CN109100315B - Wavelength selection method based on noise-signal ratio - Google Patents
Wavelength selection method based on noise-signal ratio Download PDFInfo
- Publication number
- CN109100315B CN109100315B CN201810953943.4A CN201810953943A CN109100315B CN 109100315 B CN109100315 B CN 109100315B CN 201810953943 A CN201810953943 A CN 201810953943A CN 109100315 B CN109100315 B CN 109100315B
- Authority
- CN
- China
- Prior art keywords
- noise
- spectrum
- signal ratio
- wavelength
- model
- 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.)
- Active
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
- G01N21/25—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
Landscapes
- Physics & Mathematics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses a wavelength selection method based on a noise-signal ratio, which comprises the following steps: (1) providing a noise-to-signal ratio spectrum, which corresponds to the degree of scattering of the full spectrum; (2) based on the spectrum of the noise-signal ratio, a wavelength selection method of the preferential combination of the noise-signal ratio is provided, and the optimal wavelength selection is realized. The method has the advantages of wide professional application range, sufficient information extraction, good prediction effect and the like, and provides an effective solution for the design of the light splitting system in a special analysis instrument.
Description
Technical Field
The invention relates to the technical field of wavelength screening of spectral analysis, in particular to a wavelength selection method based on a noise-signal ratio.
Background
The molecular spectrum mainly comprises ultraviolet-visible, near-infrared, mid-infrared and other spectral regions. With the rapid development of modern measurement techniques and chemometrics, molecular spectroscopy has become an effective technical means for rapid detection of samples. And in particular Near Infrared (NIR) spectra, which reflect the frequency doubled and frequency doubled absorption of the vibrations of the hydrogen-containing functional group X-H of the molecule (e.g., C-H, N-H, O-H, etc.), can be measured for most types of samples without the need for pretreatment (or simple treatment). Therefore, it has an advantage of rapid detection analysis and has been successfully applied to many fields.
Direct measurement of samples has the advantage of being rapid and simple, while overcoming methodological difficulties. In general, for complex analytes with multiple components, wavelength model optimization is a key technique to improve spectral prediction capability in order to overcome noise interference of complex background and other unknown components. However, for a complicated analysis object having a correlation with scattering, such as blood viscosity of blood, it is necessary to extract not only information on a spectral absorption-related wavelength but also information on a spectral scattering-related wavelength, which occurs mainly in a wavelength range where a spectral signal-to-noise ratio is low. The existing wavelength screening method is mainly used for selecting the wavelength with high signal-to-noise ratio and is difficult to be applied to the detection of the analysis object of the type.
Disclosure of Invention
The invention aims to overcome the defects and shortcomings of the prior art, provides a wavelength selection method based on a noise-signal ratio, focuses on quantitative light scattering wavelength information on the basis of absorption wavelength information aiming at a wavelength selection technology of a spectrum (ultraviolet-visible light, near infrared light, intermediate infrared light and the like), realizes the prediction precision of a quantitative analysis model suitable for a complex analysis object with correlation with scattering, and has an industrial prospect in various application fields of spectral analysis.
The purpose of the invention is realized by the following technical scheme: a wavelength selection method based on a noise-signal ratio comprises the following steps:
s1, collecting samples: collecting a sample, and obtaining an index measurement value of the sample by adopting a standard method;
s2, spectrum collection: repeating the measuring of the spectrum of each sample a plurality of times;
s3, obtaining a noise-signal ratio spectrum: the average spectrum [ A ] of each modeled sample was calculated separately by the spectral reproducibility experiment of S2Ave(λ)]Sum standard deviation spectrum [ ASD(λ)](ii) a The noise-to-signal ratio spectrum NSR is further defined as follows:
s4, similarity analysis of the noise-signal ratio spectrum population:
s4-1, calculating the average noise-signal ratio spectrum of the whole modeling samples; in the full-scanning spectrum region, calculating a correlation coefficient between the noise-signal ratio spectrum and the average noise-signal ratio spectrum of each modeling sample; removing a sample with the lowest correlation coefficient;
s4-2 repeating the step of S4-1 for a plurality of times until all correlation coefficients exceed the set threshold R0；
S4-3, calculating the average spectrum of the noise-signal ratio spectrum population obtained in S4-2, and calling the average spectrum as a characteristic noise-signal ratio spectrum; the corresponding value for each wavelength is referred to as the characteristic noise-to-signal ratio for that wavelength;
s5, based on the characteristic noise-signal ratio spectrum, sorting the wavelengths in the full scanning spectrum region from large to small according to the characteristic noise-signal ratio as follows, wherein the total number of the wavelengths is n:
λ1,λ2,…,λn；
s6, sequentially constructing n wavelength combination models based on the preferential combination of the noise-signal ratio as follows:
Ωi＝{λ1,λ2,…,λi},i＝1,2,…,n；
s7, establishing index modeling and inspection models on the n wavelength combination models respectively by adopting the sample spectrum determined by the S2 and adopting a plurality of methods; according to the prediction effect of the model, the optimal wavelength combination model is optimized, the number of the corresponding wavelengths is recorded as N, and the corresponding wavelength combination is as follows:
ΩN＝{λ1,λ2,…,λN}；
the wavelength model is the selected wavelength.
Preferably, the full scan spectral region may be replaced with a characteristic band specified according to a specific object.
Preferably, in step S4-1, the correlation coefficient between the noise-signal ratio spectrum and the average noise-signal ratio spectrum of each modeled sample is calculated, and the correlation coefficient between the noise-signal ratio spectrum and the average noise-signal ratio spectrum of the kth sample is calculated according to the following formula:
m is the number of wavelengths in the full scan spectrum region employed, (A)1,A2,…,Am) Is the average noise-to-signal ratio spectrum, (A)1,k,A2,k,…,Am,k) Is the noise-to-signal ratio spectrum of the kth sample, AAve,AAve,kAre respectively (A)1,A2,…,Am) And (A)1,k,A2,k,…,Am,k) Average value of (a).
Preferably, in step S7, a modeling verification model of the index is established on the n wavelength combination models by a partial least squares method or a multiple linear regression method.
Preferably, in step S2, the measurement of the spectrum of each sample is repeated a plurality of times; all samples are randomly divided into a modeling set and a testing set, and the dividing process can be repeated for M times; establishing a modeling and checking model for each division, and carrying out model optimization according to the average prediction effect; the modeling set is used for establishing a model, and the inspection set is used for inspecting the effect of the model; and calculating the model prediction effect of M divisions.
Preferably, in step S7, the optimal wavelength combination model is selected according to the predicted effect of the model: calculating the model prediction effect of M divisions, and predicting the correlation coefficient R according to the predicted root mean square error RMSEPPAverage value of (RMSEP)Ave，RP,AveStandard deviation RMSEPSD，RP,SDThe comprehensive prediction evaluation index RMSEP is provided+＝RMSEPAve+RMSEPSDAnd the method is used for optimizing an optimal wavelength combination model.
Compared with the prior art, the invention has the following advantages and beneficial effects:
1. the wavelength selection method based on the prior combination of the noise-signal ratios can effectively determine the wavelength range with low spectral signal-to-noise ratio, extracts the information of the near-infrared light scattering associated wavelength, is suitable for detecting complex analysis objects with correlation with scattering, and effectively improves the spectral prediction capability.
2. The invention also has innovation on the spectrum analysis methodology: the wavelength selection method of the noise-signal ratio is a segmented continuous method, and can effectively perform multiband combination, improve the optical characteristics of an analysis object and improve the prediction capability of a spectrum. The method has the advantages of wide professional application range, sufficient information extraction, good prediction effect and the like, and provides an effective solution for the design of the light splitting system in a special analysis instrument.
Drawings
FIG. 1 is a flow chart of an embodiment method.
FIG. 2 is a near infrared spectrum of 300 human peripheral blood samples (male 160, female 140).
Figure 3 is a population of noise to signal ratio spectra of male samples.
Fig. 4 is a population of noise to signal ratio spectra of female samples.
Fig. 5 shows the average signal-to-noise ratio spectra of men and women.
Fig. 6 is a graph of the first N wavelength combinations (for example, N781, low cut whole blood viscosity for male group) based on signal-to-noise ratio preference.
Detailed Description
The present invention will be described in further detail with reference to examples and drawings, but the present invention is not limited thereto.
Example 1
A wavelength selection method based on a noise-signal ratio comprises the following steps:
s1, collecting samples: collecting a sample, and obtaining an index measured value (reference value) of the sample by adopting a standard method;
s2, spectrum collection: repeating the measuring of the spectrum of each sample a plurality of times; all samples are randomly divided into a modeling set and a testing set;
s3, noise to signal ratio spectrum: the average spectrum [ A ] of each modeled sample was calculated separately by the spectral reproducibility experiment of S2Ave(λ)]Sum standard deviation spectrum [ ASD(λ)](ii) a The noise-to-signal ratio spectrum (NSR) is further defined as follows:
s4, similarity analysis of the noise-signal ratio spectrum population:
(1) calculating the average noise-signal ratio spectrum of the whole modeling sample; calculating the correlation coefficient of the noise-signal ratio spectrum and the average noise-signal ratio spectrum of each modeling sample in a full-scanning spectrum region (or according to a characteristic wave band specified by a specific object); and removing a sample with the lowest correlation coefficient. Wherein, the calculation formula of the correlation coefficient of the noise-signal ratio spectrum and the average noise-signal ratio spectrum of the kth sample is as follows:
m is the number of wavelengths in the full scan spectral region (or the characteristic band specified according to the specific object) employed, (A)1,A2,…,Am) Is the average noise-to-signal ratio spectrum, (A)1,k,A2,k,…,Am,k) Is the noise-to-signal ratio spectrum of the kth sample, AAve,AAve,kAre respectively (A)1,A2,…,Am) And (A)1,k,A2,k,…,Am,k) Average value of (a).
(2) Repeating the step (1) for a plurality of times until all the correlation coefficients exceed the set threshold value R0(such as R)00.900), such a population of noise-to-signal ratio spectra has good similarity.
(3) Calculating the average spectrum of the noise-signal ratio spectrum population obtained in the step (2), wherein the average spectrum effectively reflects the common characteristic of the noise-signal ratio spectrum population and is called as a characteristic noise-signal ratio spectrum; the corresponding value for each wavelength is referred to as the characteristic noise to signal ratio for that wavelength.
S5, based on the characteristic noise-signal ratio spectrum, the wavelengths (total number of wavelengths: n) of the full scanning spectrum region (or the characteristic wave band designated according to the specific object) are sorted from big to small according to the characteristic noise-signal ratio as follows:
λ1,λ2,…,λn；
s6, sequentially constructing n wavelength combination models based on the preferential combination of the noise-signal ratio as follows:
Ωi＝{λ1,λ2,…,λi},i＝1,2,…,n；
s7, establishing index modeling test models on the n wavelength combination models respectively by adopting the sample spectrum measured by the S2 and adopting methods such as Partial Least Squares (PLS) and Multiple Linear Regression (MLR); predicting the correlation coefficient R based on the predicted effect of the model (e.g. predicting the root mean square error RMSEPPEtc.) to select the optimal wavelength combination model, whichThe number of corresponding wavelengths is recorded as N, and the corresponding wavelength combination is:
ΩN＝{λ1,λ2,…,λN}；
the wavelength model achieves the best prediction effect.
Example 2
Blood is a complex analyte, and blood viscosity (called blood viscosity for short) is an important clinical diagnosis index for various diseases such as cardiovascular and cerebrovascular diseases. Blood viscosity monitoring is one of the important means for preventing and controlling cardiovascular and cerebrovascular diseases. The blood viscosity needs to be detected clinically by adopting a hemorheology method, and the method is complex and inconvenient for screening by large crowds.
The blood viscosity is closely related with the deformability and aggregability of blood erythrocytes, and thus indirectly related to hemoglobin; hemoglobin has a distinct near infrared absorption characteristic. On the other hand, when near infrared light enters a viscous whole blood sample, scattering occurs, and the scattering degree is related to blood viscosity; therefore, near infrared spectroscopy has the basis for quantitative analysis of blood viscosity. The screening of characteristic wavelength is the core problem of blood viscosity near infrared analysis. The existing wavelength selection method mainly aims at the selection of hemoglobin information wavelength, and is lack of research and adoption of noise wavelength caused by light scattering. It directly affects the accuracy of near infrared spectroscopic analysis of blood viscosity.
Low-cut, medium-cut and high-cut whole blood viscosity are blood viscosity parameters with the most clinical reference value. The present embodiment illustrates the applicability of the proposed wavelength selection method based on the preferential combination of noise-to-signal ratios by taking the near infrared spectrum quantitative analysis of three indicators of low-cut, medium-cut and high-cut whole blood viscosity as an example. The wavelength selection method based on the preferential combination of the noise-signal ratio is more suitable for blood viscosity detection by comparing the existing effective wavelength screening methods, namely moving window partial least squares (MW-PLS), competitive adaptive re-weighted sampling PLS (CARS-PLS) and continuous projection algorithm PLS (SPA-PLS). But the embodiments of the present invention are not limited thereto.
The implementation steps are as follows:
s1, collecting samples: 300 human peripheral blood samples (male 160, female 140) were collected in the same hospital, and blood viscosity parameters (low cut, medium cut, high cut whole blood viscosity) of the samples were determined by clinical standard methods;
s2, spectrum collection: the spectrum of each sample was measured in duplicate, as shown in fig. 2; as the number of red blood cells of the male and the female is different, the male and the female are modeled in groups in order to avoid the interference of different groups, improve the sample homogeneity and the prediction accuracy. Further, in order to achieve a stable and objective modeling effect, all samples were randomly divided into a modeling set (male 80, female 70) and a testing set (male 80, female 70), and this division process was repeated 10 times. And establishing a modeling and checking model for each division, and carrying out model optimization according to the average prediction effect. The modeling set is used for establishing a model, and the inspection set is used for inspecting the effect of the model.
S3, noise to signal ratio spectrum: the average spectrum [ A ] of each modeled sample was calculated separately by the spectral reproducibility experiment of S2Ave(λ)]Sum standard deviation spectrum [ ASD(λ)](ii) a Calculating a noise-signal ratio spectrum (NSR) of each modeling sample according to the formula (1);
s4, similarity analysis of the noise-signal ratio spectrum population:
(1) calculating the average noise-signal ratio spectrum of the whole modeling sample; calculating a correlation coefficient between the noise-signal ratio spectrum and the average noise-signal ratio spectrum of each modeling sample according to a formula (2) by adopting a spectrum region (1900-2000nm) near a water absorption peak; and removing a sample with the lowest correlation coefficient.
(2) The step of (1) is repeated a plurality of times until all correlation coefficients satisfy the threshold, such that the population of spectra with good similarity noise-to-signal ratio is shown in fig. 3 and 4.
(3) Calculating the average spectrum of the noise-signal ratio spectrum population obtained in the step (2), namely the characteristic noise-signal ratio spectrum, as shown in fig. 5; the corresponding value for each wavelength is referred to as the characteristic noise to signal ratio for that wavelength.
S5, based on the characteristic noise-signal ratio spectrum, sorting the wavelengths (total wavelength: n: 1050) of the full scanning spectrum region (400-:
λ1,λ2,…,λ1050；
s6, sequentially constructing 1050 wavelength combination models based on the preferential combination of the noise-signal ratio as follows:
Ωi＝{λ1,λ2,…,λi},i＝1,2,…,1050；
s7, establishing a modeling test model of the index on the n wavelength combination models respectively by adopting the sample spectrum measured by the S2 and a Partial Least Squares (PLS); calculating the model prediction effect (predicting root mean square error RMSEP, predicting correlation coefficient R) of 10 divisionsP) Average value of (RMSEP)Ave，RP,Ave) Standard deviation (RMSEP)SD，RP,SD) The comprehensive prediction evaluation index RMSEP is provided+(RMSEP+＝RMSEPAve+RMSEPSD) And the number of the corresponding wavelengths is recorded as N. FIG. 6 is a graph showing the first N wavelength combinations based on the priority of the noise-to-signal ratio (taking the male group, low cut whole blood viscosity, N781 as an example, corresponding to the wavelength combinations: 598-&2068-2346nm)。
At the same time, three important wavelength selection methods: and comparing a continuous moving window partial least square method (MW-PLS), a discrete competitive adaptive re-weighted sampling PLS (CARS-PLS) and a continuous projection algorithm PLS (SPA-PLS), and obtaining results shown in tables 1-3.
TABLE 1 modeling Effect of Low cut Whole blood viscosity for Male and female sample groups based on the NSRP-PLS method and other 3 methods
TABLE 2 modeling Effect of cut Whole blood viscosity in Male and female sample groups based on the NSRP-PLS method and other 3 methods
TABLE 3 modeling Effect of high-cut Whole blood viscosity for Male and female sample groups based on the NSRP-PLS method and other 3 methods
As can be seen from tables 1-3, the prediction effect obtained by the NSRP-PLS method is obviously superior to that obtained by the other three methods for the three indexes of low-cut, medium-cut and high-cut whole blood viscosity.
The prediction effect of the noise-signal ratio priority partial least square method (NSRP-PLS) based on the invention is obviously superior to the prior three effective wavelength screening methods (MW-PLS, CARS-PLS and SPA-PLS). It is worth noting that the NSRP-PLS method selects the wavelength according to the noise-signal ratio, an optimized segmented continuous wavelength model can be formed naturally, and the continuous wavelength screening method (MW-PLS) cannot be realized directly, so that the wavelength screening mode and the application range are widened, and the method has important reference values in the aspects of establishing a high-precision analysis model, designing a special spectrum instrument and the like.
The above embodiments are preferred embodiments of the present invention, but the present invention is not limited to the above embodiments, and any other changes, modifications, substitutions, combinations, and simplifications which do not depart from the spirit and principle of the present invention should be construed as equivalents thereof, and all such changes, modifications, substitutions, combinations, and simplifications are intended to be included in the scope of the present invention.
Claims (6)
1. A wavelength selection method based on a noise-signal ratio is characterized by comprising the following steps:
s1, collecting samples: collecting a sample, and obtaining an index measurement value of the sample by adopting a standard method;
s2, spectrum collection: repeating the measuring of the spectrum of each sample a plurality of times;
s3, obtaining a noise-signal ratio spectrum: the average spectrum A of each modeled sample was calculated separately by the spectral reproducibility experiment of S2Ave(lambda) and standard deviation spectrum ASD(λ); the noise-to-signal ratio spectrum NSR is further defined as follows:
s4, similarity analysis of the noise-signal ratio spectrum population:
s4-1, calculating the average noise-signal ratio spectrum of the whole modeling samples; in the full-scanning spectrum region, calculating a correlation coefficient between the noise-signal ratio spectrum and the average noise-signal ratio spectrum of each modeling sample; removing a sample with the lowest correlation coefficient;
s4-2 repeating the step of S4-1 for a plurality of times until all correlation coefficients exceed the set threshold R0；
S4-3, calculating the average spectrum of the noise-signal ratio spectrum population obtained in S4-2, and calling the average spectrum as a characteristic noise-signal ratio spectrum; the corresponding value for each wavelength is referred to as the characteristic noise-to-signal ratio for that wavelength;
s5, based on the characteristic noise-signal ratio spectrum, sorting the wavelengths in the full scanning spectrum region from large to small according to the characteristic noise-signal ratio as follows, wherein the total number of the wavelengths is n:
λ1，λ2，…，λn；
s6, sequentially constructing n wavelength combination models based on the preferential combination of the noise-signal ratio as follows:
Ωi＝{λ1，λ2，…，λi}，i＝1，2，…，n；
s7, establishing a modeling inspection model of indexes on the n wavelength combination models respectively by adopting a plurality of methods according to the sample spectrum determined by the S2; according to the prediction effect of the model, the optimal wavelength combination model is optimized, the number of the corresponding wavelengths is recorded as N, and the corresponding wavelength combination is as follows:
ΩN＝{λ1，λ2，…，λN}；
the wavelength model is the selected wavelength.
2. The method of noise-to-signal ratio based wavelength selection according to claim 1, wherein the full scan spectral region is replaced with a characteristic band specified according to a specific object.
3. The method for selecting a wavelength according to claim 1, wherein in step S4-1, the correlation coefficient between the snr spectra and the average snr spectra of each modeled sample is calculated, and the correlation coefficient between the snr spectra and the average snr spectra of the kth sample is calculated as follows:
m is the number of wavelengths of the employed full scan spectral region, A1，A2，…，AmIs an average noise-to-signal ratio spectrum, A1，k，A2，k，…，Am，kIs the noise-to-signal ratio spectrum of the kth sample, AAve，AAve，kAre respectively A1，A2，…，AmAnd A1，k，A2，k，…，Am，kAverage value of (a).
4. The method for selecting a wavelength according to claim 1, wherein in step S7, a modeling verification model of the index is established on the n wavelength combination models by using a partial least squares method or a multiple linear regression method.
5. The method for selecting a wavelength according to claim 1, wherein in step S2, the measurement of the spectrum of each sample is repeated a plurality of times; all samples are randomly divided into a modeling set and a testing set, and the dividing process is repeated for M times; establishing a modeling inspection model for each division, and carrying out model optimization according to the average prediction effect; the modeling set is used for establishing a model, and the inspection set is used for inspecting the effect of the model; and calculating the model prediction effect of M divisions.
6. The method for selecting a wavelength according to claim 5, wherein in step S7, based on the predicted effect of the model, the optimal wavelength combination model is selected: calculating the model prediction effect of M divisions, and predicting the correlation coefficient R according to the predicted root mean square error RMSEPPAverage value of (RMSEP)Ave，RP，AveStandard deviation RMSEPSD，RP，SDThe comprehensive prediction evaluation index RMSEP is provided+＝RMSEPAve+RMSEPSDAnd the method is used for optimizing an optimal wavelength combination model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810953943.4A CN109100315B (en) | 2018-08-21 | 2018-08-21 | Wavelength selection method based on noise-signal ratio |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810953943.4A CN109100315B (en) | 2018-08-21 | 2018-08-21 | Wavelength selection method based on noise-signal ratio |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109100315A CN109100315A (en) | 2018-12-28 |
| CN109100315B true CN109100315B (en) | 2020-11-13 |
Family
ID=64850540
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810953943.4A Active CN109100315B (en) | 2018-08-21 | 2018-08-21 | Wavelength selection method based on noise-signal ratio |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109100315B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110208248B (en) * | 2019-06-28 | 2021-11-19 | 南京林业大学 | Method for identifying abnormal measurement signal of Raman spectrum |
| CN111474128B (en) * | 2020-05-12 | 2023-03-14 | 暨南大学 | Spectral wavelength combination method based on spectral separation degree |
| CN111474124B (en) * | 2020-05-29 | 2023-04-07 | 暨南大学 | Spectral wavelength selection method based on compensation |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5846190A (en) * | 1995-10-10 | 1998-12-08 | Hewlett-Packard Company | Method of and apparatus for recognizing falsified pulse oximetry measurements |
| CN1900693A (en) * | 2006-07-17 | 2007-01-24 | 北京联合大学应用文理学院 | Detecting method for monacolin compound content in red yeast vinegar |
| JP2008203234A (en) * | 2007-02-23 | 2008-09-04 | Matsushita Electric Works Ltd | Blood component concentration analysis method and device |
| WO2013165888A2 (en) * | 2012-04-30 | 2013-11-07 | Mayo Foundation For Medical Education And Research | Spectrometric systems and methods for improved focus localization of time-and space-varying measurements |
| EP2899533A1 (en) * | 2014-01-22 | 2015-07-29 | Siemens Aktiengesellschaft | Wavelength modulation spectroscopy method with a filter for the demodulated measurement signal and the simulated signal |
| CN106092893A (en) * | 2016-08-17 | 2016-11-09 | 暨南大学 | A kind of wavelength method for optimizing of spectrum discriminant analysis |
| CN106153561A (en) * | 2016-06-21 | 2016-11-23 | 中南大学 | The many metal ion inspections of uv-vis spectra based on wavelength screening |
| CN106418635A (en) * | 2015-08-12 | 2017-02-22 | 福建武夷烟叶有限公司 | Equipment and process for reducing threshing and redrying nicotine variation coefficient |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA2642722A1 (en) * | 2006-02-16 | 2007-08-30 | Clean Earth Technologies, Llc | Method for spectral data classification and detection in diverse lighting conditions |
-
2018
- 2018-08-21 CN CN201810953943.4A patent/CN109100315B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5846190A (en) * | 1995-10-10 | 1998-12-08 | Hewlett-Packard Company | Method of and apparatus for recognizing falsified pulse oximetry measurements |
| CN1900693A (en) * | 2006-07-17 | 2007-01-24 | 北京联合大学应用文理学院 | Detecting method for monacolin compound content in red yeast vinegar |
| JP2008203234A (en) * | 2007-02-23 | 2008-09-04 | Matsushita Electric Works Ltd | Blood component concentration analysis method and device |
| WO2013165888A2 (en) * | 2012-04-30 | 2013-11-07 | Mayo Foundation For Medical Education And Research | Spectrometric systems and methods for improved focus localization of time-and space-varying measurements |
| EP2899533A1 (en) * | 2014-01-22 | 2015-07-29 | Siemens Aktiengesellschaft | Wavelength modulation spectroscopy method with a filter for the demodulated measurement signal and the simulated signal |
| CN106418635A (en) * | 2015-08-12 | 2017-02-22 | 福建武夷烟叶有限公司 | Equipment and process for reducing threshing and redrying nicotine variation coefficient |
| CN106153561A (en) * | 2016-06-21 | 2016-11-23 | 中南大学 | The many metal ion inspections of uv-vis spectra based on wavelength screening |
| CN106092893A (en) * | 2016-08-17 | 2016-11-09 | 暨南大学 | A kind of wavelength method for optimizing of spectrum discriminant analysis |
Non-Patent Citations (3)
| Title |
|---|
| Practicality of wavelength selection to improve signal-to-noise ratio in near-infrared spectroscopy;Hiroki Sato et al.;《NeuroImage》;20041231;第21卷;1554-1562页 * |
| Waveband selection of reagent-free determination for thalassemia screening indicators using Fourier transform infrared spectroscopy with attenuated total reflection;Xiaoli Long et al.;《Journal of Biomedical Optics》;20140819;第19卷(第8期);087004-1至087004-11页 * |
| 可见-近红外光谱测定血红蛋白的等效波段选择;刘振尧等;《光学精密工程》;20121031;第20卷(第10期);2170-2175页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109100315A (en) | 2018-12-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN100590417C (en) | Plant blade physiologic index non-destructive test method | |
| CN109100315B (en) | Wavelength selection method based on noise-signal ratio | |
| Poon et al. | Quantitative reagent-free detection of fibrinogen levels in human blood plasma using Raman spectroscopy | |
| CN101620180B (en) | Method for rapidly detecting tea quality through near infrared technology | |
| CN103776777B (en) | Method for identifying ginsengs with different growth patterns by using near infrared spectrum technology and determining content of components in ginsengs | |
| CN107703097B (en) | Method for constructing model for rapidly predicting crude oil property by using near-infrared spectrometer | |
| CN106596456A (en) | Changeable moving window based selection method of near infrared spectral region of solution | |
| CN106018331B (en) | The method for estimating stability and pretreatment optimization method of multi-channel spectral system | |
| CN109211829A (en) | A method of moisture content in the near infrared spectroscopy measurement rice based on SiPLS | |
| CN110749565A (en) | Method for rapidly identifying storage years of Pu' er tea | |
| CN102937575B (en) | Watermelon sugar degree rapid modeling method based on secondary spectrum recombination | |
| CN106950192A (en) | A kind of method of Contents of Main Components quick detection in vegetable protein beverage based on near-infrared spectral analysis technology | |
| CN111537469A (en) | Apple quality rapid nondestructive testing method based on near-infrared technology | |
| CN101788459B (en) | Quasi-continuous spectroscopic wavelength combination method | |
| CN115372309A (en) | Associated effective characteristic spectrum selection method and oil product index content quick detection method | |
| CN110672578A (en) | Model universality and stability verification method for polar component detection of frying oil | |
| CN109001182B (en) | Raman spectrum nondestructive testing method for alcohol content in closed container | |
| CN104020124B (en) | Based on absorbance light splitting wavelength screening technique preferentially | |
| CN107247033B (en) | Identify the method for Huanghua Pear maturity based on rapid decay formula life cycle algorithm and PLSDA | |
| CN105806803A (en) | Multi-index collaborative analysis wavelength combination and selection method thereof | |
| CN113030011A (en) | Rapid nondestructive testing method and system for sugar content of fruits | |
| CN108489928B (en) | Method for detecting textile fiber components by short-wave infrared extinction spectrum | |
| WO2023123329A1 (en) | Method and system for extracting net signal in near-infrared spectrum | |
| CN109520950A (en) | The insensitive chemical component spectral method of detection of a kind of pair of spectral shift | |
| CN112326574B (en) | Spectrum wavelength selection method based on Bayesian classification |
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 |