CN116242802A - Method for detecting density of white spirit base by utilizing near infrared spectrum technology - Google Patents
Method for detecting density of white spirit base by utilizing near infrared spectrum technology Download PDFInfo
- Publication number
- CN116242802A CN116242802A CN202310270818.4A CN202310270818A CN116242802A CN 116242802 A CN116242802 A CN 116242802A CN 202310270818 A CN202310270818 A CN 202310270818A CN 116242802 A CN116242802 A CN 116242802A
- Authority
- CN
- China
- Prior art keywords
- white spirit
- near infrared
- density
- spectrum
- spirit base
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 151
- 238000002329 infrared spectrum Methods 0.000 title claims abstract description 84
- 238000005516 engineering process Methods 0.000 title claims abstract description 45
- 238000001228 spectrum Methods 0.000 claims abstract description 84
- 238000012937 correction Methods 0.000 claims abstract description 59
- 238000004458 analytical method Methods 0.000 claims abstract description 47
- 235000014101 wine Nutrition 0.000 claims abstract description 47
- 230000002159 abnormal effect Effects 0.000 claims abstract description 35
- 230000036961 partial effect Effects 0.000 claims abstract description 27
- 238000012216 screening Methods 0.000 claims abstract description 22
- 238000003379 elimination reaction Methods 0.000 claims abstract description 19
- 230000008030 elimination Effects 0.000 claims abstract description 18
- 230000002860 competitive effect Effects 0.000 claims abstract description 17
- 238000009499 grossing Methods 0.000 claims abstract description 15
- 239000000796 flavoring agent Substances 0.000 claims abstract description 9
- 238000010561 standard procedure Methods 0.000 claims abstract description 9
- 238000004433 infrared transmission spectrum Methods 0.000 claims abstract description 7
- 239000011159 matrix material Substances 0.000 claims description 28
- 238000010238 partial least squares regression Methods 0.000 claims description 24
- 238000004422 calculation algorithm Methods 0.000 claims description 22
- 230000003595 spectral effect Effects 0.000 claims description 17
- 230000000694 effects Effects 0.000 claims description 12
- 238000004364 calculation method Methods 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 9
- 238000004497 NIR spectroscopy Methods 0.000 claims description 7
- 238000001739 density measurement Methods 0.000 claims description 7
- 238000011156 evaluation Methods 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 6
- 238000000326 densiometry Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 5
- 230000003044 adaptive effect Effects 0.000 claims description 4
- 230000003247 decreasing effect Effects 0.000 claims description 4
- 235000020097 white wine Nutrition 0.000 claims description 4
- 238000013480 data collection Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 3
- 238000003860 storage Methods 0.000 claims description 2
- 238000001514 detection method Methods 0.000 abstract description 48
- 238000004519 manufacturing process Methods 0.000 abstract description 20
- 230000001276 controlling effect Effects 0.000 abstract description 2
- 230000001105 regulatory effect Effects 0.000 abstract description 2
- LFQSCWFLJHTTHZ-UHFFFAOYSA-N Ethanol Chemical compound CCO LFQSCWFLJHTTHZ-UHFFFAOYSA-N 0.000 description 15
- 230000001476 alcoholic effect Effects 0.000 description 11
- 238000002203 pretreatment Methods 0.000 description 8
- 239000000126 substance Substances 0.000 description 8
- 238000012544 monitoring process Methods 0.000 description 7
- 238000005070 sampling Methods 0.000 description 6
- 238000013528 artificial neural network Methods 0.000 description 5
- 238000000605 extraction Methods 0.000 description 5
- 238000012417 linear regression Methods 0.000 description 5
- 238000005259 measurement Methods 0.000 description 5
- 230000007547 defect Effects 0.000 description 4
- 238000011161 development Methods 0.000 description 4
- 230000018109 developmental process Effects 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- 238000012706 support-vector machine Methods 0.000 description 4
- 230000009897 systematic effect Effects 0.000 description 4
- 238000013124 brewing process Methods 0.000 description 3
- 238000012569 chemometric method Methods 0.000 description 3
- 238000002790 cross-validation Methods 0.000 description 3
- 230000001419 dependent effect Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 230000006872 improvement Effects 0.000 description 3
- 238000012628 principal component regression Methods 0.000 description 3
- UFHFLCQGNIYNRP-UHFFFAOYSA-N Hydrogen Chemical compound [H][H] UFHFLCQGNIYNRP-UHFFFAOYSA-N 0.000 description 2
- 206010063385 Intellectualisation Diseases 0.000 description 2
- 238000010521 absorption reaction Methods 0.000 description 2
- 238000003491 array Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 235000013305 food Nutrition 0.000 description 2
- 239000001257 hydrogen Substances 0.000 description 2
- 229910052739 hydrogen Inorganic materials 0.000 description 2
- 238000002156 mixing Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000004801 process automation Methods 0.000 description 2
- 238000004445 quantitative analysis Methods 0.000 description 2
- 238000011897 real-time detection Methods 0.000 description 2
- 230000002829 reductive effect Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 240000004808 Saccharomyces cerevisiae Species 0.000 description 1
- 240000006394 Sorghum bicolor Species 0.000 description 1
- 235000011684 Sorghum saccharatum Nutrition 0.000 description 1
- 238000002835 absorbance Methods 0.000 description 1
- 238000000862 absorption spectrum Methods 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 238000000149 argon plasma sintering Methods 0.000 description 1
- 235000013339 cereals Nutrition 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010924 continuous production Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000000855 fermentation Methods 0.000 description 1
- 230000004151 fermentation Effects 0.000 description 1
- 235000019634 flavors Nutrition 0.000 description 1
- 239000007789 gas Substances 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012845 near infrared spectroscopy analysis Methods 0.000 description 1
- 238000001320 near-infrared absorption spectroscopy Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 238000000513 principal component analysis Methods 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000007430 reference method Methods 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 238000012827 research and development Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000004611 spectroscopical analysis Methods 0.000 description 1
- 238000010183 spectrum analysis Methods 0.000 description 1
- 238000010025 steaming Methods 0.000 description 1
- 230000004083 survival effect Effects 0.000 description 1
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
- G01N21/25—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
- G01N21/01—Arrangements or apparatus for facilitating the optical investigation
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
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 belongs to the technical field of spectrum detection and analysis, and discloses a method for detecting the density of white spirit base liquor by utilizing a near infrared spectrum technology, which takes fen-flavor white spirit base liquor as an object and collects the near infrared transmission spectrum of the white spirit base liquor; detecting the density of the base wine sample by using a national standard method, and detecting and eliminating abnormal values by using a Markov distance method; sample pretreatment is carried out by adopting a mode of combining first-order conduction, convolution smoothing, multi-element scattering correction and first-order conduction and multi-element scattering correction; establishing a near infrared rapid analysis model of the white spirit base density by adopting a partial least square regression method; selecting interval partial least square method, non-information variable elimination method, continuous projection method and competitive self-adaptive weighting method to screen near infrared spectrum of white spirit base liquor, and screening wavelength variable with high correlation with white spirit base liquor density from the spectrum which is optimally pretreated. The invention realizes the function of online prediction by the near infrared spectrum technology, is used for timely regulating and controlling the white spirit production process and improves the white spirit production quality.
Description
Technical Field
The invention belongs to the technical field of spectrum detection and analysis, and particularly relates to a method for detecting white spirit base density by utilizing a near infrared spectrum technology.
Background
At present, the distilled spirit has thousands of years history in China, is a traditional distilled spirit in China, and is mainly prepared from sorghum or grain mixture serving as a raw material through the steps of steaming, mixing with yeast, fermenting, distilling, ageing, blending and the like. The main components of the white spirit in China are ethanol, water and various flavor substances, the alcohol content is one of the most important indexes for controlling the quality of the white spirit, and the detection of the white spirit is a very main link. The measurement of the alcoholic strength of the white spirit in China is carried out according to GB 5009.225-2016 'determination of the concentration of ethanol in national Standard wine for food safety', and the standard method for measuring the alcoholic strength comprises an alcohol meter method, a density bottle method, a digital densimeter method and the like. The current common method for measuring the alcoholic strength of the white spirit still depends on density measurement, but the original density detection method has various problems such as low working efficiency, easy uncertainty introduction, high sample consumption, long analysis time and the like.
The near infrared spectrum technology (near infrared spectrometry, NIR) is a detection technology developed recently, has the characteristics of high detection speed, high detection precision, simultaneous detection of multiple indexes, on-line monitoring and the like, and establishes a near infrared spectrum model of a sample by collecting near infrared spectrum data of the sample so as to detect an unknown sample.
The white spirit brewing is an online continuous process, and the traditional national standard method for detecting the white spirit base alcohol degree is an offline detection mode, so that the real-time detection cannot be realized, and the purposes of full-process automation and intellectualization of the white spirit brewing process are achieved. There is an urgent need for a rapid, efficient, stable and reliable measurement technique for real-time online measurement of white wine density.
The prior method for measuring the alcoholicity of the white spirit production site still relies on manual work to adopt a traditional detection method, and the concentration of a sample is roughly measured by using an alcohol meter, so that the detection result is unstable and errors are easy to generate. In order to obtain a more accurate measurement result, an off-line measurement mode is required, and a high-cost instrument such as a gas chromatograph is generally utilized to perform detection in a laboratory, so that time and labor are wasted. The conventional analysis methods have the defects of complicated steps, pretreatment of samples, time and labor consumption and the like before the test, are not suitable for measuring a large number of liquor samples, cannot realize real-time online detection, and restrict the rapid development of Chinese liquor detection technology.
The analysis information which is important to bear in the near infrared spectrum region is frequency multiplication and frequency combination information of the hydrogen-containing groups in the molecules, and the characteristic information of the hydrogen-containing groups in the sample can be obtained by scanning the near infrared spectrum of the sample, so that the near infrared absorption spectrum of the substance can be obtained by detecting the condition that near infrared light is absorbed by an instrument, and the spectrum is subjected to certain pretreatment and optimization, so that the influence of various non-target factors on the spectrum is eliminated or weakened, and the required result is detected. Because of the particularity of white spirit brewing, the online monitoring of the whole production process cannot be realized at present, the offline measurement is basically adopted for detecting the alcoholic strength of the white spirit base, the time consumption is long, the real-time regulation and optimization is difficult to carry out, and the production quality of the white spirit cannot be completely guaranteed.
Through the above analysis, the problems and defects existing in the prior art are as follows:
(1) The current method for measuring the alcoholic strength of the white spirit according to the density has various problems such as low working efficiency, easy uncertainty introduction, high sample consumption, long analysis time and the like.
(2) The traditional national standard method for detecting the alcoholic strength of the white spirit base is an off-line detection mode, cannot realize real-time detection, and achieves the purposes of full-process automation and intellectualization in the white spirit brewing process.
(3) At present, the detection result of measuring the alcohol degree by manually adopting the traditional detection method is unstable and is easy to generate errors; the off-line measuring instrument has higher cost, is time-consuming and labor-consuming, and is not suitable for measuring a large quantity of wine samples
(4) At present, online monitoring of the white spirit brewing process cannot be realized, the detection of the white spirit base alcohol degree by adopting offline is long in time consumption, real-time regulation and control optimization is difficult to carry out, and the production quality of white spirit cannot be completely guaranteed.
Disclosure of Invention
Aiming at the problems existing in the prior art, the invention provides a method for detecting the density of white spirit base liquor by utilizing a near infrared spectrum technology.
The invention is realized in such a way that the method for detecting the density of the white spirit base by utilizing the near infrared spectrum technology comprises the following steps: taking a fen-flavor white spirit base wine as an object, and collecting a near infrared transmission spectrum of the white spirit base wine; detecting the density of the base wine sample by using a national standard method, and detecting and eliminating abnormal values by using a Markov distance method; preprocessing a white spirit base wine sample, and establishing a near infrared rapid analysis model of white spirit base wine density by adopting a partial least squares regression method; and screening the near infrared spectrum of the white spirit base liquor, and screening from the optimally pretreated spectrum to obtain the wavelength variable which has higher correlation with the density of the white spirit base liquor.
Further, the method for detecting the density of the white spirit base by utilizing the near infrared spectrum technology comprises the following steps:
step one, white spirit density measurement: detecting the density of the white spirit base liquor sample according to a digital densitometry method;
step two, near infrared spectrum acquisition: collecting the near infrared spectrum of the white spirit base liquor by using a near infrared spectrometer;
step three, eliminating abnormal points: detecting and removing abnormal values of the spectrum data of the sample by using a Markov distance method;
step four, spectrum pretreatment: the method comprises the steps of preprocessing a white spirit base wine sample by adopting a mode of combining first-order guide, convolution smoothing, multi-element scattering correction and first-order guide and multi-element scattering correction;
step five, screening characteristic wave bands: selecting a space partial least square method, an information-free variable elimination method, a continuous projection method and a competitive self-adaptive weighting method, selecting a wavelength variable with higher correlation with the white spirit base wine density from the optimally preprocessed near infrared spectrum, and carrying out predictive analysis on the white spirit base wine density.
Step six, model construction and performance evaluation: adopting partial least square regression method to construct near infrared analysis model of white spirit base density, and adopting decision coefficient R 2 The experimental results are analyzed and compared by the root mean square correction error and the root mean square prediction error, and the near red is comprehensively obtained Evaluating the effect of the external analysis model;
in the second step, a Fourier transform near infrared spectrometer of Matrix-F is adopted to collect the near infrared transmission spectrum of the white spirit base wine. Preheating a near infrared spectrometer for 1h, and setting instrument parameters; collecting the background in the empty beaker by using a Fourier transform near infrared spectrometer of Matrix-F preheated for 1h, and pouring a white spirit base wine sample into the submerged data collection port; and (3) carrying out spectrum acquisition on the white wine sample by using a Fourier transform near infrared spectrometer of Matrix-F, and adopting spectrum signal acquisition software OPUS7.8 as spectrum data processing software.
Further, the parameters of the Fourier transform near infrared spectrometer of Matrix-F were set as follows: the resolution of the instrument during scanning was 8cm -1 The number of scans was 32, the spectral range was 12000cm wavenumber -1 ～4000cm -1 。
When the fen-flavor liquor base liquor samples are selected, at least 50 samples are selected from a single-component system; the content range of the selected modeling sample is larger than the range analyzed later; the modeled samples were uniform throughout the content range; the prepared modeled sample concentration was not linearly increased or decreased; the method for measuring the component content can obtain the result.
In the third step, abnormal values are detected and removed by the mahalanobis distance method. When abnormal points are identified by utilizing the mahalanobis distance, the sample set is divided into a correction set and a prediction set, the mahalanobis distance calculation is performed by utilizing the sample spectrum data of the correction set, and the abnormal points are judged by a method of setting a threshold value.
Wherein, the calculation process of the mahalanobis distance is as follows: dividing sample data into a correction set and a prediction set, and calculating to obtain an average spectrum of a sample; centering the spectrum matrix of the correction set, and subtracting the average spectrum matrix from the spectrum matrix of the correction set to obtain a Markov matrix; calculating to obtain a Markov matrix from the correction set sample to the average spectrum; setting a threshold value, and identifying abnormal points in the correction set sample through the threshold value.
In the fifth step, the spectral data corresponding to the key variables extracted by the interval partial least square method, the non-information variable elimination method, the continuous projection method and the competitive self-adaptive weighting method and the full-band spectral data after pretreatment are used as input data of PLS modeling, and the liquor base liquor density is predicted and analyzed.
Another object of the present invention is to provide a system for detecting a density of a white spirit base using a near infrared spectrum technology, using the method for detecting a density of a white spirit base using a near infrared spectrum technology, the system for detecting a density of a white spirit base using a near infrared spectrum technology comprising:
the white spirit density measuring module is used for detecting the density of the white spirit base liquor sample according to a digital densitometry method;
The near infrared spectrum acquisition module is used for acquiring the near infrared spectrum of the white spirit base wine by using a near infrared spectrometer;
the abnormal point removing module is used for detecting and removing abnormal values of the spectrum data of the sample by adopting a mahalanobis distance method;
the spectrum pretreatment module is used for carrying out pretreatment on the white spirit base wine sample by adopting a mode of combining first-order conduction, convolution smoothing, multi-element scattering correction and first-order conduction and multi-element scattering correction;
the characteristic wave band screening module is used for selecting interval partial least square, non-information variable elimination, continuous projection algorithm and competitive self-adaptive weighting algorithm, selecting wavelength variable with higher correlation with the white spirit base density from the optimal preprocessed near infrared spectrum, and carrying out predictive analysis on the white spirit base density.
The model construction and performance evaluation module is used for constructing a near infrared analysis model of the white spirit base density by adopting a partial least square regression method and adopting a decision coefficient R 2 Analyzing and comparing the experimental result, and evaluating the model effect on the whole;
another object of the present invention is to provide a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and the computer program when executed by the processor causes the processor to execute the steps of the method for detecting the density of white spirit base using the near infrared spectrum technology.
Another object of the present invention is to provide a computer readable storage medium storing a computer program, which when executed by a processor, causes the processor to perform the steps of the method for detecting the density of white spirit base using near infrared spectrum technology.
The invention further aims to provide an information data processing terminal which is used for realizing the system for detecting the white spirit base density by utilizing the near infrared spectrum technology.
In combination with the technical scheme and the technical problems to be solved, the technical scheme to be protected has the following advantages and positive effects:
first, aiming at the technical problems in the prior art and the difficulty of solving the problems, the technical problems solved by the technical proposal of the invention are analyzed in detail and deeply by tightly combining the technical proposal to be protected, the results and data in the research and development process, and the like, and some technical effects brought after the problems are solved have creative technical effects. The specific description is as follows:
the current common method for measuring the alcoholic strength of the white spirit is still based on density measurement, and the invention provides a method for detecting the density of a white spirit base by utilizing a near infrared spectrum technology. The invention takes the fen-flavor liquor base liquor as an analysis object, and establishes a near infrared rapid analysis model of the liquor base liquor density by collecting the near infrared transmission spectrum of the liquor base liquor, detecting the density of a base liquor sample by using a national standard method and adopting a partial least square method.
In the white spirit production process, the alcohol degree is still measured on the production site by adopting a traditional detection method manually, the detection method is complex, the detection time is long, the time and the labor are consumed, and the purpose of online detection cannot be realized. The invention provides a method for detecting the density of white spirit base liquor by a near infrared spectrum technology, which provides a reference for designing an online detection system for the white spirit production process, and has the following key improvement points:
1. the near infrared spectrum technology is applied to the prediction of the white spirit density, and has the advantages of simple operation, nondestructive detection, on-line detection and the like.
2. Sample pretreatment is carried out by adopting first-order guide (FD), convolution Smoothing (SG), multiple Scattering Correction (MSC) and FD+MSC in combination to reduce noise and systematic error, and the optimal pretreatment method is selected in comparison.
3. Selecting Interval Partial Least Squares (iPLS), non-information variable elimination (UVE), continuous projection algorithm (SPA) and competitive self-adaptive re-weighting algorithm (CARS) to screen the near infrared spectrum of the white spirit base wine, and selecting the wavelength variable which has higher correlation with the density of the white spirit base wine from the optimally preprocessed spectrum to remove the defects of large calculation amount of a model and low efficiency.
The current common method for measuring the alcoholic strength of the white spirit still depends on density measurement, but the original density detection method has various problems such as low working efficiency, easy uncertainty introduction, high sample consumption, long analysis time and the like. The near infrared spectrum technology (NIR) has the characteristics of high detection speed, high detection precision, simultaneous detection of multiple indexes, on-line monitoring and the like, and the near infrared spectrum technology is applied to the detection of the density of white spirit base wine, the optimal pretreatment method and the optimal band screening algorithm are compared and selected, the function of on-line prediction is realized through the near infrared spectrum technology so as to timely regulate and control the white spirit production process, the white spirit production quality is improved, and a theoretical basis is provided for the analysis and development of an on-line monitoring system in the white spirit production process.
Secondly, the technical scheme is regarded as a whole or from the perspective of products, and the technical scheme to be protected has the following technical effects and advantages:
according to the invention, the density of the white spirit base is predicted by adopting a near infrared spectrum technology, and the alcoholic strength of the white spirit is obtained in time according to the obtained result so as to adjust the production process, so that the purposes of improving the quality of the white spirit, realizing high quality and high yield and low consumption are achieved, and a thought is provided for developing a white spirit fermentation online detection system.
Thirdly, as inventive supplementary evidence of the claims of the present invention, the following important aspects are also presented:
the technical scheme of the invention solves the technical problems that people are always desirous of solving but are not successful all the time:
aiming at the problem that the alcoholic strength of the white spirit base liquor cannot be detected in real time in the current white spirit industry, the white spirit base liquor density is detected by adopting a near infrared spectrum technology based on the detection of the white spirit density, and the online detection of a part of the white spirit production process is realized.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the embodiments of the present invention will be briefly described below, and it is obvious that the drawings described below are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a method for detecting the density of white spirit base by using a near infrared spectrum technology according to an embodiment of the invention;
fig. 2 is a main flow chart of near infrared spectrum analysis according to an embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the following examples in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
Aiming at the problems existing in the prior art, the invention provides a method for detecting the density of white spirit base by utilizing a near infrared spectrum technology, and the invention is described in detail below with reference to the accompanying drawings.
According to the embodiment of the invention, the fen-flavor white spirit base wine is taken as an analysis object, and the near infrared transmission spectrum of the white spirit base wine is collected; detecting the density of the base wine sample by using a national standard method, and detecting and eliminating abnormal values by using a Markov distance method; sample pretreatment is carried out by adopting a mode of combining first-order conduction, convolution smoothing, multi-element scattering correction and first-order conduction and multi-element scattering correction; establishing a near infrared rapid analysis model of the white spirit base density by adopting a partial least square regression method; selecting an interval partial least square method, an information-free variable elimination method, a continuous projection method and a competitive self-adaptive weighting method to screen the near infrared spectrum of the white spirit base wine, and screening from the optimally pretreated spectrum to obtain a wavelength variable which has higher correlation with the density of the white spirit base wine, thereby realizing the density detection of the white spirit base wine.
As shown in fig. 1, the method for detecting the density of white spirit base by using the near infrared spectrum technology provided by the embodiment of the invention comprises the following steps:
s101, white spirit density measurement: detecting the density of the white spirit base liquor sample according to a digital densitometry method;
S102, near infrared spectrum acquisition: collecting the near infrared spectrum of the white spirit base liquor by using a near infrared spectrometer;
s103, eliminating abnormal points: detecting and removing abnormal values of the spectrum data of the sample by using a Markov distance method;
s104, spectrum pretreatment: the method comprises the steps of preprocessing a white spirit base wine sample by adopting a mode of combining first-order guide, convolution smoothing, multi-element scattering correction and first-order guide and multi-element scattering correction;
s105, characteristic wave band screening: selecting a space partial least square method, an information-free variable elimination method, a continuous projection method and a competitive self-adaptive weighting method, selecting a wavelength variable with higher correlation with the white spirit base wine density from the optimally preprocessed near infrared spectrum, and carrying out predictive analysis on the white spirit base wine density.
S106, model construction and performance evaluation: adopting partial least square regression method to construct near infrared analysis model of white spirit base density, and adopting decision coefficient R 2 Analyzing and comparing the experimental result, and evaluating the near infrared analysis model effect on the whole;
in step S102 provided by the embodiment of the invention, preheating a Fourier transform near infrared spectrometer of Matrix-F for 1h, and setting instrument parameters: the Fourier transform near infrared spectrometer resolution of Matrix-F at the time of scanning was 8cm -1 The number of scans was 32, the spectral range was 12000cm wavenumber -1 ～4000cm -1 The method comprises the steps of carrying out a first treatment on the surface of the Collecting back in empty beaker with Fourier transform near infrared spectrometer of Matrix-F preheated for 1hPouring the white spirit base wine sample to the submerged data acquisition port; and (3) carrying out spectrum acquisition on the white wine sample by using a Fourier transform near infrared spectrometer of Matrix-F, and adopting spectrum signal acquisition software OPUS7.8 as spectrum data processing software. When the fen-flavor liquor base liquor samples are selected, at least 50 samples are selected from a single-component system; the content range of the selected modeling sample is larger than the range analyzed later; the modeled samples were uniform throughout the content range; the prepared modeled sample concentration was not linearly increased or decreased; the method for measuring the component content can obtain the result.
In step S103 provided by the embodiment of the present invention, abnormal values are detected and removed by using a mahalanobis distance method. When abnormal points are identified by utilizing the mahalanobis distance, the sample set is divided into a correction set and a prediction set, the mahalanobis distance calculation is performed by utilizing the sample spectrum data of the correction set, and the abnormal points are judged by a method of setting a threshold value.
Wherein, the calculation process of the mahalanobis distance is as follows: dividing sample data into a correction set and a prediction set, and calculating to obtain an average spectrum of a sample; centering the spectrum matrix of the correction set, and subtracting the average spectrum matrix from the spectrum matrix of the correction set to obtain a Markov matrix; calculating to obtain a Markov matrix from the correction set sample to the average spectrum; setting a threshold value, and identifying abnormal points in the correction set sample through the threshold value.
The characteristic band screening in step S105 provided by the embodiment of the present invention includes:
and taking spectral data corresponding to key variables extracted by the interval partial least square method, the non-information variable elimination method, the continuous projection method and the competitive self-adaptive weighting method and the pretreated full-band spectral data as input data of PLS modeling, and carrying out predictive analysis on the white spirit base density.
The system for detecting the density of the white spirit base by utilizing the near infrared spectrum technology provided by the embodiment of the invention comprises the following steps:
the near infrared spectrum acquisition module is used for acquiring the near infrared spectrum of the white spirit base wine by using a near infrared spectrometer;
the white spirit density measuring module is used for detecting the density of the white spirit base liquor sample according to a digital densitometry method;
the abnormal point removing module is used for detecting and removing abnormal values of the spectrum data of the sample by adopting a mahalanobis distance method;
the spectrum pretreatment module is used for carrying out pretreatment on the white spirit base wine sample by adopting a mode of combining first-order conduction, convolution smoothing, multi-element scattering correction and first-order conduction and multi-element scattering correction;
the model construction and performance evaluation module is used for constructing a near infrared analysis model of the white spirit base density by adopting a partial least square regression method and adopting a decision coefficient R 2 Analyzing and comparing the experimental result, and evaluating the near infrared analysis model effect on the whole;
the characteristic wave band screening module is used for selecting interval partial least square, non-information variable elimination, continuous projection algorithm and competitive self-adaptive weighting algorithm, selecting wavelength variable with higher correlation with the white spirit base density from the optimal preprocessed near infrared spectrum, and carrying out predictive analysis on the white spirit base density.
In order to prove the inventive and technical value of the technical solution of the present invention, this section is an application example on specific products or related technologies of the claim technical solution.
1. Near infrared spectroscopy
Near infrared spectroscopy is somewhat different from conventional spectroscopy. The conventional spectrum analysis only needs to establish a working curve or mathematical relation between the absorbance of one or a plurality of wavelength points and the to-be-measured, but the same wavelength point in the near infrared spectrum can contain information of various substances, that is, one substance can contain information at a plurality of wavelength points in the near infrared spectrum region, so that the information in the near infrared spectrum region is complex, spectrum peaks can be overlapped with each other, the signal to noise ratio is low, the analysis of the near infrared spectrum is very difficult, and the obtained effective information is low. Therefore, the near infrared spectrum analysis must use a spectrum preprocessing method to correct the spectrum background, and combine with a chemometric method, such as principal component regression (principle component regression, PCR), partial least squares regression (PLS), and the like to pre-establish a mathematical model of the relationship between the spectrum and the to-be-measured spectrum. Chemometric methods combined with near infrared spectral quantitative analysis belong to the category of multivariate calibration, which is generally divided into two categories: a linear correction method and a nonlinear correction method. The linear correction method is a very mature method developed in near infrared spectrum analysis and mainly comprises a multiple linear regression method, a principal component regression method, a partial least square method and the like. Multiple linear regression (multiple linear regression, MLR) is a method often used for near infrared spectrum early analysis, and an optical filter spectroscopic system is adopted in an early near infrared spectrum instrument because of technical limitations and the like, so that a full-band spectrogram cannot be acquired. With the development of technical theory, novel spectroscopic systems are developed and utilized, and acquisition of full-band spectrograms is slowly realized by means of near infrared spectroscopic technology of the systems, so that PLS and PCR are widely applied to near infrared spectroscopic analysis, and a partial least square method is a standard method used in near infrared content analysis. However, various complex samples appear in the actual detection process, and the linear regression method is not applicable when detecting these samples, so the nonlinear correction method is used for establishing the near infrared spectrum analysis model, and good results can be obtained. The nonlinear correction method mainly comprises a support vector machine (support vector regression, SVR), an artificial neural network (artificial neural network, ANN) and the like. The near infrared spectrum analysis mainly comprises the processes of sample collection, determination of a reference value (physicochemical index), near infrared spectrum acquisition, spectrum data preprocessing, model establishment and optimization, model verification and the like (see figure 2).
1.1 characteristic band screening
The prediction model built by the full spectrum has large calculated amount and complex model, and when the full spectrum has the areas of low signal-to-noise ratio, serious spectrum overlap, strong absorption wave band of water and other component background, the prediction effect is often poor. Too many variables increase the complexity of the prediction model, and too few variables are easy to miss detailed information, so that the accuracy and precision of the method can be improved by adopting characteristic wavelength extraction. The invention combines the characteristic extraction methods of interval partial least squares regression (interval partial least square, iPLS), competitive self-adaptive weighting (competitive adaptive reweighted sampling, CARS), continuous projection algorithm (successive Projections Algorithm, SPA) and no-information variable elimination (uninformative variable elimination, UVE) to select the full spectrum interval.
Interval partial least squares regression (iPLS) is to divide a spectrum region participating in operational screening into a certain number of intervals (subintervals), then calculate cross-validation mean square error (root mean square error of cross validation, RMSECV) values of each subinterval, reject the interval corresponding to the maximum value each time, build PLS models on the rest intervals and give corresponding RMSECV values, and circulate in this way until a single subinterval remains to build PLS models, and in this one-to-one elimination process, a plurality of intervals corresponding to the minimum RMSECV values of each PLS model are the optimized optimal modeling combination interval.
The non-information variable cancellation method (UVE) refers to the method of the original spectrum data matrix X n×m Adding a data matrix R identical to the rows and columns of the data matrix R n×m I.e. adding noise to a set of interference spectral information, combined to form a new matrix X Rn×2m And then, a partial least square model is established by a one-by-one elimination method, the stability of each wavelength variable is analyzed according to the regression coefficient, and the variable which does not provide information is eliminated to improve the robustness of the model.
The calculation steps of the UVE method are as follows:
(1) Random generator matrix R n×m And it is combined with spectrum matrix X n×m Is combined into XR n×2m ；
(2) Establishing a partial least squares regression model of the XR and the physical and chemical value matrix Y of the component content of the detection sample, removing one sample at a time by cross validation, and finally obtaining a regression coefficient matrix B n×2m ；
(3) Calculation of B by column n×2m Mean (b) j ) And standard deviation std (b) j ) Calculating the stability S of the variable according to equation (1) j ；
(4) In [ m+1,2m]Calculating the threshold S according to the formula (2) in the range max ；
S max ＝max(abs(S k )),k＝m+1,m+2,...,2m (2)
(5) And deleting the non-information wavelength variable with smaller absolute value of stability than a set threshold value within the range of [1, m ], namely, the non-information wavelength variable with lower modeling contribution rate, and using the rest variable as an effective modeling variable in a final regression model.
The continuous projection algorithm (SPA) is a forward selection algorithm that finds the wavelength variable set with the lowest linear correlation degree in the original spectrum data as the final modeling variable. The method aims at summarizing most of spectral information of a test sample by using data of a few wavelength points, and reduces complexity of model fitting.
The SPA algorithm operates as follows:
(1) Randomly selecting a column x from the correction set spectral data vector j Denoted as x k(0) ；
(2) The variables that are not selected are denoted S,
(3) Calculating the selected column vector x according to equation (3) k(0) Projection p on variable space of S set xj ；
(4) Extracting a wavelength point corresponding to the maximum value in the n-1 projections, and recording as k (n), k (n) =arg (max||p xj ||)；
(5) Let x j ＝p xj ；
(6) Let n=n+1, if N < N, execute step (3) again, continue the loop.
The final set of wavelength points is { k (N), n=1, 2, …, N-1}, N being the total number of wavelength points to be extracted.
Competitive adaptive re-weighting (CARS), a new class of feature variable extraction methods, which are based on the theory of "survival of the fitter. In the process of N times of Monte Carlo (MC) sampling, an Exponential Decay Function (EDF) is utilized to play a key role, and in a preferential manner of superior and inferior elimination, a variable with smaller absolute value weight of a standardized regression coefficient (namely a wavelength point with low importance degree) in an adaptive re-weighted sampling (ARS) is eliminated, and a wavelength point with large weight of the standardized regression coefficient is reserved. And selecting the variable in the corresponding set when the RMSECV value reaches the lowest as the characteristic wavelength selected by the CARS method.
The CARS screening key variables comprises the following steps:
(1) Firstly, randomly selecting a modeling sample according to an MC sampling method, and establishing a regression model;
(2) Removing variables with smaller absolute values of regression coefficients by using EDF, reserving wavelength variables with large predicted contribution rate to the content of the component to be detected, and calculating the variable number reserved by the ith MC sample through a formula (4);
r i ＝ae -ki (4)
since all variables are retained at sample 1 MC, r is 1 =1; there are m wavelength points in total, leaving only 2 variables in the nth sample, so r N =2/m. In this way, in the formula (4),
(3) The variable corresponding to the relatively large absolute value in the regression coefficient is reserved through ARS sampling. Calculating the weight omega of each wavelength variable regression coefficient according to the formula (5) j Thereby measuring the importance of each wavelength variable;
(4) Comparing the magnitudes of N groups of RMSECV values obtained after the MC sampling for N times, wherein the variable set contained in the RMSECV minimum state is the optimal characteristic wavelength of the component to be detected.
1.2 Spectrum pretreatment
The obtained spectrum has serious absorption peak overlapping phenomenon in a short-wave band region, and can not directly reflect the effective information of the sample components, so that the analysis from the original spectrogram is very difficult. Meanwhile, in order to reduce interference caused by incoherent factors in the spectrum, such as light scattering, baseline drift, influence of instrument response, sample non-uniformity, moisture and the like, pretreatment of the white spirit base wine is required. The pretreatment methods commonly used include smooth noise reduction, vector normalization, additional scattering correction, first derivative, second derivative and the like, and the main purpose of the pretreatment method is to reduce noise and systematic errors to improve the effective information rate in the spectrum.
1.3 outlier rejection
Abnormal points are that the prediction accuracy of a model built after some samples are added into the model in the modeling process is greatly reduced, and the model is generally divided into two types, wherein one type is that the chemical value analysis or scanned spectrum of the abnormal samples has larger error, and the effective information rate of modeling data can be improved by removing the samples of the model in a concentrated manner, so that the prediction accuracy of the model is improved; the second type is that the chemical value and the scanning spectrum data of the samples are accurate, but the samples have larger difference from most of the modeled samples, the chemical value or the spectrum data of the samples are more special than other samples, and the addition of the samples to the model establishment can reduce the prediction accuracy of the model to the general samples, but can improve the applicability of the model. The main methods for eliminating abnormal points include a mahalanobis distance method, a lever value method, a principal component score method and the like.
1.4 modeling
The information of each spectrum of the near infrared spectrum is seriously overlapped, the spectrum peak is wide, and the near infrared spectrum quantitative analysis of the complex sample can be established by a corresponding chemometric method. Common methods for establishing quantitative models include Multiple Linear Regression (MLR), principal component regression (Principal component analysis, PCA), partial least squares regression (PLS), artificial Neural Networks (ANN), support vector machines (SVR), and the like.
The invention adopts partial least squares regression as a modeling method. The purpose of partial least squares regression is to find some linear combinations in the interpretation variable space to better interpret the variation information of the reaction variables.
There are q dependent variables { Y1, ··yq } and p independent variables { X1, ·xp, n sample points, and thus a self-variable data table x= (X1, ··xp) n×p, and a dependent variable data table y= (Y1, ··yq) n×q.
Partial least squares regression extracts the components t1 and u1 in X and Y, respectively, and when extracting these two components, there are two requirements for regression analysis:
(1) t1 and u1 should carry as large as possible the variation information in their respective data tables;
(2) the correlation degree between t1 and u1 can be maximized.
These two requirements indicate that t1 and u1 should represent tables X and Y as well as possible, while the component t1 of the independent variable has the strongest interpretation of the component u1 of the dependent variable. After the first components t1 and u1 are extracted, partial least squares regression achieves regression of X to t1 and Y to t1, respectively. If the regression equation has reached satisfactory accuracy, the algorithm terminates; otherwise, the component extraction of the second round is performed by using the residual information of X interpreted by t1 and the residual information of Y interpreted by t 1. This is repeated until a satisfactory accuracy is achieved. If m components t1, ··tm, partial least squares regression is finally extracted for X, then the regression of yk, k=1, 2, ·q to t1, ·tm will be performed, and then expressed again as a regression equation for yk with respect to the original variables X1, ·xp.
1.5 evaluation of model Performance
The invention adopts the determination coefficient R 2 The experimental results were compared analytically with root mean square correction errors (Root Mean Squared Error of Calibration, RMSEC) and root mean square prediction errors (Root Mean Squared Error of Prediction, RMSEP), and the model effect was evaluated as a whole.
2. Experimental procedure
2.1 near infrared Spectrum acquisition
The invention adopts a Fourier transform near infrared spectrometer of Matrix-F. In order to ensure the stability of the instrument, firstly, an infrared spectrometer is turned on to preheat for 1h; then the instrument is arrangedParameters: the resolution of the instrument during scanning was 8cm -1 The number of scans was 32, the spectral range was 12000cm wavenumber -1 ～4000cm -1 The method comprises the steps of carrying out a first treatment on the surface of the Firstly, a background in an empty beaker is collected by using a near infrared which is preheated for 1h, then a white spirit base wine sample (which can submerge a data collection port) is poured, a near infrared spectrometer is used for carrying out spectrum collection on the white spirit sample, and spectrum signal collection software OPUS7.8 is used as spectrum data processing software. Care should be taken in selecting experimental samples: (1) at least 50 samples of the one-component system; (2) the content range of the selected modeling sample is larger than the range analyzed later; (3) the modeling sample should be uniform throughout the content range; (4) the prepared modeling sample is used for ensuring that the concentration is not linearly increased or decreased; (5) ensuring that the reference method used to measure the component content gives reliable results.
2.2 determination of Density of white spirit
The density of the sample is detected according to the digital densimeter method in GB 5009.225-2016 (determination of ethanol concentration in national Standard for food safety) and the alcohol content can be obtained by looking up a table in the follow-up process.
2.3 outlier rejection
The invention adopts the mahalanobis distance method to detect and reject abnormal values. The method for identifying abnormal points by using the mahalanobis distance comprises the steps of dividing a sample set into a correction set and a prediction set, calculating the mahalanobis distance by using the sample spectrum data of the correction set, and judging the abnormal points by a method of setting a threshold value. The algorithm for the mahalanobis distance is as follows:
(1) Dividing the sample data into a correction set and a prediction set; (2) calculating an average spectrum of the sample; (3) Centering the spectrum matrix of the correction set, namely subtracting the average spectrum matrix from the spectrum matrix of the correction set; (4) obtaining a matrix of the Markov; (5) Calculating a Markov matrix from the correction set sample to the average spectrum; (6) Setting a threshold value, and identifying abnormal points in the correction set sample through the threshold value.
2.4 Spectrum pretreatment
Sample pretreatment is performed by using first-order derivative (FD), convolution smoothing (convolution smoothing, SG), multiple scatter correction (multivariate scattering correction, MSC) and fd+msc in combination, respectively, to reduce noise and systematic errors. As can be seen from table 1, the regression model established by the different spectrum pretreatment methods has different effects and the prediction results for the samples are different. Through comparison, the mean square error (RMSEC) of the correction set sample is close to and very small than the predicted mean square error (RMSEP) data of the verification set sample after the first derivative and the multiple scattering correction treatment, and the comprehensive performance of the model is optimal.
TABLE 1 influence of spectral pretreatment methods on prediction models
2.5 characteristic band screening
The invention selects Interval Partial Least Square (iPLS), no information variable elimination (UVE), continuous projection algorithm (SPA) and competitive self-adaptive weighting algorithm (CARS), and selects wavelength variable with higher correlation with white spirit base density from the spectrum of optimal pretreatment.
In order to investigate the influence of the characteristic wavelength selection method iPLS, UVE, SPA and CARS on the performance of the PLS model, the spectral data corresponding to the key variables extracted by the four methods and the pretreated full-band spectral data are taken as input data of PLS modeling, the liquor base liquor density is subjected to predictive analysis, and the model parameters of the iPLS, UVE, SPA and CARS models are listed in table 2.
TABLE 2 influence of different band screening methods on prediction models
From table 2, in the aspect of predicting the white spirit base density, the model established by adopting the iPLS, UVE, SPA and CARS method extracted key variables optimizes the PLS prediction model to a certain extent. Compared with a model without feature screening, the method has the advantages that errors are obviously reduced, so that the feature variable screening method can remove redundant wavelengths with low correlation degree with white spirit density from all-band wavelengths, extract key variables with high contribution rate, and has high model running speed and high precision. In the four feature variable extraction methods, 247 key variables are screened out by a UVE algorithm, and the established PLS model is relatively most stable.
In the white spirit production process, the alcohol degree is still measured on the production site by adopting a traditional detection method manually, the detection method is complex, the detection time is long, the time and the labor are consumed, and the purpose of online detection cannot be realized. The invention provides a method for detecting the density of white spirit base liquor by a near infrared spectrum technology, which provides a reference for designing an online detection system for the white spirit production process, and has the following key improvement points:
1. the near infrared spectrum technology is applied to the prediction of the white spirit density, and has the advantages of simple operation, nondestructive detection, on-line detection and the like.
2. Sample pretreatment is carried out by adopting first-order guide (FD), convolution Smoothing (SG), multiple Scattering Correction (MSC) and FD+MSC in combination to reduce noise and systematic error, and the optimal pretreatment method is selected in comparison.
3. Selecting Interval Partial Least Squares (iPLS), non-information variable elimination (UVE), continuous projection algorithm (SPA) and competitive self-adaptive re-weighting algorithm (CARS) to screen the near infrared spectrum of the white spirit base wine, and selecting the wavelength variable which has higher correlation with the density of the white spirit base wine from the optimally preprocessed spectrum to remove the defects of large calculation amount of a model and low efficiency.
The current common method for measuring the alcoholic strength of the white spirit still depends on density measurement, but the original density detection method has various problems such as low working efficiency, easy uncertainty introduction, high sample consumption, long analysis time and the like. The near infrared spectrum technology (NIR) has the characteristics of high detection speed, high detection precision, simultaneous detection of multiple indexes, on-line monitoring and the like, the near infrared spectrum technology is applied to the detection of the density of white spirit base wine, the optimal pretreatment method and the optimal band screening algorithm are selected in comparison, the function of on-line prediction is realized through the near infrared spectrum technology, the white spirit production process is regulated and controlled in time, the white spirit production quality is improved, and a theoretical basis is provided for the analysis and development of an on-line monitoring system in the white spirit production process.
It should be noted that the embodiments of the present invention can be realized in hardware, software, or a combination of software and hardware. The hardware portion may be implemented using dedicated logic; the software portions may be stored in a memory and executed by a suitable instruction execution system, such as a microprocessor or special purpose design hardware. Those of ordinary skill in the art will appreciate that the apparatus and methods described above may be implemented using computer executable instructions and/or embodied in processor control code, such as provided on a carrier medium such as a magnetic disk, CD or DVD-ROM, a programmable memory such as read only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The device of the present invention and its modules may be implemented by hardware circuitry, such as very large scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., as well as software executed by various types of processors, or by a combination of the above hardware circuitry and software, such as firmware.
The foregoing is merely illustrative of specific embodiments of the present invention, and the scope of the invention is not limited thereto, but any modifications, equivalents, improvements and alternatives falling within the spirit and principles of the present invention will be apparent to those skilled in the art within the scope of the present invention.
Claims (10)
1. A method for detecting the density of white spirit base by utilizing a near infrared spectrum technology is characterized in that the method for detecting the density of white spirit base by utilizing the near infrared spectrum technology comprises the following steps: taking a fen-flavor white spirit base wine as an object, and collecting a near infrared transmission spectrum of the white spirit base wine; detecting the density of the base wine sample by using a national standard method, and detecting and eliminating abnormal values by using a Markov distance method; preprocessing a white spirit base wine sample, and establishing a near infrared rapid analysis model of white spirit base wine density by adopting a partial least squares regression method; and screening the near infrared spectrum of the white spirit base liquor, and screening from the optimally pretreated spectrum to obtain the wavelength variable which has higher correlation with the density of the white spirit base liquor.
2. The method for detecting the density of white spirit base using near infrared spectroscopy according to claim 1, wherein the method for detecting the density of white spirit base using near infrared spectroscopy comprises the steps of:
Step one, white spirit density measurement: detecting the density of the white spirit base liquor sample according to a digital densitometry method;
step two, near infrared spectrum acquisition: collecting the near infrared spectrum of the white spirit base liquor by using a near infrared spectrometer;
step three, eliminating abnormal points: detecting and removing abnormal values of the spectrum data of the sample by using a Markov distance method;
step four, spectrum pretreatment: the method comprises the steps of preprocessing a white spirit base wine sample by adopting a mode of combining first-order guide, convolution smoothing, multi-element scattering correction and first-order guide and multi-element scattering correction;
step five, screening characteristic wave bands: selecting a space partial least square method, an information-free variable elimination method, a continuous projection method and a competitive self-adaptive weighting method, selecting a wavelength variable with higher correlation with the white spirit base wine density from the optimally preprocessed near infrared spectrum, and carrying out predictive analysis on the white spirit base wine density.
Step six, model construction and performance evaluation: adopting partial least square regression method to construct near infrared analysis model of white spirit base density, and adopting decision coefficient R 2 And analyzing and comparing the experimental result, and evaluating the near infrared analysis model effect on the whole.
3. The method for detecting the density of white spirit base by using a near infrared spectrum technology as claimed in claim 2, wherein in the first step, a near infrared transmission spectrum of the white spirit base is collected by using a Matrix-F Fourier transform near infrared spectrometer; preheating a near infrared spectrometer for 1h, and setting instrument parameters; collecting the background in the empty beaker by using a Fourier transform near infrared spectrometer of Matrix-F preheated for 1h, and pouring a white spirit base wine sample into the submerged data collection port; and (3) carrying out spectrum acquisition on the white wine sample by using a Fourier transform near infrared spectrometer of Matrix-F, and adopting spectrum signal acquisition software OPUS7.8 as spectrum data processing software.
4. A method for detecting white spirit base density by near infrared spectroscopy according to claim 3, wherein the parameters of the Matrix-F fourier transform near infrared spectrometer are set as follows: the resolution of the instrument during scanning was 8cm -1 The number of scans was 32, the spectral range was 12000cm wavenumber -1 ～4000cm -1 ；
When the fen-flavor liquor base liquor samples are selected, at least 50 samples are selected from a single-component system; the content range of the selected modeling sample is larger than the range analyzed later; the modeled samples were uniform throughout the content range; the prepared modeled sample concentration was not linearly increased or decreased; the method for measuring the component content can obtain the result.
5. The method for detecting the density of white spirit base by utilizing the near infrared spectrum technology as set forth in claim 2, wherein in the third step, abnormal values are detected and removed by a mahalanobis distance method; when abnormal points are identified by utilizing the mahalanobis distance, dividing a sample set into a correction set and a prediction set, calculating the mahalanobis distance by utilizing the sample spectrum data of the correction set, and judging the abnormal points by a method of setting a threshold value;
wherein, the calculation process of the mahalanobis distance is as follows: dividing sample data into a correction set and a prediction set, and calculating to obtain an average spectrum of a sample; centering the spectrum matrix of the correction set, and subtracting the average spectrum matrix from the spectrum matrix of the correction set to obtain a Markov matrix; calculating to obtain a Markov matrix from the correction set sample to the average spectrum; setting a threshold value, and identifying abnormal points in the correction set sample through the threshold value.
6. The method for detecting the density of white spirit base by utilizing the near infrared spectrum technology as set forth in claim 2, wherein in the sixth step, the spectral data corresponding to the key variables extracted by the interval partial least square method, the non-information variable elimination method, the continuous projection method and the competitive adaptive weighting method and the preprocessed full-band spectral data are used as input data of PLS modeling to perform predictive analysis on the density of white spirit base.
7. A system for detecting the density of white spirit base using near infrared spectrum technology, which applies the method for detecting the density of white spirit base using near infrared spectrum technology according to any one of claims 1 to 6, characterized in that the system for detecting the density of white spirit base using near infrared spectrum technology comprises:
the white spirit density measuring module is used for detecting the density of the white spirit base liquor sample according to a digital densitometry method;
the near infrared spectrum acquisition module is used for acquiring the near infrared spectrum of the white spirit base wine by using a near infrared spectrometer;
the abnormal point removing module is used for detecting and removing abnormal values of the spectrum data of the sample by adopting a mahalanobis distance method;
the spectrum pretreatment module is used for carrying out pretreatment on the white spirit base wine sample by adopting a mode of combining first-order conduction, convolution smoothing, multi-element scattering correction and first-order conduction and multi-element scattering correction;
the characteristic wave band screening module is used for selecting interval partial least square, non-information variable elimination, continuous projection algorithm and competitive self-adaptive weighting algorithm, selecting wavelength variable with higher correlation with the white spirit base density from the optimal preprocessed near infrared spectrum, and carrying out predictive analysis on the white spirit base density.
The model construction and performance evaluation module is used for constructing a near infrared analysis model of the white spirit base density by adopting a partial least square regression method and adopting a decision coefficient R 2 And analyzing and comparing the experimental results, and evaluating the model effect on the whole.
8. A computer device comprising a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to perform the steps of the method of detecting wine base intensity using near infrared spectroscopy as claimed in any one of claims 1 to 6.
9. A computer readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform the steps of the method for detecting white spirit base density using near infrared spectroscopy as claimed in any one of claims 1 to 6.
10. An information data processing terminal, wherein the information data processing terminal is used for realizing the system for detecting the density of white spirit base by utilizing the near infrared spectrum technology according to claim 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310270818.4A CN116242802A (en) | 2023-03-20 | 2023-03-20 | Method for detecting density of white spirit base by utilizing near infrared spectrum technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310270818.4A CN116242802A (en) | 2023-03-20 | 2023-03-20 | Method for detecting density of white spirit base by utilizing near infrared spectrum technology |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116242802A true CN116242802A (en) | 2023-06-09 |
Family
ID=86627799
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310270818.4A Pending CN116242802A (en) | 2023-03-20 | 2023-03-20 | Method for detecting density of white spirit base by utilizing near infrared spectrum technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116242802A (en) |
-
2023
- 2023-03-20 CN CN202310270818.4A patent/CN116242802A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Hu et al. | Rapid detection of three quality parameters and classification of wine based on Vis-NIR spectroscopy with wavelength selection by ACO and CARS algorithms | |
| Urbano-Cuadrado et al. | Near infrared reflectance spectroscopy and multivariate analysis in enology: Determination or screening of fifteen parameters in different types of wines | |
| Liu et al. | Detection of organic acids and pH of fruit vinegars using near-infrared spectroscopy and multivariate calibration | |
| Cuadrado et al. | Comparison and joint use of near infrared spectroscopy and Fourier transform mid infrared spectroscopy for the determination of wine parameters | |
| CN101413885A (en) | Near-infrared spectrum method for rapidly quantifying honey quality | |
| Laghi et al. | FTIR spectroscopy and direct orthogonal signal correction preprocessing applied to selected phenolic compounds in red wines | |
| CN107703097B (en) | Method for constructing model for rapidly predicting crude oil property by using near-infrared spectrometer | |
| CN110726694A (en) | Characteristic wavelength selection method and system of spectral variable gradient integrated genetic algorithm | |
| CN110987846B (en) | Nitrate concentration prediction method based on iPLS-PA algorithm | |
| Cozzolino | Sample presentation, sources of error and future perspectives on the application of vibrational spectroscopy in the wine industry | |
| CN111398213A (en) | Method for judging eligibility of fermented grain model | |
| Cavdaroglu et al. | Detection of vinegar adulteration with spirit vinegar and acetic acid using UV–visible and Fourier transform infrared spectroscopy | |
| Pizarro et al. | Prediction of total and volatile acidity in red wines by Fourier-transform mid-infrared spectroscopy and iterative predictor weighting | |
| Schorn-Garcia et al. | ATR-MIR spectroscopy as a process analytical technology in wine alcoholic fermentation–a tutorial | |
| Ning et al. | Rapid identification and quantitative pit mud by near infrared Spectroscopy with chemometrics | |
| Cavaglia et al. | Detection of bacterial spoilage during wine alcoholic fermentation using ATR-MIR and MCR-ALS | |
| Zhu et al. | Prediction of wood property in Chinese Fir based on visible/near-infrared spectroscopy and least square-support vector machine | |
| CN113030011A (en) | Rapid nondestructive testing method and system for sugar content of fruits | |
| CN116952896A (en) | Near infrared rapid detection method for total nitrogen in saccharified wort of brewery | |
| CN107356556A (en) | A kind of double integrated modelling approach of Near-Infrared Spectra for Quantitative Analysis | |
| CN114624402B (en) | Quality evaluation method for snail rice noodle sour bamboo shoots based on near infrared spectrum | |
| Shen et al. | Evaluation of near‐infrared and mid‐infrared spectroscopy for the determination of routine parameters in Chinese rice wine | |
| CN116242802A (en) | Method for detecting density of white spirit base by utilizing near infrared spectrum technology | |
| Yang et al. | Raman spectroscopy applied to online monitoring of a bioreactor: Tackling the limit of detection | |
| CN113607683A (en) | Automatic modeling method for near infrared spectrum quantitative analysis |
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 |