CN111007040A - Near infrared spectrum rapid evaluation method for rice taste quality - Google Patents
Near infrared spectrum rapid evaluation method for rice taste quality Download PDFInfo
- Publication number
- CN111007040A CN111007040A CN201911382675.6A CN201911382675A CN111007040A CN 111007040 A CN111007040 A CN 111007040A CN 201911382675 A CN201911382675 A CN 201911382675A CN 111007040 A CN111007040 A CN 111007040A
- Authority
- CN
- China
- Prior art keywords
- rice
- support vector
- vector machine
- near infrared
- characteristic wavelength
- 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
Images
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
- G01N21/47—Scattering, i.e. diffuse reflection
- G01N21/4738—Diffuse reflection, e.g. also for testing fluids, fibrous materials
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- 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
- G01N21/47—Scattering, i.e. diffuse reflection
- G01N21/4738—Diffuse reflection, e.g. also for testing fluids, fibrous materials
- G01N2021/4764—Special kinds of physical applications
- G01N2021/4769—Fluid samples, e.g. slurries, granulates; Compressible powdery of fibrous samples
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2201/00—Features of devices classified in G01N21/00
- G01N2201/12—Circuits of general importance; Signal processing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Theoretical Computer Science (AREA)
- Biophysics (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Evolutionary Biology (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Computational Biology (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Analytical Chemistry (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Chemical & Material Sciences (AREA)
- Biochemistry (AREA)
- Physiology (AREA)
- Genetics & Genomics (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention relates to a near infrared spectrum rapid evaluation method of rice taste quality, which is a method for food detection by using a near infrared spectrum, and the method comprises the steps of processing a collected sample, preprocessing the spectrum by using the near infrared diffuse reflection spectrum as an original spectrum, preliminarily optimizing the characteristic wavelength of the near infrared spectrum based on a competitive adaptive reweighting sampling algorithm, and synchronously optimizing the parameters of a support vector machine based on a quantum genetic simulation annealing algorithm and the characteristic wavelength of the near infrared spectrum to obtain the optimal characteristic wavelength; establishing a regression model by using the optimal characteristic wavelength, evaluating the precision of the regression model and quickly establishing an evaluation model; and performing spectral scanning on polished rice needing to be measured, and inputting spectral data into an evaluation model according to characteristic wavelengths to finish quality evaluation. The invention uses the quantum genetic simulated annealing algorithm to synchronously optimize the parameters of the support vector machine and the characteristic wavelength of the near infrared spectrum, thereby effectively improving the detection precision and efficiency of the regression model of the support vector machine for the taste quality of rice.
Description
Technical Field
The invention relates to a method for detecting food by utilizing near infrared spectrum, in particular to a near infrared spectrum rapid evaluation method for rice taste quality.
Background
With the improvement of living standard, people put higher requirements on the edible safety and the taste quality of rice, and the taste quality becomes one of the most important indexes for improving the rice quality. At present, the taste quality of rice is mainly evaluated by a sensory evaluation method, a physicochemical index evaluation method and an instrument analysis evaluation method. The sensory evaluation method is the most intuitive evaluation performed by directly evaluating rice cooked into cooked rice by professionals, and is easily influenced by subjective preferences of the evaluating staff and has low repeatability. The physical and chemical index evaluation method is used for evaluating the taste quality of rice based on the correlation relationship between the physical and chemical indexes of the rice and the taste quality, but the physical and chemical indexes have high detection cost and large workload, and the detection result is easy to be greatly output due to the system error in the analysis process. The instrument analysis and evaluation method mainly adopts two instruments, namely a rice grain taste meter and a rice taste meter to evaluate the taste quality of rice. The rice grain taste meter firstly measures the amylose, protein, water and fatty acid content of rice based on a near infrared spectrum technology, and then inputs the result into a regression model between the pre-established physicochemical index and the taste quality to realize the evaluation of the taste quality of the rice.
Heilongjiang, as the main production area of high-quality japonica rice in China, has become the national grain-safe ballast stone. In order to cultivate and screen new varieties of rice with excellent taste quality, it is necessary to develop a method for rapidly detecting the taste quality of high-quality northeast rice in Heilongjiang province. Aiming at the current situation that the Heilongjiang high-quality rice lacks special detection equipment for taste quality, based on the application feasibility of a near infrared spectrum analysis technology in the field of rapid detection of rice moisture, amylose, protein, fat, gel consistency and alkali elimination value and the obvious correlation between the indexes and the rice taste quality, the Heilongjiang rice taste quality direct rapid evaluation method based on the near infrared spectrum analysis technology is provided, the trade order of the Heilongjiang rice is effectively maintained, the benefits of rice farmers are protected, the income of farmers is improved, and the method plays an important role in promoting the structure adjustment of the Heilongjiang agricultural industry and promoting the development of the high-end rice industry.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a near infrared spectrum rapid evaluation method for rice taste quality.
The near infrared spectrum rapid evaluation method for the taste quality of rice is realized by the following steps:
(1) sample collection and processing
Collecting japonica rice samples of different producing areas, collecting no less than 30 samples in each producing area, collecting no less than 150 total samples, selecting, hulling, and milling the collected samples to obtain rice polished rice samples, sealing and bagging for later use;
(2) near infrared spectral collection
The near-infrared diffuse reflection spectrum of the rice grains is collected by adopting a German Bruker TANGO near-infrared spectrometer, and the spectrum collection range is 3946-11542 cm-1Resolution of 8.0cm-1Scanning a sample for 32 times, wherein the sample loading mode is scanning by a sample cup rotating table of 50mm, the sample loading height is about 30mm, and the average value of 3-5 times of scanning is taken as the original spectrum of the sample;
(3) taste quality assessment
Evaluating the taste quality of the collected and prepared rice sample by adopting a national standard method, and grading by adopting a percent system mode;
(4) spectral preprocessing and sample set partitioning
Removing abnormal samples by adopting a residual mean-variance distribution diagram based on Monte Carlo cross validation, selecting more than 120 samples from a data set with the abnormal samples removed as a correction set and a validation set by adopting a random selection mode, dividing the correction set and the validation set according to a ratio of 3:1 by adopting a Kernard-Stone method, and taking the rest samples as independent test sets; preprocessing the spectrum data of the correction set by adopting a Savitzky-Golay convolution smoothing method, a multivariate scattering correction method, a standard regular transformation method, a derivative processing method, a wavelet transformation method and a combination method thereof, and selecting a spectrum preprocessing method based on the minimum cross validation root-mean-square error of a partial least square regression model of the correction set;
(5) preliminary near infrared spectrum characteristic wavelength optimization based on competitive adaptive re-weighting sampling algorithm
Executing a plurality of rounds of a competitive adaptive reweighting sampling algorithm on the correction set data obtained in the step (4), and selecting a repeatedly selected wavelength point corresponding to the correction set when the root mean square error is minimum in cross validation as a primary optimal result of the characteristic wavelength obtained by the competitive adaptive reweighting sampling algorithm;
(6) support vector machine parameter and near infrared spectrum characteristic wavelength synchronous optimization based on quantum genetic simulated annealing algorithm
Synchronously optimizing parameters of the support vector machine and the characteristic wavelength of the near infrared spectrum obtained in the step (5) by adopting a quantum genetic simulated annealing algorithm, wherein the quantum genetic simulated annealing algorithm adopts multi-quantum bit coding, and the punishment parameter, the kernel function parameter, the insensitive loss function parameter and the characteristic wavelength coding structure of the support vector machine are as follows if the number of the initially selected characteristic wavelength points of the near infrared spectrum of the competitive adaptive re-weighted sampling algorithm is k:
wherein (α)mn,βmn) Is the probability amplitude of the quantum state, and | αmn|2+|βmn|21,2,3,4, 1,2, …, k, and (α) when initializing the populationmn,βmn) Is composed of
In the decoding process of the quantum genetic simulated annealing algorithm, a large amount of the quantum genetic simulated annealing algorithm is usedSub-bit encoding quantum probability collapsed into binary sequence a1a2…akb1b2…bkc1c2…ckd1d2…dkBinary coding corresponding to the k-bit penalty parameter, kernel function parameter, insensitive loss function parameter and characteristic wavelength; the punishment parameter, the kernel function parameter and the insensitive loss function parameter are decoded by adopting binary real numbers, and the characteristic wavelength coding directly determines whether the corresponding wavelength point participates in the operation according to the values of binary bits '1' and '0';
the quantum genetic simulated annealing algorithm adopts the cross validation of a regression model of a support vector machine to obtain a target function f (x) and a fitness function fit (x) ═ exp [ - (f (x)) fmin)/t]Wherein f isminIs the minimum objective function value of the current generation population, and t is a temperature parameter; the evolution process of the quantum genetic simulated annealing algorithm comprises gambling wheel selection with an optimal retention strategy, quantum crossing, quantum variation, quantum revolving gate disturbance solution construction and Metropolis selection replication evolution;
when synchronous optimization of support vector machine parameters and near infrared spectrum characteristic wavelengths is carried out, encoding and population initialization are carried out, after quantum probability collapse is carried out on chromosomes in a population, decoding operation is carried out to obtain punishment parameters, kernel function parameters and insensitive loss function parameter values; inputting the spectral data of the correction set into a regression model of a support vector machine according to the characteristic wavelength coding value, establishing the regression model according to the support vector machine parameters obtained by decoding, calculating a target function and a fitness function, and performing genetic evolution operation according to the fitness function value to obtain a next generation population; after the cooling operation is executed, the iterative processes of decoding, calculating a fitness function, genetic evolution and generating a new population are continuously executed until an algorithm termination condition is met, and the optimal parameter combination and the optimal characteristic wavelength of the support vector machine are obtained;
(7) evaluation model building
Inputting the spectral data of the correction set into a support vector machine according to the optimal characteristic wavelength, training a regression model of the support vector machine by using the optimal parameters of the optimized support vector machine, establishing a regression model of the near infrared spectrum support vector machine for the taste quality of rice, and evaluating the precision of the regression model by adopting a verification set and an independent test set; when the detection precision of the established support vector machine regression model meets the requirement, outputting a corresponding model to complete establishment of a rice taste quality rapid evaluation model;
(8) evaluation of taste quality of Rice
And performing near-infrared diffuse reflection spectrum scanning on the polished rice needing to measure the taste quality, and inputting the preprocessed spectrum data into an evaluation model according to characteristic wavelengths to finish the rapid evaluation of the taste quality.
The invention relates to a near infrared spectrum rapid evaluation method for rice taste quality, which adopts a support vector machine to establish a nonlinear regression model between rice near infrared spectrum data and the taste quality of the rice so as to establish a near infrared spectrum rapid evaluation method for high-quality rice taste quality. Aiming at the dual requirements of the support vector machine parameter optimization and the near infrared spectrum characteristic wavelength optimization, a quantum genetic simulation annealing algorithm is applied to synchronously optimize the support vector machine parameters and the near infrared spectrum characteristic wavelength, so that the optimal parameters of the support vector machine and the optimal characteristic wavelength variable suitable for modeling of the support vector machine are obtained, and the detection precision and efficiency of the rice taste quality support vector machine regression model are effectively improved.
Drawings
FIG. 1 is a schematic flow chart of a near infrared spectrum rapid evaluation method for rice taste quality;
FIG. 2 is a schematic diagram of a synchronous optimization process of support vector machine parameters and near infrared spectrum characteristic wavelengths based on a quantum genetic simulated annealing algorithm.
Detailed Description
The near infrared spectrum rapid evaluation method for the taste quality of rice is realized by the following steps
(1) Sample collection and processing
The method comprises the steps of collecting five typical high-quality rice production places of Zhenhang, Wuchang, Fangzheng, Xiangshui and Jiansanjiang province japonica rice samples, collecting 30 samples in each production place, and collecting 150 rice samples in total. After rice selection, hulling, rice hulling and milling are finished uniformly, the rice polished rice sample is prepared and sealed and bagged for later use.
(2) Near infrared spectral collection
The near-infrared diffuse reflection spectrum of the rice grains is collected by adopting a German Bruker TANGO near-infrared spectrometer, and the spectrum collection range is 3946-11542 cm-1Resolution of 8.0cm-1The sample is scanned 32 times, the sample loading mode is that a sample cup rotating platform is scanned 50mm, the sample loading height is about 30mm, and the average value of 3 times of scanning is taken as the original spectrum of the sample.
(3) Taste quality assessment
The taste quality of the collected and prepared rice sample is evaluated according to GB/T15682-2008 'grain and oil detection, sensory evaluation method of rice and rice cooking edible quality', and the rice sample is scored in a percentage mode.
(4) Spectral preprocessing and sample set partitioning
Removing abnormal samples by adopting a residual mean-variance distribution diagram based on Monte Carlo cross validation, selecting 120 samples from a data set with the abnormal samples removed as a correction set and a validation set by adopting a random selection mode, taking the rest samples as independent test sets, and dividing the 120 samples into the correction set and the validation set according to a ratio of 3:1 by adopting a Kernard-Stone method; and preprocessing the spectral data by adopting Savitzky-Golay convolution smoothing, multivariate scattering correction, standard regular transformation, derivative processing, wavelet transformation and a combination method thereof on the correction set samples, and selecting a spectral preprocessing method based on the minimum cross validation root mean square error of the partial least square regression model of the correction set.
(5) Preliminary near infrared spectrum characteristic wavelength optimization based on competitive adaptive re-weighting sampling algorithm
The competitive adaptive re-weighting sampling algorithm obtains a series of variable combinations based on a partial least square algorithm regression coefficient, and selects a subset with the minimum cross validation root mean square error as an optimal characteristic wavelength. However, two random factors, namely Monte Carlo sampling and adaptive weighted sampling, are introduced in the iterative search process of the characteristic wavelength, so that the consistency of the preferred characteristic wavelength every time is difficult to ensure. In order to solve the problem that the optimized result of the competitive adaptive re-weighting sampling algorithm is inconsistent, the performance of the regression model is further improved by executing the algorithm for multiple times and selecting the characteristic wavelength points selected repeatedly for multiple times as the optimized result. And executing 500 rounds of the competitive adaptive reweighting sampling algorithm, and selecting a repeatedly selected wavelength point corresponding to the minimum root mean square error of the cross validation of the correction set as a primary preference result of the characteristic wavelength obtained by the competitive adaptive reweighting sampling algorithm.
(6) Synchronous optimization of support vector machine parameters and near infrared spectrum characteristic wavelengths based on a quantum genetic simulated annealing algorithm aims at the requirements of support vector machine regression model parameter optimization and the problem that a few wavelength points with weak correlation exist in preliminary optimization results of near infrared spectrum characteristic wavelengths of a competitive adaptive reweighted sampling algorithm, and the quantum genetic simulated annealing algorithm is adopted to synchronously optimize the support vector machine parameters and the near infrared spectrum characteristic wavelengths. The quantum genetic simulated annealing algorithm adopts multi-quantum bit coding, and the number of the near infrared spectrum characteristic wavelength points initially selected by the competitive adaptive reweighted sampling algorithm is set as k, so that the punishment parameters, the kernel function parameters, the insensitive loss function parameters and the characteristic wavelength coding structure of the support vector machine are as follows:
wherein (α)mn,βmn) Is the probability amplitude of the quantum state, and | αmn|2+|βmn|2When k. is population initialization, 1,2,3,4, 1,2, …, then (α)mn,βmn) Is composed of
In the decoding process of the quantum genetic simulated annealing algorithm, the multi-quantum bit codes are collapsed into a binary sequence a through quantum probability1a2…akb1b2…bkc1c2…ckd1d2…dkCorresponding to k-bit penalty parameter, kernel function parameter and insensitive loss functionBinary encoding of the number parameter and the characteristic wavelength. The penalty parameter, the kernel function parameter and the insensitive loss function parameter are decoded by adopting binary real numbers, and the characteristic wavelength coding directly determines whether the corresponding wavelength point participates in the operation according to the values of binary bits '1' and '0'.
The quantum genetic simulated annealing algorithm adopts the cross validation of a regression model of a support vector machine to obtain a target function f (x) and a fitness function fit (x) ═ exp [ - (f (x)) fmin)/t]Wherein f isminAnd t is the minimum objective function value of the current generation population and is a temperature parameter. The evolution process of the quantum genetic simulated annealing algorithm comprises round-robin selection with an optimal retention strategy, quantum crossing, quantum variation, quantum rotating gate perturbation solution construction and Metropolis selective replication evolution.
When synchronous optimization of support vector machine parameters and near infrared spectrum characteristic wavelengths is carried out, encoding and population initialization are carried out, after quantum probability collapse is carried out on chromosomes in a population, decoding operation is carried out, and penalty parameters, kernel function parameters and insensitive loss function parameter values are obtained. And inputting the spectral data of the correction set into a regression model of a support vector machine according to the characteristic wavelength coding value, establishing the regression model according to the support vector machine parameters obtained by decoding, calculating a target function and a fitness function, and performing genetic evolution operation according to the fitness function value to obtain a next generation population. And after the cooling operation is executed, continuously executing the iterative processes of decoding, calculating a fitness function, genetic evolution and generating a new population until an algorithm termination condition is met to obtain the optimal parameter combination and the optimal characteristic wavelength of the support vector machine.
(7) Evaluation model establishment and evaluation
And inputting the spectral data of the correction set into a support vector machine according to the optimal characteristic wavelength, training a regression model of the support vector machine by using the optimal parameters of the optimized support vector machine, establishing a regression model of the near infrared spectrum support vector machine for the rice taste quality, and evaluating the precision of the regression model by adopting a verification set and an independent test set. If the evaluation results of the verification set and the independent test set do not meet the test precision requirement, re-executing (6), and synchronously optimizing the parameters of the support vector machine and the characteristic wavelength of the near infrared spectrum; and when the detection precision of the established support vector machine regression model meets the requirement, outputting a corresponding model, and completing establishment of the fast taste quality evaluation model of the Heilongjiang rice.
(8) Evaluation of taste quality of Rice
The method comprises the steps of selecting, hulling and rice milling the rice to be measured for taste quality, collecting near infrared diffuse reflection spectrum, preprocessing spectrum data, and inputting the preprocessed spectrum data into a regression model according to the optimized characteristic wavelength, so that the taste quality can be rapidly evaluated. And the collected polished rice sample is directly subjected to near-infrared diffuse reflection spectrum scanning, and the preprocessed spectrum data is input into an evaluation model according to characteristic wavelengths, so that the rapid evaluation of the taste quality can be completed.
Claims (1)
1. The near infrared spectrum rapid evaluation method for the taste quality of rice is realized by the following steps:
(1) sample collection and processing
Collecting japonica rice samples of different producing areas, collecting no less than 30 samples in each producing area, collecting no less than 150 total samples, selecting, hulling, and milling the collected samples to obtain rice polished rice samples, sealing and bagging for later use;
(2) near infrared spectral collection
The near-infrared diffuse reflection spectrum of the rice grains is collected by adopting a German Bruker TANGO near-infrared spectrometer, and the spectrum collection range is 3946-11542 cm-1Resolution of 8.0cm-1Scanning a sample for 32 times, wherein the sample loading mode is scanning by a sample cup rotating table of 50mm, the sample loading height is about 30mm, and the average value of 3-5 times of scanning is taken as the original spectrum of the sample;
(3) taste quality assessment
Evaluating the taste quality of the collected and prepared rice sample by adopting a national standard method, and grading by adopting a percent system mode;
(4) spectral preprocessing and sample set partitioning
Removing abnormal samples by adopting a residual mean-variance distribution diagram based on Monte Carlo cross validation, selecting more than 120 samples from a data set with the abnormal samples removed as a correction set and a validation set by adopting a random selection mode, dividing the correction set and the validation set according to a ratio of 3:1 by adopting a Kernard-Stone method, and taking the rest samples as independent test sets; preprocessing the spectrum data of the correction set by adopting a Savitzky-Golay convolution smoothing method, a multivariate scattering correction method, a standard regular transformation method, a derivative processing method, a wavelet transformation method and a combination method thereof, and selecting a spectrum preprocessing method based on the minimum cross validation root-mean-square error of a partial least square regression model of the correction set;
(5) preliminary near infrared spectrum characteristic wavelength optimization based on competitive adaptive re-weighting sampling algorithm
Executing a plurality of rounds of a competitive adaptive reweighting sampling algorithm on the correction set data obtained in the step (4), and selecting a repeatedly selected wavelength point corresponding to the correction set when the root mean square error is minimum in cross validation as a primary optimal result of the characteristic wavelength obtained by the competitive adaptive reweighting sampling algorithm;
the method is characterized in that:
(6) support vector machine parameter and near infrared spectrum characteristic wavelength synchronous optimization based on quantum genetic simulated annealing algorithm
Synchronously optimizing parameters of the support vector machine and the characteristic wavelength of the near infrared spectrum obtained in the step (5) by adopting a quantum genetic simulated annealing algorithm, wherein the quantum genetic simulated annealing algorithm adopts multi-quantum bit coding, and the punishment parameter, the kernel function parameter, the insensitive loss function parameter and the characteristic wavelength coding structure of the support vector machine are as follows if the number of the initially selected characteristic wavelength points of the near infrared spectrum of the competitive adaptive re-weighted sampling algorithm is k:
wherein (α)mn,βmn) Is the probability amplitude of the quantum state, and | αmn|2+|βmn|21,2,3,4, 1,2, …, k, and (α) when initializing the populationmn,βmn) Is composed of
In the decoding process of the quantum genetic simulated annealing algorithm, the multi-quantum bit codes are collapsed into a binary sequence a through quantum probability1a2…akb1b2…bkc1c2…ckd1d2…dkBinary coding corresponding to the k-bit penalty parameter, kernel function parameter, insensitive loss function parameter and characteristic wavelength; the punishment parameter, the kernel function parameter and the insensitive loss function parameter are decoded by adopting binary real numbers, and the characteristic wavelength coding directly determines whether the corresponding wavelength point participates in the operation according to the values of binary bits '1' and '0';
the quantum genetic simulated annealing algorithm adopts the cross validation of a regression model of a support vector machine to obtain a target function f (x) and a fitness function fit (x) ═ exp [ - (f (x)) fmin)/t]Wherein f isminIs the minimum objective function value of the current generation population, and t is a temperature parameter; the evolution process of the quantum genetic simulated annealing algorithm comprises gambling wheel selection with an optimal retention strategy, quantum crossing, quantum variation, quantum revolving gate disturbance solution construction and Metropolis selection replication evolution;
when synchronous optimization of support vector machine parameters and near infrared spectrum characteristic wavelengths is carried out, encoding and population initialization are carried out, after quantum probability collapse is carried out on chromosomes in a population, decoding operation is carried out to obtain punishment parameters, kernel function parameters and insensitive loss function parameter values; inputting the spectral data of the correction set into a regression model of a support vector machine according to the characteristic wavelength coding value, establishing the regression model according to the support vector machine parameters obtained by decoding, calculating a target function and a fitness function, and performing genetic evolution operation according to the fitness function value to obtain a next generation population; after the cooling operation is executed, the iterative processes of decoding, calculating a fitness function, genetic evolution and generating a new population are continuously executed until an algorithm termination condition is met, and the optimal parameter combination and the optimal characteristic wavelength of the support vector machine are obtained;
(7) evaluation model building
Inputting the spectral data of the correction set into a support vector machine according to the optimal characteristic wavelength, training a regression model of the support vector machine by using the optimal parameters of the optimized support vector machine, establishing a regression model of the near infrared spectrum support vector machine for the taste quality of rice, and evaluating the precision of the regression model by adopting a verification set and an independent test set; when the detection precision of the established support vector machine regression model meets the requirement, outputting a corresponding model to complete establishment of a rice taste quality rapid evaluation model;
(8) evaluation of taste quality of Rice
And performing near-infrared diffuse reflection spectrum scanning on the polished rice needing to measure the taste quality, and inputting the preprocessed spectrum data into an evaluation model according to characteristic wavelengths to finish the rapid evaluation of the taste quality.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911382675.6A CN111007040B (en) | 2019-12-27 | 2019-12-27 | Near infrared spectrum rapid evaluation method for rice taste quality |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911382675.6A CN111007040B (en) | 2019-12-27 | 2019-12-27 | Near infrared spectrum rapid evaluation method for rice taste quality |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111007040A true CN111007040A (en) | 2020-04-14 |
| CN111007040B CN111007040B (en) | 2020-10-23 |
Family
ID=70119305
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911382675.6A Active CN111007040B (en) | 2019-12-27 | 2019-12-27 | Near infrared spectrum rapid evaluation method for rice taste quality |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111007040B (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111965137A (en) * | 2020-08-18 | 2020-11-20 | 山东金璋隆祥智能科技有限责任公司 | Method for measuring content and moisture of components in cold heat clearing granules |
| CN112285057A (en) * | 2020-11-27 | 2021-01-29 | 常州金坛江南制粉有限公司 | Method for rapidly detecting water content of water-milled glutinous rice flour based on near infrared spectrum technology |
| CN112881323A (en) * | 2021-01-14 | 2021-06-01 | 云南中烟工业有限责任公司 | Quality evaluation method of cigarette mainstream smoke |
| CN112947332A (en) * | 2021-02-04 | 2021-06-11 | 天津国科医工科技发展有限公司 | Triple quadrupole mass spectrometer parameter optimization method based on simulated annealing |
| CN113125377A (en) * | 2021-03-30 | 2021-07-16 | 武汉理工大学 | Method and device for detecting diesel oil property based on near infrared spectrum |
| CN114136918A (en) * | 2021-11-29 | 2022-03-04 | 中国科学院合肥物质科学研究院 | Near-infrared-based rice taste quality evaluation method |
| CN114200093A (en) * | 2021-10-27 | 2022-03-18 | 吉林大学 | Rice taste double-blind quality identification sorting system and method |
| CN114471786A (en) * | 2021-12-15 | 2022-05-13 | 怀宁县皖河长青米业有限公司 | Cooling device based on rice processing |
| CN115114838A (en) * | 2022-07-22 | 2022-09-27 | 黑龙江八一农垦大学 | Spectral characteristic wavelength selection method based on particle swarm algorithm thought and simulated annealing strategy |
| WO2023092647A1 (en) * | 2021-11-24 | 2023-06-01 | 江苏大学 | Rapid quantification evaluation method for taste characteristics of fried rice |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105067784A (en) * | 2015-07-22 | 2015-11-18 | 中国水稻研究所 | Method for determiningtaste quality of rice by using physical and chemical property indexes of rice quality |
| CN106485049A (en) * | 2016-09-09 | 2017-03-08 | 黑龙江大学 | A kind of detection method of the NIRS exceptional sample based on Monte Carlo cross validation |
| CN107677638A (en) * | 2017-09-29 | 2018-02-09 | 贵州大学 | It is a kind of first roasting cigarette rate containing stalk quick determination method based on near-infrared spectrum technique |
| CN108037256A (en) * | 2017-12-05 | 2018-05-15 | 中国水稻研究所 | The rapid assay methods of rice eating-quality |
| CN108732128A (en) * | 2018-05-30 | 2018-11-02 | 山东省花生研究所 | A method of detection shelled peanut eats organoleptic quality |
| CN110163444A (en) * | 2019-05-28 | 2019-08-23 | 河北工程大学 | A kind of water demand prediction method based on GASA-SVR |
-
2019
- 2019-12-27 CN CN201911382675.6A patent/CN111007040B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105067784A (en) * | 2015-07-22 | 2015-11-18 | 中国水稻研究所 | Method for determiningtaste quality of rice by using physical and chemical property indexes of rice quality |
| CN106485049A (en) * | 2016-09-09 | 2017-03-08 | 黑龙江大学 | A kind of detection method of the NIRS exceptional sample based on Monte Carlo cross validation |
| CN107677638A (en) * | 2017-09-29 | 2018-02-09 | 贵州大学 | It is a kind of first roasting cigarette rate containing stalk quick determination method based on near-infrared spectrum technique |
| CN108037256A (en) * | 2017-12-05 | 2018-05-15 | 中国水稻研究所 | The rapid assay methods of rice eating-quality |
| CN108732128A (en) * | 2018-05-30 | 2018-11-02 | 山东省花生研究所 | A method of detection shelled peanut eats organoleptic quality |
| CN110163444A (en) * | 2019-05-28 | 2019-08-23 | 河北工程大学 | A kind of water demand prediction method based on GASA-SVR |
Non-Patent Citations (7)
| Title |
|---|
| LIU JINMING ET AL.: "A Method for Missing Data Recovery of Air Pollutants Monitoring in Henhouse Based on QGSA-SVM", 《INTERNATIONAL JOURNAL OF SMART HOME》 * |
| MAO ZAN ET AL.: "Parameter Optimization of SVM Based on HQGA", 《2010 SIXTH INTERNATIONAL CONFERENCE ON NATURAL COMPUTATION》 * |
| RAVIPAT LAPCHAROENSUK ET AL.: "Eating quality of cooked rice determination using Fourier transform near infrared spectroscopy", 《JOURNAL OF INNOVATIVE OPTICAL HEALTH SCIENCES》 * |
| 刘金明等: "基于GSA-SVM的畜禽舍废气监测缺失数据恢复方法", 《东北农业大学学报》 * |
| 刘金明等: "玉米秸秆纤维素和半纤维素NIRS特征波长优选", 《光谱学与光谱分析》 * |
| 孟庆虹等: "可见光-近红外光谱预测粳稻食味的适应性研究", 《中国粮油学报》 * |
| 聂黎行等: "近红外光谱技术结合竞争自适应重加权采样算法用于中药定量分析", 《中国实验方剂学杂志》 * |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111965137A (en) * | 2020-08-18 | 2020-11-20 | 山东金璋隆祥智能科技有限责任公司 | Method for measuring content and moisture of components in cold heat clearing granules |
| CN112285057A (en) * | 2020-11-27 | 2021-01-29 | 常州金坛江南制粉有限公司 | Method for rapidly detecting water content of water-milled glutinous rice flour based on near infrared spectrum technology |
| CN112881323A (en) * | 2021-01-14 | 2021-06-01 | 云南中烟工业有限责任公司 | Quality evaluation method of cigarette mainstream smoke |
| CN112947332B (en) * | 2021-02-04 | 2022-12-16 | 威高国科质谱医疗科技(天津)有限公司 | Triple quadrupole mass spectrometer parameter optimization method based on simulated annealing |
| CN112947332A (en) * | 2021-02-04 | 2021-06-11 | 天津国科医工科技发展有限公司 | Triple quadrupole mass spectrometer parameter optimization method based on simulated annealing |
| CN113125377A (en) * | 2021-03-30 | 2021-07-16 | 武汉理工大学 | Method and device for detecting diesel oil property based on near infrared spectrum |
| CN113125377B (en) * | 2021-03-30 | 2024-02-23 | 武汉理工大学 | Method and device for detecting property of diesel based on near infrared spectrum |
| CN114200093A (en) * | 2021-10-27 | 2022-03-18 | 吉林大学 | Rice taste double-blind quality identification sorting system and method |
| WO2023092647A1 (en) * | 2021-11-24 | 2023-06-01 | 江苏大学 | Rapid quantification evaluation method for taste characteristics of fried rice |
| CN114136918A (en) * | 2021-11-29 | 2022-03-04 | 中国科学院合肥物质科学研究院 | Near-infrared-based rice taste quality evaluation method |
| CN114136918B (en) * | 2021-11-29 | 2023-11-14 | 中国科学院合肥物质科学研究院 | Near infrared-based rice taste quality evaluation method |
| CN114471786A (en) * | 2021-12-15 | 2022-05-13 | 怀宁县皖河长青米业有限公司 | Cooling device based on rice processing |
| CN115114838A (en) * | 2022-07-22 | 2022-09-27 | 黑龙江八一农垦大学 | Spectral characteristic wavelength selection method based on particle swarm algorithm thought and simulated annealing strategy |
| CN115114838B (en) * | 2022-07-22 | 2023-02-07 | 黑龙江八一农垦大学 | Spectral characteristic wavelength selection method based on particle swarm algorithm thought and simulated annealing strategy |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111007040B (en) | 2020-10-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111007040B (en) | Near infrared spectrum rapid evaluation method for rice taste quality | |
| CN107464020B (en) | Method for rapidly screening rice product processing raw materials | |
| Fernández-Novales et al. | Shortwave-near infrared spectroscopy for determination of reducing sugar content during grape ripening, winemaking, and aging of white and red wines | |
| CN110907393B (en) | Method and device for detecting saline-alkali stress degree of plants | |
| Liu et al. | Applying near-infrared spectroscopy and chemometrics to determine total amino acids in herbicide-stressed oilseed rape leaves | |
| Guimarães et al. | Use of NIRS to predict composition and bioethanol yield from cell wall structural components of sweet sorghum biomass | |
| CN108169165B (en) | Maltose mixture quantitative analysis method based on terahertz spectrum and image information fusion | |
| CN109409579A (en) | The method of BP neural network prediction Raw material processing suitability | |
| CN108254324B (en) | Method and device for rapidly detecting freshness of crab meat | |
| Shi et al. | Combination of near-infrared spectroscopy and key wavelength-based screening algorithm for rapid determination of rice protein content | |
| He et al. | Study on the identification of resistance of rice blast based on near infrared spectroscopy | |
| CN112595675B (en) | Hyperspectrum-based rice grain quality functional gene analysis method | |
| Sinelli et al. | Prediction of semolina technological quality by FT-NIR spectroscopy | |
| Cato et al. | Wheat quality: wheat breeding and quality testing in Australia | |
| CN112881333B (en) | Near infrared spectrum wavelength screening method based on improved immune genetic algorithm | |
| Siebenmorgen et al. | Assessing rice milling quality | |
| CN111950882B (en) | Wine sample grading method based on comprehensive sensory analysis | |
| CN112763448A (en) | ATR-FTIR technology-based method for rapidly detecting content of polysaccharides in rice bran | |
| Lebot | Near infrared spectroscopy for quality evaluation of root crops: practical constraints, preliminary studies and future prospects | |
| CN112101817A (en) | Method for evaluating homogenization degree of batch-processed flue-cured tobacco finished products | |
| CN105241824A (en) | Method for quantitatively detecting solid fermentation index distribution difference through hyperspectral image technology | |
| CN114136918B (en) | Near infrared-based rice taste quality evaluation method | |
| CN114693047A (en) | Method for evaluating taste quality of rice and artificial neural network | |
| Lever et al. | Australian wheat for the sponge and dough bread making process | |
| ZHANG et al. | Comparison of near infrared spectroscopy models for determining protein and amylose contents between calibration samples of recombinant inbred lines and conventional varieties of rice |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |