CN105651727B - The method that near-infrared spectrum analysis based on JADE and ELM differentiates apple shelf life - Google Patents
The method that near-infrared spectrum analysis based on JADE and ELM differentiates apple shelf life Download PDFInfo
- Publication number
- CN105651727B CN105651727B CN201511005467.6A CN201511005467A CN105651727B CN 105651727 B CN105651727 B CN 105651727B CN 201511005467 A CN201511005467 A CN 201511005467A CN 105651727 B CN105651727 B CN 105651727B
- Authority
- CN
- China
- Prior art keywords
- shelf life
- learning machine
- sample
- samples
- near infrared
- 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.)
- Expired - Fee Related
Links
- 238000004458 analytical method Methods 0.000 title claims abstract description 34
- 238000000034 method Methods 0.000 title claims abstract description 29
- 238000002329 infrared spectrum Methods 0.000 title claims abstract description 26
- 239000010977 jade Substances 0.000 title claims abstract description 19
- 238000001228 spectrum Methods 0.000 claims abstract description 29
- 238000013441 quality evaluation Methods 0.000 claims abstract description 3
- 210000002569 neuron Anatomy 0.000 claims description 17
- 238000012549 training Methods 0.000 claims description 15
- 230000006870 function Effects 0.000 claims description 12
- 230000005284 excitation Effects 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 abstract description 10
- 239000000523 sample Substances 0.000 description 29
- 239000000203 mixture Substances 0.000 description 5
- 238000004611 spectroscopical analysis Methods 0.000 description 5
- 230000006835 compression Effects 0.000 description 4
- 238000007906 compression Methods 0.000 description 4
- 235000013399 edible fruits Nutrition 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000000354 decomposition reaction Methods 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 230000003595 spectral effect Effects 0.000 description 3
- 238000003860 storage Methods 0.000 description 3
- 239000000126 substance Substances 0.000 description 3
- 238000002790 cross-validation Methods 0.000 description 2
- 230000006378 damage Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 235000021022 fresh fruits Nutrition 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 239000005416 organic matter Substances 0.000 description 2
- 238000000926 separation method Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 102100020895 Ammonium transporter Rh type A Human genes 0.000 description 1
- 101100301844 Arabidopsis thaliana RH50 gene Proteins 0.000 description 1
- 102000004190 Enzymes Human genes 0.000 description 1
- 108090000790 Enzymes Proteins 0.000 description 1
- VGGSQFUCUMXWEO-UHFFFAOYSA-N Ethene Chemical compound C=C VGGSQFUCUMXWEO-UHFFFAOYSA-N 0.000 description 1
- 239000005977 Ethylene Substances 0.000 description 1
- 238000004497 NIR spectroscopy Methods 0.000 description 1
- 101150107345 Rhag gene Proteins 0.000 description 1
- 235000010724 Wisteria floribunda Nutrition 0.000 description 1
- 241000607479 Yersinia pestis Species 0.000 description 1
- 238000010521 absorption reaction Methods 0.000 description 1
- TZCXTZWJZNENPQ-UHFFFAOYSA-L barium sulfate Inorganic materials [Ba+2].[O-]S([O-])(=O)=O TZCXTZWJZNENPQ-UHFFFAOYSA-L 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 239000003153 chemical reaction reagent Substances 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000012850 discrimination method Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 235000013305 food Nutrition 0.000 description 1
- 238000003306 harvesting Methods 0.000 description 1
- 238000012880 independent component analysis Methods 0.000 description 1
- 239000004615 ingredient Substances 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 239000002932 luster Substances 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000000813 microbial effect Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 210000004218 nerve net Anatomy 0.000 description 1
- 238000009659 non-destructive testing Methods 0.000 description 1
- 238000012628 principal component regression Methods 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 239000013074 reference sample Substances 0.000 description 1
- 230000029058 respiratory gaseous exchange Effects 0.000 description 1
- 239000004575 stone Substances 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
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
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/359—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light using near infrared light
-
- 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
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N2021/3595—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light using FTIR
Landscapes
- Physics & Mathematics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses a kind of methods that near-infrared spectrum analysis based on JADE and ELM differentiates apple shelf life, include the following steps：(1) sample is collected, acquires sample spectra, sample near-infrared is obtained and diffuses modal data, and original near infrared spectrum data is compressed using wavelet transform；(2) compressed spectrum data using eigenmatrix joint approximate diagonalization algorithm are decomposed, obtains the mixed battle array of isolated component matrix reconciliation；(3) operating limit learning machine method, using the mixed battle array of solution as mode input, corresponding shelf life establishes extreme learning machine analysis model as output；(4) quality evaluation of model measures the shelf life of sample to be identified.The present invention can quickly differentiate apple shelf life, enrich chemometrics method, have a good application prospect.
Description
Technical field
The present invention relates to Infrared Non-destructive Testing technical field more particularly to a kind of feature based matrix joint approximate diagonalizations
(JADE) and the near-infrared spectrum analysis of extreme learning machine (ELM) differentiate apple shelf life method.
Background technology
Apple is the main fruit that northern China is abounded with, and the activity of internal certain enzymes becomes strong after apple harvesting, fruit breathing
Rate and Ethylene Production Rate become larger, and saprogenicity microbial reproduction is accelerated, and quality decline is caused finally to be rotted and loses business valency
Value.Therefore, strengthen the monitoring of apple shelf life in storage, transport, sales section, advantageously reduce the loss of apple fresh fruit simultaneously
Ensure edible quality of the fresh fruit during shelf, increase economic efficiency.
Modern near infrared spectroscopic method is a kind of detection method of quick nondestructive, and principle is to hydrogeneous in organic matter
The frequency multiplication sum of fundamental frequencies of group X-H generates absorption, and the physical and chemical index of organic matter is measured by stoechiometric process, is calculated with effective mathematics
By physical and chemical index and establishment of spectrum functional relation, it has been widely used in the quantitative analysis of agricultural product method.Presently, there are
The fruit shelf life discrimination method based near infrared spectroscopy, sample radix needed for modeling is big, the training time is long, a large amount of samples
The acquisition of physics and chemistry value takes time and effort.To this situation, it is badly in need of a kind of stronger model of universality, the effective shelf for differentiating apple
Phase.
Since near infrared spectrum can be regarded as the linear combination of a variety of main component spectrum, in recent years, there is scholar by " blind source
Problem is introduced into near-infrared spectral analytical method for separation (BBS) ", it is intended to by the spectrum of these main components from the mixed light of complexity
It is separated in spectrum.JADE algorithms are the algebraically independent component analysis methods of a kind of numerical stability, strong robustness, suitable for spectrum
Decomposition.Meanwhile modeled with reference to extreme learning machine, ELM is a kind of new algorithm for single hidden layer feedforward neural network, the calculation
Method randomly generates the connection weight of input layer and hidden layer and the threshold value of hidden layer neuron, and in the training process without adjusting
It is whole, only hidden layer neuron number need to be set, unique optimal solution can be obtained, there is pace of learning is fast, generalization ability is good etc.
Advantage.
Present invention employs the methods that JADE and extreme learning machine classification are combined, and allow near infrared spectroscopic method effective
Discriminating apple shelf life, enrich stoechiometric process and carried for crops shelf life discriminating in near-infrared spectrum analysis field
Theoretical premise and technical support are supplied.
Invention content
In view of the above-mentioned deficiencies in the prior art, it is an object of the present invention to provide a kind of near infrared spectrum based on JADE and ELM
The method of analysis and identification apple shelf life.
The present invention is achieved by the following technical solutions：A kind of near-infrared spectrum analysis based on JADE and ELM differentiates
The method of apple shelf life, includes the following steps：
(1) sample is collected, including shelf life is n days samples, shelf life is m days samples and sample to be identified, wherein, m and n
Respectively less than 30, and n is more than or equal to 7 with m differences；Acquire three kinds of samples near infrared spectrum, and to these three near infrared spectrums into
The processing of row wavelet transform, obtains three kinds of compressed near infrared spectrum data matrixes；
(2) three kinds of compressed near infrared spectrum data matrixes that step (1) obtains are subjected to eigenmatrix joint approximation
Diagonalization (JADE) is decomposed, and obtains the mixed battle array of three kinds of solutions；
(3) operating limit learning machine (ELM) method establishes initial threshold learning machine analysis model, is n days samples by shelf life
Product and shelf life are the practical shelf life of m days samples and are n days samples by the shelf life that step (2) obtains and shelf life is m
Mixed mode input of the battle array as initial threshold learning machine analysis model of solution of its sample, and then obtain the analysis of optimal limit learning machine
Model；
4) solution of sample to be identified is mixed into battle array input optimal limit learning machine analysis model, obtains the shelf of sample to be identified
Phase.
Further, the step (1) is implemented as follows：
Three kinds of samples are scanned using near infrared spectrometer, the near infrared spectrum of three kinds of samples is obtained, is calculated using K-S
Method presses 3:1 quantity ratio is n days samples by shelf life and shelf life is that m days samples are randomly divided into training set sample and forecast set sample
Product, wherein, calibration set sample is used for model training, and forecast set sample is used for the quality evaluation of model；By the near red of three kinds of samples
External spectrum carries out wavelet transform processing, and wavelet transform selects wavelet mother function, and for dbn, n is vanishing moment, and vanishing moment takes
2,3 layers are decomposed to, obtains three kinds of compressed near infrared spectrum data matrixes.
Further, the step (3) is implemented as follows：By shelf life be n days samples and shelf life is m days samples
Practical shelf life and their solution mix battle array as extreme learning machine analysis model, and operating limit learning machine method establishes initial pole
Limit learning machine analysis model；During optimal limit learning machine analysis model is obtained, using " Sigmoidal " as the limit
General hidden layer excitation function in habit machine analysis model, by hidden layer neuron number initializing set be 5, and with 5 for step-length according to
Secondary to increase to 50, repetition training 20 times, obtains the model of optimal limit learning machine analysis model under each hidden neuron value
Parameter, so as to obtain optimal limit learning machine analysis model.
The beneficial effects of the invention are as follows：Using JADE algorithms, choose optimal isolated component number and establish calibration model；Using
ELM algorithms establish model, need arrange parameter few, are easy and fast to train, and improve model accuracy, choose optimal models.With existing skill
Art is compared, such as multiple linear regression, principal component regression, and the isolated component obtained through JADE algorithms is more nearly actual spectrum,
Institute's established model has more practical significance.Entire measurement process does not consume chemical reagent, and test is quick, uses manpower and material resources sparingly, and batch is surveyed
Result is accurate during examination, greatly improves detection efficiency.The links such as this method can be stored in fruit quotient, transport, sale are promoted the use of.
Description of the drawings
Fig. 1 is shelf life method flow diagram of the present invention；
Fig. 2 is two different sample primary light spectrograms of shelf life of the present invention.
Specific embodiment
The present invention provides the near infrared spectroscopic methods that a kind of apple shelf life accurately differentiates.Below in conjunction with the accompanying drawings 1, it is attached
The present invention is further described for Fig. 2 and embodiment.Embodiment is illustrated for the present invention, and not the invention is limited.
The technical scheme is that collecting sample and adopting spectrum, original spectrum is pre-processed, uses wavelet transform first
Near infrared spectrum data is effectively compressed, obtains the moderate matrix of data volume, which using JADE algorithms is decomposed, is obtained
Mixed battle array is conciliate to isolated component matrix, using the mixed battle array of solution as mode input, is modeled by extreme learning machine algorithm.Entire scheme stream
Journey figure is as shown in Figure 1.
The original spectrum directly acquired by near infrared spectrometer, data volume is huge, and repeatability is high, redundancy weight, and by
Noise jamming.Using small wave converting method, compressed spectrum data, and remove spectral noise, remain spectrum main information and
Substantially reduce data volume, in practice it has proved that, for probability density, wavelet coefficient after wavelet transformation is than the superelevation of original signal
This property is stronger, is more suitable for blind source separating, this step need to select rational wavelet function and the wavelet decomposition number of plies.
JADE algorithms are used to decompose data after wavelet compression.JADE algorithms are one kind of blind source separation algorithm, it is therefore an objective to will
Independent element in mixed signal is separated, and this method robustness is good, can usually obtain relatively stable source estimated result.With
When spectrum data matrix is decomposed, independent element and the mixed battle array of corresponding solution can be obtained.Every a line of independent element matrix is suitable
Pure material information in a kind of spectral information of statistical iteration ingredient and blend sample, corresponding solution mix battle array reflection
Substance proportion in initial data, i.e. contribution of the isolated component to entire sample near infrared spectrum.
Extreme learning machine algorithm is a kind of new type nerve net to be grown up by single hidden layer feedforward neural network (SLFN)
Network algorithm, overcoming needs constantly to calculate in traditional artificial neural network training process, adjusts connection weight between each layer, is hidden
The shortcomings that neuron number containing layer and hidden layer neuron threshold value.The hidden layer neuron number of extreme learning machine network needs
It is determined before training, the biasing of weights and hidden layer neuron between input layer and hidden layer need to only be set at random, and be instructed
Practice process without adjustment.Therefore hidden layer neuron number is first determined before network training, it, can be by hidden layer neuron number in implementation
Mesh is initialized as 5, and increases to training set sample number for step-length with 5, to determine best hidden layer neuron number.This hair
Using Sigmoidal functions as the excitation function of ELM network hidden layer neurons in bright.
Embodiment
1. sample collection and spectra collection
It is red fuji apple for examination apple, place of production Shandong Province is suitble to fresh food, but shelf life is short, is easily waited in storage, transport
It is damaged in journey.Median size is selected, color and luster is close and the apple 80 of no disease and pests harm and mechanical damage, the picking same day transport
Toward laboratory (25 DEG C, RH50%-60%) storages at room temperature.Due to apple from picking later at room temperature, need one balance
Process, composition transfer is larger, stores two days, carries out first time spectra collection.After the completion of adopting spectrum, sample is continued to store, seven days
Afterwards, second of spectra collection is carried out.As shown in Figure 2, spectrum of the dotted line for the acquisition for the first time of certain a sample in figure, solid line are it
The spectrum of second of acquisition, it can be seen from the figure that there is differences for the sample near infrared spectrum of different shelf life.It acquires twice
160 spectroscopic datas are obtained, including two kinds of shelf life classifications, all data Kennard/Stone (K/S) algorithm is pressed 3:
1 quantity ratio is divided into training set sample and forecast set sample, and K/S algorithms are to calculate sample according to sample room spectrum Euclidean distance
Difference, it can obtain most representative training set sample accordingly.Spectra collection instrument is Thermo Nicolet companies of the U.S.
The Nexus870 type Fourier transform near infrared instrument of production, band are useful for irreflexive Smart near-IR attachmentes.Compose area's acquisition
Range：800nm-1850nm, experiment carry out at room temperature.12 different locations are carried out along equator to each sample in experiment
Scanning, uses BaSO4Change piece as reference sample, take its averaged spectrum.To avoid the interference of stray light, used when acquiring spectrum
1.5mm gasket shadings.Model foundation software is carried out based on Matlab2013a.
2. spectroscopic data is handled
To compress near infrared spectrum data, wavelet transform, wavelet basis function choosing are carried out to collected original spectrum
Db2 is selected, decomposes to 3 layers, data account for about the 12% of former data volume after compression, and data contain the big portion of initial data after compression
Divide information.
Matrix after compressing is decomposed with JADE algorithms.Setting isolated component number is needed in JADE algorithms, which represents to be decomposed
The number of independent element in matrix is related to the validity of JADE algorithms and modeling speed and precision.By the first of isolated component number
Initial value is set as 5, to find best isolated component number, isolated component number is increased to 10 successively, and analyzed for step-length with 1,
It is 7 to obtain isolated component optimal value.According to isolated component value, matrix has obtained isolated component after decomposing compression using JADE algorithms
Matrix and corresponding solution mix battle array.
3. establish extreme learning machine analysis model
In extreme learning machine modeling analysis, the selection of One hidden layer neuron number is most important to the Classification and Identification of model.
In the present embodiment, " Sigmoidal ", " Sine " and " Hardlim " function are first chosen respectively and is encouraged as ELM models hidden layer
Function.Through testing it is found that when general hidden layer excitation function is selected as " Sigmoidal " function, Model checking performance is relatively stable, and
With higher discrimination precision.To obtain optimum model parameter, hidden layer neuron number is set as 5, is increased successively for step-length with 5
60 are added to, the repetition training 10 times under each neuron value, most ELM hidden nodes are determined as 30 at last.With through JADE points
The solution obtained after solution mixes battle array as mode input, and model cross-validation investigates the generalization ability of extreme learning machine model, mould
Type cross validation accuracy rate is 98.75%.40 spectrum of forecast set sample are imported into model, to its forecasting shelf life, experience
Card, prediction result have fabulous linear relationship with actual conditions, and predictablity rate is up to 97.5%, so as to demonstrate extreme learning machine
The correctness of analysis model.
4. the shelf life of sample to be identified differentiates
After the spectra collection of sample to be identified, gradually pre-processed using wavelet transform, JADE decomposition methods, it is defeated
Enter into extreme learning machine analysis model, export sample shelf life to be identified.
Claims (1)
1. a kind of method that near-infrared spectrum analysis based on JADE and ELM differentiates apple shelf life, which is characterized in that including with
Lower step：
（1）Sample is collected, including shelf life is n days samples, shelf life is m days samples and sample to be identified, wherein, m and n are small
In 30, and n is more than or equal to 7 with m differences；Acquire three kinds of samples near infrared spectrum, and these three near infrared spectrums are carried out from
Wavelet transform process is dissipated, obtains three kinds of compressed near infrared spectrum data matrixes；
（2）By step（1）Three kinds of obtained compressed near infrared spectrum data matrixes carry out eigenmatrix joint approximate diagonal
Change（JADE）It decomposes, obtains the mixed battle array of three kinds of solutions；
（3）Operating limit learning machine（ELM）Method establishes initial threshold learning machine analysis model, by shelf life for n days samples with
Shelf life is for the practical shelf life of m days samples and by step（2）Obtained shelf life is n days samples and shelf life is m days samples
Mixed mode input of the battle array as initial threshold learning machine analysis model of solution of product, and then obtain optimal limit learning machine analysis mould
Type；
4）The solution of sample to be identified is mixed into battle array input optimal limit learning machine analysis model, obtains the shelf life of sample to be identified；
The step（1）It is implemented as follows：
Three kinds of samples are scanned using near infrared spectrometer, the near infrared spectrum of three kinds of samples is obtained, is pressed using K-S algorithms
3:1 quantity ratio is n days samples by shelf life and shelf life is that m days samples are randomly divided into training set sample and forecast set sample,
Wherein, calibration set sample is used for model training, and forecast set sample is used for the quality evaluation of model；By the near infrared light of three kinds of samples
Spectrum carries out wavelet transform processing, and wavelet transform selects wavelet mother function, and for dbn, n is vanishing moment, and vanishing moment takes 2,
3 layers are decomposed to, obtains three kinds of compressed near infrared spectrum data matrixes；
The step（3）It is implemented as follows：The practical shelf life that by shelf life be n days samples and shelf life is m days samples with
And their solution mixes battle array as extreme learning machine analysis model, operating limit learning machine method establishes the analysis of initial threshold learning machine
Model；During optimal limit learning machine analysis model is obtained, using " Sigmoidal " as extreme learning machine analysis model
In general hidden layer excitation function, be 5 by hidden layer neuron number initializing set, and successively increased with 5 for step-length to 50,
Repetition training 20 times, obtains the model parameter of optimal limit learning machine analysis model under each hidden neuron value, so as to obtain
Optimal limit learning machine analysis model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201511005467.6A CN105651727B (en) | 2015-12-28 | 2015-12-28 | The method that near-infrared spectrum analysis based on JADE and ELM differentiates apple shelf life |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201511005467.6A CN105651727B (en) | 2015-12-28 | 2015-12-28 | The method that near-infrared spectrum analysis based on JADE and ELM differentiates apple shelf life |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105651727A CN105651727A (en) | 2016-06-08 |
| CN105651727B true CN105651727B (en) | 2018-06-12 |
Family
ID=56478230
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201511005467.6A Expired - Fee Related CN105651727B (en) | 2015-12-28 | 2015-12-28 | The method that near-infrared spectrum analysis based on JADE and ELM differentiates apple shelf life |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105651727B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108507954A (en) * | 2018-03-20 | 2018-09-07 | 南京财经大学 | A kind of minitype fruits quality and shelf life on-line measuring device and method |
| CN113447084A (en) * | 2021-08-31 | 2021-09-28 | 季华实验室 | Detection device, system, method and storage medium for estimating shelf life of food |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101520412A (en) * | 2009-03-23 | 2009-09-02 | 中国计量学院 | Near infrared spectrum analyzing method based on isolated component analysis and genetic neural network |
| CN103018181A (en) * | 2012-12-14 | 2013-04-03 | 江苏大学 | Soft measurement method based on correlation analysis and ELM neural network |
| CN104374738A (en) * | 2014-10-30 | 2015-02-25 | 中国科学院半导体研究所 | Qualitative analysis method for improving identification result on basis of near-infrared mode |
| CN105092509A (en) * | 2015-08-20 | 2015-11-25 | 东北大学 | Sample component measurement method based on PCR-ELM algorithm |
| CN105158200A (en) * | 2015-07-16 | 2015-12-16 | 天津工业大学 | Modeling method capable of improving accuracy of qualitative near-infrared spectroscopic analysis |
-
2015
- 2015-12-28 CN CN201511005467.6A patent/CN105651727B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101520412A (en) * | 2009-03-23 | 2009-09-02 | 中国计量学院 | Near infrared spectrum analyzing method based on isolated component analysis and genetic neural network |
| CN103018181A (en) * | 2012-12-14 | 2013-04-03 | 江苏大学 | Soft measurement method based on correlation analysis and ELM neural network |
| CN104374738A (en) * | 2014-10-30 | 2015-02-25 | 中国科学院半导体研究所 | Qualitative analysis method for improving identification result on basis of near-infrared mode |
| CN105158200A (en) * | 2015-07-16 | 2015-12-16 | 天津工业大学 | Modeling method capable of improving accuracy of qualitative near-infrared spectroscopic analysis |
| CN105092509A (en) * | 2015-08-20 | 2015-11-25 | 东北大学 | Sample component measurement method based on PCR-ELM algorithm |
Non-Patent Citations (5)
| Title |
|---|
| Application of independent components analysis with the JADE algorithm and NIR hyperspectral imaging for revealing food adulteration;Puneet Mishra 等;《Journal of Food Engineering》;20150710;第168卷;第7-15页 * |
| 基于ELM和可见／近红外光谱的鲜枣动态分类检测;杨一 等;《光谱学与光谱分析》;20150731;第35卷(第7期);第1870-1874页 * |
| 独立成分分析结合近红外光谱分析用于多组分同时测定;张世芝 等;《计算机与应用化学》;20101228;第27卷(第12期);第1684-1686页 * |
| 独立组分分析在红外光谱分析中的应用;毕贤 等;《高等学校化学学报》;20040630;第25卷(第6期);第1023-1027页 * |
| 近红外光谱结合极限学习机识别贮藏期的损伤猕猴桃;郭文川 等;《光学精密工程》;20131030;第21卷(第10期);第2722页第1栏4行至第18行，第2725页第1栏第1行至第2栏第11行 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105651727A (en) | 2016-06-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Li et al. | Advances in non-destructive early assessment of fruit ripeness towards defining optimal time of harvest and yield prediction—A review | |
| Jiang et al. | Comparison of algorithms for wavelength variables selection from near-infrared (NIR) spectra for quantitative monitoring of yeast (Saccharomyces cerevisiae) cultivations | |
| Zhang et al. | Rapid determination of leaf water content using VIS/NIR spectroscopy analysis with wavelength selection | |
| Rizzo et al. | Fruit ripeness classification: A survey | |
| CN106596416A (en) | Chilled fresh meat quality non-destructive testing method based on hyperspectral imaging technology | |
| CN107121407B (en) | The method that near-infrared spectrum analysis based on PSO-RICAELM identifies Cuiguan pear maturity | |
| CN101915744A (en) | Near infrared spectrum nondestructive testing method and device for material component content | |
| Zhang et al. | Detection of canopy chlorophyll content of corn based on continuous wavelet transform analysis | |
| Pang et al. | Hyperspectral imaging coupled with multivariate methods for seed vitality estimation and forecast for Quercus variabilis | |
| CN106092990A (en) | A kind of three-dimensional fluorescence spectrum discrimination method of lycium barbarum | |
| CN107655851A (en) | A kind of method based on near-infrared spectrum technique quick detection lysine content | |
| CN108007881A (en) | A kind of aquaculture water quality total nitrogen content detection method based on spectral technique | |
| CN110082310A (en) | A kind of near infrared band EO-1 hyperion diagnostic method of rubber tree LTN content | |
| CN110363119A (en) | Based on wavelet transformation-random forests algorithm tobacco mildew method for quickly identifying | |
| Scalisi et al. | Maturity prediction in yellow peach (Prunus persica L.) cultivars using a fluorescence spectrometer | |
| CN110569605A (en) | Non-glutinous rice leaf nitrogen content inversion model method based on NSGA2-ELM | |
| CN110398473A (en) | A kind of rapid test paper detection method and system | |
| CN105651727B (en) | The method that near-infrared spectrum analysis based on JADE and ELM differentiates apple shelf life | |
| CN102135496A (en) | Infrared spectrum quantitative analysis method and infrared spectrum quantitative analysis device based on multi-scale regression | |
| Kang et al. | Nondestructive testing model of mango dry matter based on fluorescence hyperspectral imaging technology | |
| CN107247033B (en) | Identify the method for Huanghua Pear maturity based on rapid decay formula life cycle algorithm and PLSDA | |
| Liu et al. | Rapid determination of rice protein content using near-infrared spectroscopy coupled with feature wavelength selection | |
| Zhang et al. | Hyperspectral model based on genetic algorithm and SA-1DCNN for predicting Chinese cabbage chlorophyll content | |
| Zhang et al. | Hyperspectral imaging combined with CNN for maize variety identification | |
| Zhou et al. | Research on hyperspectral regression method of soluble solids in green plum based on one-dimensional deep convolution network |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180612 |