rice sample.
Under identical experiment condition, the spectrum of the plasma of 6 kinds of rice of collection, every kind of 100 spectrum of sample collection,
Therefore 6 kinds rice gather 600 spectrum altogether.
3. selected characteristic spectral line.
Same 13 characteristic spectral line C-N (0,0) 388.34nm selecting Main elements, C I 247.86nm, N I
746.83nm, O I 777.19nm, H I 656.29nm, C-C (0,0) 516.52nm, Mg II 279.55nm, Mn I
403.08nm, Ca I 422.67nm, Si I 288.16nm, Al I 394.40nm, Na I 588.95nm and K I
766.49nm.Characteristic spectral line intensity level in every group of data is normalized divided by Ca I 422.67nm the intensity of spectral line value, final
To 600 groups of spectroscopic datas, every group of spectroscopic data comprises 12 situational variables.
4. characteristic spectral line combination
Same by characteristic spectral line the strongest for selected every kind of element and many after identity element intensity normalization near one another
Bar characteristic spectral line (C-N (0,0) 388.34nm+C-N (1,1) 387.14nm+C-N (2,2) 386.19nm+C-N (3,3)
385.47nm+C-N (4,4) 385.09nm, N I 746.83nm+N I 744.23nm+N I 742.36nm, O I 777.19nm
+ O I 777.42nm, Mg II 279.55nm+Mg II 280.27nm+Mg I 285.21nm, Mn I 403.08nm+Mn I
403.31nm+Mn I 403.45nm, Al I394.40nm+Al I 396.15nm, Na I588.95nm+Na I589.59nm and
KI766.49nm+KI 769.90nm) combination, increase the otherness between different sources rice, be similarly obtained 600 groups of spectrum numbers
According to every group of spectroscopic data comprises 12 situational variables.
5.SVM disaggregated model is set up.
Carry out mathematical modeling using LIBSVM tool box under the software environment of Matlab2010b version.Single features are composed
C, g parameter of line is respectively 12.1257 and 2.2974；C, g parameter of assemblage characteristic spectral line is respectively 64 and 0.43528.Finally
It is 93.67% with the training set discrimination that SVM training pattern obtains single features spectral line, forecast set discrimination is 88.67%,
The training set discrimination of multiline combination is 94.33%, and forecast set discrimination is 93%, obtains higher discrimination, specifically
Different sources recognition result and the SVM recognition result contrast combined with spectral line as shown in table 6.Therefore pass through multiple features spectral line
Combination can improve the degree of accuracy of LIBS rice Production area recognition.
6. other weighted value control methods contrast
The LIBS technical result that contrast spectral line combined method combines SVM with spectral line weights adjustment is as shown in table 4, using spectral line
Combination is to the discrimination 92.33% that the discrimination 93% of rice is higher than spectral line weights adjustment.
Embodiment 3
1. sample preparation.
The implementation case chooses the rice sample (as table 7) of the 20 kinds of different sources in 10 different provinces, sample specific product
Title and the place of production are as follows：Golden globefish THAI Fragrant rice (Foshan), Hubei Guan Miaoshan special product organic rice (Hubei Zhijiang), seven riverheads
(Heilungkiang Suihua), the soft rice of beautiful shrimp Wang Xiang (Dongguan, Guangdong), world grain people (Chaoyang), Bama of Guangxi orthodox school glutinous rice (Guangxi
Bar horse), disk mansion Feng Jin rice (Liaoning Panjin), the little sago of five cereals (Guangzhou Guangdong), rice fragrance of a flower rice (Heilungkiang five, 5 constant virtues
Often), (Siping is double for baby's diatery supplement (Heilungkiang Ning'an lotus flower village), extraction woods glutinous rice (Heilungkiang Qiqihar), northeast brown rice
The Liao Dynasty), cross people Yu Gong (Anqing), northeast Xiangshui County rice (Heilungkiang village of Ning'an Xiangshui County) Hunan pond rice (Fauna of Taoyuan, Nw Hunan), Wan Nian
Tribute rice (Jiangxi Shangrao), Chongming Island rice (Jiangsu Taizhou), a river autumn selenium-rich rice (Jiangxi Ji'an), Zhuxi tribute rice (Hubei ten
Weir Zhuxi) and vegetarian diet cat Old Taylor rice (Heilungkiang Qiqihar).Place of production classification is carried out to them using the inventive method.In order to
Simplified analysis process, is numbered respectively to different sources rice, label value is followed successively by：1,2,3,4,5,6,7,8,9,10,11,
12,13,14,15,16,17,18,19,20；The corresponding place of production is with place name abbreviation Foshan (FS), Zhijiang (ZJ), Suihua (SH), east
Tabernaemontanus bulrush (DG), Chaoyang (CY), rivers and ponds (HC), Panjin (PJ), Guangzhou (GZ), 5 constant virtues (WC), lotus flower village (LHC), Qiqihar
(QQHE), Shuangliao (SL), Anqing (AQ), village of Xiangshui County (XSC), the Land of Peach Blossoms (TY), Shangrao (SR), Taizhou (TZ), Ji'an (JA), Zhuxi
And Heilungkiang (HLJ) (ZX).Also with pulverizer, different sources rice is pulverized respectively, preparation uniformly powder-like
Product, then weigh 15g about powder sample adopt tablet press machine to make sheet sample in uniform thickness with 25MPa pressure to be used for
LIBS spectral detection.
2. the spectrum data gathering of rice sample.
Under identical experiment condition, the spectrum of the plasma of 20 kinds of rice of collection, every kind of 100 light of sample collection
Spectrum, therefore 20 kinds rice gather 2000 spectrum altogether.
3. selected characteristic spectral line.
Same 13 characteristic spectral line C-N (0,0) 388.34nm selecting Main elements, C I 247.86nm, N I
746.83nm, O I 777.19nm, H I 656.29nm, C-C (0,0) 516.52nm, Mg II 279.55nm, Mn I
403.08nm, Ca I 422.67nm, Si I 288.16nm, Al I 394.40nm, Na I 588.95nm and K I
766.49nm.Characteristic spectral line intensity level in every group of data is normalized divided by Ca I 422.67nm the intensity of spectral line value, final
To 2000 groups of spectroscopic datas, every group of spectroscopic data comprises 12 situational variables.
4. characteristic spectral line combination
Same by characteristic spectral line the strongest for selected every kind of element and many after identity element intensity normalization near one another
Bar characteristic spectral line (C-N (0,0) 388.34nm+C-N (1,1) 387.14nm+C-N (2,2) 386.19nm+C-N (3,3)
385.47nm+C-N (4,4) 385.09nm, N I 746.83nm+N I 744.23nm+N I 742.36nm, O I 777.19nm
+ O I 777.42nm, Mg II 279.55nm+Mg II 280.27nm+Mg I 285.21nm, Mn I 403.08nm+Mn I
403.31nm+Mn I 403.45nm, Al I394.40nm+Al I 396.15nm, Na I588.95nm+Na I589.59nm and
K I 766.49nm+K I 769.90nm) combination, increase the otherness between different sources rice, be similarly obtained 2000 groups of light
Modal data, every group of spectroscopic data comprises 12 situational variables.
5.SVM disaggregated model is set up.
Carry out mathematical modeling using LIBSVM tool box under the software environment of Matlab2010b version.Single features are composed
C, g parameter of line is respectively 222.8609 and 0.37893；C, g parameter of assemblage characteristic spectral line is respectively 388.0234 and
0.071794.The final training set discrimination obtaining single features spectral line with SVM training pattern is 86.7%, and forecast set identifies
Rate is 86.2%, and the training set discrimination of multiline combination is 90.1%, and forecast set discrimination is 89.4%, obtains higher
Discrimination, specific different sources recognition result and the SVM recognition result contrast combined with spectral line are as shown in table 8.Therefore logical
Excessive characteristic spectral line combination can improve the degree of accuracy of LIBS rice Production area recognition.6. other weighted values are adjusted
Method contrasts
The LIBS technical result that contrast spectral line combined method combines SVM with spectral line weights adjustment is as shown in table 4, using spectral line
Combination is to the discrimination 84.8% that the discrimination 89.4% of rice is higher than spectral line weights adjustment.
General principle and principal character and the advantages of the present invention of the present invention have been shown and described above.The section of the industry
Grind personnel it should be appreciated that the present invention is not restricted to the described embodiments, the simply explanation described in above-described embodiment and specification
The principle of the present invention, without departing from the spirit and scope of the present invention, the present invention also has various changes and modifications, these
Changes and improvements both fall within scope of the claimed invention.Claimed scope by appending claims and
Its equivalent thereof.
Table 1 is different sources rice inventory
Table 2 by adopted the combining of single features spectral line and a plurality of characteristic spectral line
Table 3 is using single features spectral line and different sources rice characteristic spectral line intensity under a plurality of characteristic spectral line combination
Concrete difference value after normalization
Table 4 is the result of different sources rice svm classifier
Table 5 is same province different sources rice inventory
Table 6 is the result of same identity different sources rice svm classifier
Table 7 is different provinces different sources rice inventory
Table 8 is the result of different provinces different sources rice svm classifier
Claims (6)
1. a kind of multiline combination laser induced breakdown spectroscopy cereal crops Production area recognition method is it is characterised in that the method bag
Include svm classifier model and two processes of Production area recognition set up, specific as follows：
(1) set up svm classifier model process：
1.1. sample preparation：
The powder sample taking equal quality makes sheet sample in uniform thickness, and wherein, sample is that the cereal of n different sources is made
Thing；
1.2. the spectrum data gathering of sample：
LIBS spectrum data gathering is carried out to the cereal crops of n different sources, every kind of place of production cereal crops sample collection m etc.
Gas ions spectrum, obtains N=n × m plasma spectrometry altogether, wherein, the light spectrum amplitude number of every kind of sample that m gathers for LIBS, m
=1,2 ..., 1000；
1.3. selected characteristic spectral line
The plasma spectrometry of n different sources cereal crops of collection is analyzed, the s choosing in each Main elements is individual
Strong characteristic spectral line, as Algorithm Analysis index, reads each characteristic spectral line intensity level, selects one of Main elements characteristic spectrum
Line strength makees normalized to other s-1 Main elements characteristic spectrum intensity, and wherein s-1 represents index number to be analyzed；
1.4. spectral line combination
Characteristic spectral line of the same race near s in selected Main elements the strongest characteristic spectral line is combined, finally gives q
Index to be analyzed after combination；
1.5.SVM disaggregated model is set up
The data of linearly inseparable is mapped in higher dimensional space by way of constructing kernel mappings function, thus empty in higher-dimension
Between linearly distinguished, described kernel mappings function adopts RBF：
By N group cereal crops spectroscopic data, selector packet spectrum, as training set, is used for setting up supporting vector machine model, will
Remaining set spectrum as model measurement collection, for testing set up supporting vector machine model accuracy of identification；
(2) Production area recognition
First, according to the step of step 1.1 to 1.4, sample preparation, spectrum data gathering, selection are carried out to cereal crops to be identified
Characteristic spectral line and carry out spectral line combination, obtains some groups of cereal crops spectroscopic datas；
Then, using the svm classifier model of the different sources cereal crops set up in step 1.5, the spectrum number to unknown sample
According to being classified, thus obtaining the prediction of the corresponding place of production.
2. multiline combination laser induced breakdown spectroscopy cereal crops Production area recognition method according to claim 1, it is special
Levy and be, described cereal crops are rice.
3. multiline combination laser induced breakdown spectroscopy cereal crops Production area recognition method according to claim 1 and 2, its
It is characterised by, LIBS spectrum data gathering process described in described step 1.2 is：First choose in n different sources cereal crops
A place of production cereal crops carry out spectral signal collection time delay, laser energy and defocusing amount experiment parameter carry out excellent
Change, to obtain spectral intensity and the high LIBS signal of signal to noise ratio；Then stable with the spectral intensity of Main elements in cereal crops
Property and signal to noise ratio be judging quota, by optimizing the spectrum-stable degree of Main elements and maximum signal to noise ratio, comprehensive determine most preferably real
Test condition；Under optimal experiment condition, LIBS spectrum data gathering is carried out to the cereal crops of n different sources.
4. multiline combination laser induced breakdown spectroscopy cereal crops Production area recognition method according to claim 3, it is special
Levy and be, described spectral detection stability indicator is determined by calculating relative standard deviation RSD；Signal to noise ratio specifically calculates
By the spectral intensity values subtracting background signal strength values of collection divided by noise figure, select signal to noise ratio maximum.
5. multiline combination laser induced breakdown spectroscopy cereal crops Production area recognition method according to claim 1 and 2, its
It is characterised by, described step 1.5, if each example is (X in training seti,yi), i=1 ..., n, Xi,yiRepresent i-th respectively
Sample data and its corresponding label value, wherein Xi(xi1,xi2,…,xir) ∈ Rr, Rr represents training sample data collection, and r represents
The number of property value；yi∈ { 1,2,3 ... }, 1,2,3 ... for every group of property value label value；
Described selection RBF is as kernel mappings function：
Thus obtained non-linear SVM classifier equation is：
Constraints:
Wherein,Represent the sample data of forecast set；Represent the mapping function of forecast set sample；Represent two models
Number distance；L represents the number of training set；αiFor Lagrange multiplier；B is the bias factor in equation；C is punishment parameter；G is
Mapping kernel parameter.
6. multiline combination laser induced breakdown spectroscopy cereal crops Production area recognition method according to claim 5, it is special
Levy and be, the range of choice of c, g：C=2e, g=2w, e, w ∈ { -10, -9 ..., 9,10 }.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610802039.4A CN106404748B (en) | 2016-09-05 | 2016-09-05 | A kind of multiline combination laser induced breakdown spectroscopy cereal crops Production area recognition method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610802039.4A CN106404748B (en) | 2016-09-05 | 2016-09-05 | A kind of multiline combination laser induced breakdown spectroscopy cereal crops Production area recognition method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106404748A true CN106404748A (en) | 2017-02-15 |
| CN106404748B CN106404748B (en) | 2019-03-05 |
Family
ID=57998336
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610802039.4A Active CN106404748B (en) | 2016-09-05 | 2016-09-05 | A kind of multiline combination laser induced breakdown spectroscopy cereal crops Production area recognition method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106404748B (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107525797A (en) * | 2017-07-27 | 2017-12-29 | 上海交通大学 | A kind of LIBS analysis methods of micron dimension powdered rubber trace element |
| CN107677647A (en) * | 2017-09-25 | 2018-02-09 | 重庆邮电大学 | Chinese medicine place of production discrimination method based on principal component analysis and BP neural network |
| CN108780048A (en) * | 2018-05-24 | 2018-11-09 | 深圳达闼科技控股有限公司 | A kind of method, detection device and the readable storage medium storing program for executing of determining detection device |
| CN109142251A (en) * | 2018-09-17 | 2019-01-04 | 平顶山学院 | The LIBS quantitative analysis method of random forest indirect labor's neural network |
| CN109916991A (en) * | 2019-04-09 | 2019-06-21 | 新疆大学 | A method of based on metallic element combination PLS-DA Model checking hop varieties and the place of production |
| CN111401444A (en) * | 2020-03-16 | 2020-07-10 | 深圳海关食品检验检疫技术中心 | Method and device for predicting origin of red wine, computer equipment and storage medium |
| CN111770705A (en) * | 2017-12-28 | 2020-10-13 | 数眠公司 | Bed with presence detection feature |
| CN111965167A (en) * | 2020-08-20 | 2020-11-20 | 天津大学 | Method and device for predicting element composition and calorific value of solid waste |
| CN112051256A (en) * | 2020-07-22 | 2020-12-08 | 中国地质大学（武汉） | CNN model-based LIBS (laser induced breakdown spectroscopy) measurement method and system for content of element to be measured |
| CN112188637A (en) * | 2018-03-23 | 2021-01-05 | Oppo广东移动通信有限公司 | Wireless communication method, user equipment and network equipment |
| CN112782151A (en) * | 2021-02-22 | 2021-05-11 | 湖北工程学院 | Data processing method for improving classification accuracy of laser-induced breakdown spectroscopy |
| CN114579635A (en) * | 2022-03-04 | 2022-06-03 | 北京三月雨文化传播有限责任公司 | Big data information analysis processing system based on cloud computing |
| TWI825490B (en) * | 2017-05-04 | 2023-12-11 | 美商菲爾薇解析公司 | Devices and methods for endpoint detection in manufacturing process by near infrared spectroscopy and machine learning techniques, and related computer-readable medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103488874A (en) * | 2013-09-01 | 2014-01-01 | 西北大学 | Improved support vector machine-LIBS (laser-induced breakdown spectroscopy) combined sorting method for steel materials |
| CN104697965A (en) * | 2015-03-10 | 2015-06-10 | 西北大学 | Method for recognizing slag variety by combining with laser-induced breakdown spectroscopy based on least squares support vector machine |
| CN104730041A (en) * | 2013-12-20 | 2015-06-24 | 武汉新瑞达激光工程有限责任公司 | Method and apparatus for improving plastic identification precision of laser probe |
| CN104964950A (en) * | 2015-06-10 | 2015-10-07 | 长江大学 | Multi-element wave peak-based laser-induced breakdown spectroscopy rock fragment type identification method |
| CN105181678A (en) * | 2015-09-07 | 2015-12-23 | 长江大学 | Identification method of rice varieties based on laser-induced breakdown spectroscopy (LIBS) |
| CN105806827A (en) * | 2016-03-11 | 2016-07-27 | 华中科技大学 | Method for identifying plastics by virtue of laser probe based on non-metallic element |
-
2016
- 2016-09-05 CN CN201610802039.4A patent/CN106404748B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103488874A (en) * | 2013-09-01 | 2014-01-01 | 西北大学 | Improved support vector machine-LIBS (laser-induced breakdown spectroscopy) combined sorting method for steel materials |
| CN104730041A (en) * | 2013-12-20 | 2015-06-24 | 武汉新瑞达激光工程有限责任公司 | Method and apparatus for improving plastic identification precision of laser probe |
| CN104697965A (en) * | 2015-03-10 | 2015-06-10 | 西北大学 | Method for recognizing slag variety by combining with laser-induced breakdown spectroscopy based on least squares support vector machine |
| CN104964950A (en) * | 2015-06-10 | 2015-10-07 | 长江大学 | Multi-element wave peak-based laser-induced breakdown spectroscopy rock fragment type identification method |
| CN105181678A (en) * | 2015-09-07 | 2015-12-23 | 长江大学 | Identification method of rice varieties based on laser-induced breakdown spectroscopy (LIBS) |
| CN105806827A (en) * | 2016-03-11 | 2016-07-27 | 华中科技大学 | Method for identifying plastics by virtue of laser probe based on non-metallic element |
Non-Patent Citations (8)
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI825490B (en) * | 2017-05-04 | 2023-12-11 | 美商菲爾薇解析公司 | Devices and methods for endpoint detection in manufacturing process by near infrared spectroscopy and machine learning techniques, and related computer-readable medium |
| CN107525797A (en) * | 2017-07-27 | 2017-12-29 | 上海交通大学 | A kind of LIBS analysis methods of micron dimension powdered rubber trace element |
| CN107677647A (en) * | 2017-09-25 | 2018-02-09 | 重庆邮电大学 | Chinese medicine place of production discrimination method based on principal component analysis and BP neural network |
| CN111770705A (en) * | 2017-12-28 | 2020-10-13 | 数眠公司 | Bed with presence detection feature |
| CN112188637A (en) * | 2018-03-23 | 2021-01-05 | Oppo广东移动通信有限公司 | Wireless communication method, user equipment and network equipment |
| US11895642B2 (en) | 2018-03-23 | 2024-02-06 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Wireless communication method for sidelink resource scheduling, user equipment, and network device |
| CN112188637B (en) * | 2018-03-23 | 2022-03-29 | Oppo广东移动通信有限公司 | Wireless communication method, user equipment and network equipment |
| CN108780048A (en) * | 2018-05-24 | 2018-11-09 | 深圳达闼科技控股有限公司 | A kind of method, detection device and the readable storage medium storing program for executing of determining detection device |
| CN108780048B (en) * | 2018-05-24 | 2020-07-07 | 深圳达闼科技控股有限公司 | Method for determining detection equipment, detection device and readable storage medium |
| CN109142251A (en) * | 2018-09-17 | 2019-01-04 | 平顶山学院 | The LIBS quantitative analysis method of random forest indirect labor's neural network |
| CN109916991A (en) * | 2019-04-09 | 2019-06-21 | 新疆大学 | A method of based on metallic element combination PLS-DA Model checking hop varieties and the place of production |
| CN111401444B (en) * | 2020-03-16 | 2023-11-03 | 深圳海关食品检验检疫技术中心 | Method and device for predicting red wine origin, computer equipment and storage medium |
| CN111401444A (en) * | 2020-03-16 | 2020-07-10 | 深圳海关食品检验检疫技术中心 | Method and device for predicting origin of red wine, computer equipment and storage medium |
| CN112051256A (en) * | 2020-07-22 | 2020-12-08 | 中国地质大学（武汉） | CNN model-based LIBS (laser induced breakdown spectroscopy) measurement method and system for content of element to be measured |
| CN111965167A (en) * | 2020-08-20 | 2020-11-20 | 天津大学 | Method and device for predicting element composition and calorific value of solid waste |
| CN112782151A (en) * | 2021-02-22 | 2021-05-11 | 湖北工程学院 | Data processing method for improving classification accuracy of laser-induced breakdown spectroscopy |
| CN112782151B (en) * | 2021-02-22 | 2023-01-13 | 湖北工程学院 | Data processing method for improving classification accuracy of laser-induced breakdown spectroscopy |
| CN114579635A (en) * | 2022-03-04 | 2022-06-03 | 北京三月雨文化传播有限责任公司 | Big data information analysis processing system based on cloud computing |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106404748B (en) | 2019-03-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106404748B (en) | A kind of multiline combination laser induced breakdown spectroscopy cereal crops Production area recognition method | |
| CN106770194B (en) | Cereal crops place of production discrimination method based on wavelet transformation laser induced breakdown spectroscopy | |
| CN102072885B (en) | Machine vision-based paddy neck blast infection degree grading method | |
| CN104374752A (en) | Rapid detection method for nutrient elements of crops based on collinear laser-induced breakdown spectroscopy | |
| CN110398466A (en) | Crop growth state monitoring method based on remote-sensing inversion | |
| CN105445229A (en) | Nitrogen balance spectroscopy-based wheat spring nitrogenous fertilizer application method, and construction method of nitrogen topdressing amount model thereof | |
| CN109916838A (en) | A kind of detection method of the rice seed germination ability based on high light spectrum image-forming and artificial neural network | |
| CN107421911A (en) | A kind of preprocess method of the soil nitrogen detection based on portable near infrared spectrometer | |
| CN102519886A (en) | Method for detecting contents of chlorophyll a and carotinoid in crop laminas | |
| CN109001147A (en) | A method of fresh tea leaves geography information is differentiated using near infrared spectrum | |
| CN108181262A (en) | A kind of method using Near Infrared Spectroscopy for Rapid Sargassum horneri content of cellulose | |
| CN102507504A (en) | Hyperspectral monitoring method for pine wilt disease of masson pines | |
| CN106290292A (en) | A kind of utilize the method for carotenoid content in copolymerization Jiao's microscopic Raman detection Folium Camelliae sinensis | |
| Brambilla et al. | Application of a low-cost RGB sensor to detect basil (Ocimum basilicum L.) nutritional status at pilot scale level | |
| CN110398473A (en) | A kind of rapid test paper detection method and system | |
| CN110378925A (en) | A kind of ecological water estimation method of reserve of airborne LiDAR and multispectral romote sensing technology | |
| CN109520965A (en) | A method of lysine content is detected based near infrared spectrum characteristic extractive technique | |
| CN111751347A (en) | Barley leaf pigment imaging method under powdery mildew stress based on Raman spectrum | |
| Chu et al. | Determination of total flavonoid and polysaccharide content in Anoectochilus formosanus in response to different light qualities using hyperspectral imaging | |
| Luo et al. | Effect of Cu stress on minerals in rice by analyzing husk based on laser-induced breakdown spectroscopy | |
| CN108007917A (en) | Hilbert method establishes nitrogen content raman spectroscopy measurement model method in rice strain | |
| CN110188955A (en) | A kind of large scale autumn grain crops yield evaluation method based on chlorophyll fluorescence remote sensing | |
| CN106770016A (en) | The Protein quantitative analysis NIR transmitted spectrum measuring methods of single seed paddy seed | |
| Chang et al. | Red-edge ratio normalized vegetation index for remote estimation of green biomass | |
| CN110132865A (en) | Based on SAE-LSSVR crops cadmium content Vis-NIR spectral depth characteristic model method for building up |
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 | ||
| CB03 | Change of inventor or designer information |
Inventor after: Li Xiangyou Inventor after: Yang Ping Inventor after: Guo Lianbo Inventor after: Zhu Yining Inventor after: Zeng Xiaoyan Inventor after: Li Jiaming Inventor after: Yang Xinyan Inventor after: Tang Yun Inventor before: Li Xiangyou Inventor before: Yang Ping Inventor before: Guo Lianbo Inventor before: Zhu Yining Inventor before: Zeng Xiaoyan Inventor before: Lu Yongfeng Inventor before: Li Jiaming Inventor before: Yang Xinyan Inventor before: Tang Yun |
|
| CB03 | Change of inventor or designer information |