CN115049025A - Model migration method and system based on elastic segmentation standardization algorithm - Google Patents
Model migration method and system based on elastic segmentation standardization algorithm Download PDFInfo
- Publication number
- CN115049025A CN115049025A CN202210977891.0A CN202210977891A CN115049025A CN 115049025 A CN115049025 A CN 115049025A CN 202210977891 A CN202210977891 A CN 202210977891A CN 115049025 A CN115049025 A CN 115049025A
- Authority
- CN
- China
- Prior art keywords
- model
- host
- data
- spectral
- satellite
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Abstract
The invention provides a model migration method and a system based on an elastic segmentation standardization algorithm, which relate to the field of spectral analysis, and the method comprises the steps of firstly obtaining spectral data collected by a host and a satellite machine as samples and determining the wavelength range of the spectral data; carrying out full spectrum analysis and normalization operation on the spectral data; dividing each refraction spectrum data into a training data set and a verification data set; establishing a quantitative analysis model by using the host spectral data; setting the initial width of a window; training a multiple regression model; recording the spectral lines of the satellite machine and the host machine after the verification set is corrected; selecting an optimal window width according to the RMSE of the verification set to establish a conversion matrix; converting the spectral data of the satellite machine through a conversion matrix; and taking the converted satellite machine spectrum data as the input of a host machine model, and predicting by using the host machine model. The method reduces the RMSE between the spectral lines of the peak positions and improves the accuracy of model migration.
Description
Technical Field
The invention relates to the field of spectral analysis, in particular to a model migration method and system based on an elastic segmentation standardization algorithm.
Background
In recent years, with the improvement of the manufacturing level of optical spectrum instruments and the development of optical spectrum analysis algorithms and software, the laser-induced breakdown spectroscopy technology is rapidly developed. The laser-induced breakdown spectroscopy technology is benefited, so that nondestructive, multi-element simultaneous detection and rapid real-time detection can be performed on a sample, and the laser-induced breakdown spectroscopy technology is widely applied to a plurality of fields such as soil detection, coal mining, metallurgical analysis, mineral development, biomedicine and the like. Practice has proved that the method of chemometrics, neural network, machine learning and the like is applied to data processing of laser-induced breakdown spectroscopy, so that the quality and efficiency of product detection can be greatly improved, the production cost and energy consumption are reduced, and the laser-induced breakdown spectroscopy technology becomes an important detection and analysis means. Due to the defects of noise, baseline drift, self-absorption effect and the like existing in laser-induced breakdown spectroscopy data, quantitative or qualitative analysis can be carried out only by means of a stable model.
However, establishing a stable and reliable high-quality laser-induced breakdown spectroscopy analysis model usually requires a large number of spectral samples, which is costly, time-consuming and labor-consuming, and the established model is difficult to maintain. In actual production, the spectrum measured by a spectrometer is influenced by many factors, such as different environments and different instruments and processing machines, which result in certain differences between spectra, for example: the spectral model of the previous batch of samples cannot be used for data analysis of the next batch of samples; models built from the same sample cannot be shared between different instruments. These problems have severely limited the development and application of laser-induced breakdown spectroscopy. Therefore, establishing a theoretical and systematic method for high quality transfer between models for sharing between spectral models has become an important research work for laser-induced breakdown spectroscopy.
In the existing laser-induced breakdown spectroscopy technology, a mapping is found by establishing a relation between a host and a satellite spectrum, so that the spectrum of a satellite instrument is converted into spectrum data which can use a host model, and the effect that a plurality of instrument spectrums share one model is realized. In actual production, the spectrometer is changed along with scene change, environmental conditions change, instrument aging and the like, so that the laser-induced breakdown spectroscopy analysis model established before needs regular maintenance.
At present, some achievements have been achieved on a laser-induced breakdown spectroscopy model transfer technology, but in actual production, the model transfer technology is not widely used, the root cause is the prediction precision after model transfer, and the stability of the model still needs to be improved. It can be expected that the green, rapid and pollution-free laser-induced breakdown spectroscopy technology will be widely applied to various aspects of life today when the portable spectrometer is rapidly developed, and the research on the model migration method still has wide prospects.
The establishment of an analysis model is the basis of quantitative or qualitative analysis by the laser-induced breakdown spectroscopy technology, however, the establishment of a stable and reliable spectral analysis model consumes huge manpower and material resources, and the acquisition and processing of spectra consume a lot of time and energy. However, when the model is built, the model can only be used for measuring a current period of time or a single sample, which is not acceptable in practical production. The problem of commonality between models has severely hampered the development of laser-induced breakdown spectroscopy techniques. In the existing situation, there are two main reasons for the "failure" of the model, which are as follows:
1) a change in the sample to be tested. After the high-quality labeled data is used for learning and establishing an analysis model, samples to be analyzed have some differences among different batches of samples due to time differences and different environments, and if the original model is directly used for analyzing the existing samples without correction and optimization, a large error occurs, so that the model is invalid.
2) The apparatus differs. In general spectrometer manufacturers, not only one spectrometer of the same type is produced, but also the mass production is often performed. Due to slight differences in the manufacturing process and the mechanical structure, an analytical model cannot be used universally for instruments manufactured in the same model. In some research units, prediction errors are caused by different purchase time of instruments or different manufacturers purchasing the same model.
Disclosure of Invention
To solve the above problems, the present invention proposes the idea of model migration, namely: and the model can be improved and maintained under the condition of only a small amount of newly measured samples, so that the pursuit goal of model migration is realized.
The invention provides a model migration method based on an elastic segmentation standardization algorithm, aiming at improving the precision of model migration and reducing RMSE between spectral lines of a peak position.
The method comprises the following steps:
step 1: acquiring spectral data acquired by a host and a satellite as samples, and determining the wavelength range of the spectral data;
step 2: carrying out full spectrum analysis and normalization operation on the spectral data;
and 3, step 3: dividing each refraction spectrum data into a training data set and a verification data set;
and 4, step 4: establishing a quantitative analysis model by using the host spectral data;
and 5: setting the initial width of a window as WL and the maximum width as WLmax;
step 6: training a multivariate regression model by taking the spectrum of the training set of the satellite machine as input and the corresponding spectral line of the host machine as output;
and 7: recording RMSE of the satellite spectral line and the host spectral line after the verification set is corrected, wherein the window width is = width + 1;
and 8: repeating the step 5 to the step 6 until the width = WLmax;
and step 9: selecting an optimal window width according to the RMSE of the verification set to establish a conversion matrix;
step 10: converting the spectral data of the satellite machine through a conversion matrix;
step 11: and taking the converted satellite machine spectrum data as the input of a host machine model, and predicting by using the host machine model.
Preferably, each sample is subjected to a plurality of spectral data acquisitions, and an average value calculation is performed on the plurality of spectral data in the sample.
Preferably, the initial width is set to WL to realize the conversion of the satellite spectrum data.
Preferably, a multiple linear regression is used to represent the relationship between the satellite spectral data and the host spectral data.
Preferably, the window width is automatically adjusted in response to changes in the position of the spectral line.
Preferably, a functional relationship of the spectral data is established between the host and the satellite machine in a segmented manner, and the spectral data of the satellite machine is corrected by the formula:
in the formula:R i is the first of the spectral data of the hostiLine intensity values of the lines, window widths ofd，DAt the window widthd(ii) the spectrum of (a);f i to correspond toiThe transfer function of the line.
The invention also provides a model migration system based on the elastic segmentation standardization algorithm, which comprises the following components: a migration terminal, a host and a satellite;
further, the host computer is used for establishing a quantitative analysis model by using the spectral data;
furthermore, the satellite machine is used for a laser-induced spectrum acquisition system on each practical application scene, and can acquire related spectrum data according to requirements;
furthermore, the migration terminal is used for acquiring the spectral data acquired by the host computer and the satellite computer and determining the wavelength range of the spectral data;
carrying out full spectrum analysis and normalization operation on the spectrum data;
dividing each refraction spectrum data into a training data set and a verification data set;
setting the initial width of a window as WL and the maximum width as WLmax;
training a multivariate regression model by taking the spectrum of the training set of the satellite machine as input and the corresponding spectral line of the host machine as output;
recording RMSE of the spectral lines of the satellite machine and the host machine after the verification set is corrected, and enabling the window width to be width = width +1 so that the width is = WLmax;
selecting an optimal window width according to the RMSE of the verification set to establish a conversion matrix;
converting the spectral data of the satellite machine through a conversion matrix;
and taking the converted satellite machine spectrum data as the input of a host machine model, and predicting by using the host machine model.
Furthermore, a functional relation of the spectral data is established between the host and the satellite machine in a segmented mode, and the spectral data of the satellite machine is corrected by the formula:
in the formula:R i is the first of the spectral data of the hostiBy lines of the linesLine intensity value, window width isd，DIs at the width of the windowd(ii) the spectrum of (a);f i to correspond toiThe transfer function of the line.
According to the technical scheme, the invention has the following advantages:
the method disclosed by the invention has the advantages that the quantitative analysis model of the LIBS is migrated based on the elastic segmentation standardization algorithm, the RMSE between the spectral lines of the peak positions is reduced, and the model migration accuracy is improved.
The method and the segmented direct standardization algorithm are used for expressing on PLS and LR + SUAC + BPNN quantitative analysis models. The satellite data converted by the FPDS has the advantages that the prediction accuracy is greatly improved, RMSE among spectral lines and prediction accuracy of a migration model are combined, and compared with PDS, the FPDS can convert wave peaks with higher accuracy.
Drawings
In order to more clearly illustrate the technical solution of the present invention, the drawings used in the description will be briefly introduced, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained based on these drawings without creative efforts.
FIG. 1 is a flow chart of a model migration method based on an elastic segmentation normalization algorithm.
FIG. 2 is a schematic diagram of a model migration system based on an elastic segment normalization algorithm.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention aims to solve the problem that LIBS equipment of different production lines share a trained quantitative analysis model, namely, the problem that samples to be analyzed have some differences due to time differences and different environments after the analysis model is established by learning with high-quality labeled data is solved. The method avoids the phenomenon that the original model is directly used for analyzing the existing sample without correction and optimization, so that larger errors occur and the model is invalid.
The model migration method based on the elastic segmentation standardization algorithm can also solve the problem that due to small differences in processing technology and mechanical structure, instruments produced in the same model cannot be commonly used by one analysis model, so that prediction errors are caused.
Specifically, as shown in fig. 1, the model migration method based on the elastic segment normalization algorithm provided by the present invention relates to elastic segment direct normalization, and each spectral line based on the spectral data of the host device has a certain functional relationship with other spectral lines near the corresponding spectral line of the satellite device. The host computer and the satellite machine establish the functional relation of the spectral data in a sectional mode, and further correct the spectral data of the satellite machine, and the formula is as follows:
in the formula:Riis the first of the spectral data of the hostiLine intensity values of the lines, window widths ofd，DIs at the width of the windowd(ii) the spectrum of (a);f i to correspond toiThe transfer function of the line. The size of the window is often smaller than the number of transferred samples.
The transfer function for each spectral line is established by traversing all characteristic spectral lines of the spectral data collected on the host device through a sliding window. Through the transfer function, the standardization of the spectral data collected on the satellite equipment can be realized. In order to improve the fitting accuracy of the spectral line as much as possible, therefore, the width of the sliding window is made variable. The RMSE between the converted satellite spectral data and the host spectral data is reduced by varying the width of the sliding window. The width of the corresponding window is determined by observing the RMSE of each spectral line of the validation set spectrum.
That is, the method of the present invention is implemented by sklern function package in Python.
The model migration method based on the elastic segmentation standardization algorithm comprises the following steps:
(1) spectral data collected by the host and the satellite are acquired as samples, and a wavelength range of the spectral data is determined.
Illustratively, the system may be configured with 38 samples, and spectral matrices of 342 × 8188 size are obtained on two devices, respectively, with 9 spectral data for each sample. And averaging the 9 pieces of spectral data acquired by each sample, and finally obtaining a 38X 8188 spectral data matrix on each device by 38 samples.
(2) And performing full spectrum and normalization operation on the spectral data. Dividing the original spectrum data by the average value of the full spectrum to obtain normalized spectrum data;
(3) dividing each refraction spectrum data into a training set and a verification set;
(4) establishing a quantitative analysis model by using the host spectral data;
(5) setting the initial width of a window as WL and the maximum width as WLmax;
(6) training a multivariate regression model by taking the spectrum of the training set of the satellite machine as input and the corresponding spectral line of the host machine as output;
(7) recording RMSE of the satellite spectral line and the host spectral line after the verification set is corrected, wherein the window width is = width + 1;
(8) repeating the steps (5) to (6) until width = WLmax;
(9) selecting an optimal window width according to RMSE of the verification set to establish a conversion matrix;
(10) converting the spectral data of the satellite machine through a conversion matrix;
(11) and taking the converted satellite machine spectrum data as the input of a host machine model, and predicting by using the host machine model.
The method disclosed by the invention has the advantages that the quantitative analysis model of the LIBS is migrated based on the elastic segmentation standardization algorithm, the RMSE between the spectral lines of the peak positions is reduced, and the model migration accuracy is improved.
The invention also carries out result verification on the model migration method based on the elastic segmentation standardization algorithm: the invention adopts the method and the segmented direct standardization algorithm to obtain the comparison result of the predicted value and the true value of the 5-fold cross validation.
Table 1 compares the performance of the method of the invention with the piecewise direct normalization algorithm on the PLS and LR + SUAC + BPNN quantitative analysis models. The satellite data converted by the FPDS has the advantages that the prediction accuracy is greatly improved, RMSE among spectral lines and prediction accuracy of a migration model are combined, and compared with PDS, the FPDS can convert wave peaks with higher accuracy.
TABLE 1 analysis results from the method of the invention and the piecewise direct normalization algorithm
In the embodiment, the raw materials of the metallurgical slag are adopted, the method is only a preferred embodiment, and the analysis can be carried out according to different application objects during specific implementation, so that the window width can be adjusted.
Of course, for the above-described model migration method based on flexible segment normalization algorithm, those skilled in the art will appreciate that the elements and algorithm steps of the examples described in connection with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both, and that the components and steps of the examples have been described generally in terms of function in the above description for clarity of illustrating interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the technical solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
The block diagrams shown in the figures of the method and system are only functional entities and do not necessarily correspond to physically separate entities. I.e. these functional entities may be implemented in the form of software, or in one or more hardware modules or integrated circuits, or in different networks and/or processor means and/or microcontroller means.
The invention also provides a model migration system based on the elastic segmentation standardization algorithm, as shown in fig. 2, the system comprises: a migration terminal, a host and a satellite;
the mobile terminal, the host computer and the satellite machine can be connected through network communication. That is, the network is the medium used to provide the communication link between the migration terminal, the host, and the satellite. The network may include various connection types, such as wired, wireless communication links, or fiber optic cables, to name a few.
And the migration terminal is respectively in communication connection with the host and the satellite. The migration terminal receives the information of the host computer and the satellite computer and can also send the information to the host computer and the satellite computer.
The migration terminal may be a variety of electronic devices having a display screen including, but not limited to, smart phones, tablet computers, laptop and desktop computers, and the like.
In the model migration system based on the elastic segment normalization algorithm, for example, a model migration generation method such as a machine learning method or a deep learning method may be used. The host is used for establishing a quantitative analysis model by using the spectral data;
the migration terminal is used for acquiring spectral data acquired by the host and the satellite as samples and determining the wavelength range of the spectral data;
carrying out full spectrum analysis and normalization operation on the spectral data; dividing each refraction spectrum data into a training data set and a verification data set;
setting the initial width of a window as WL and the maximum width as WLmax;
training a multivariate regression model by taking the spectrum of the training set of the satellite machine as input and the corresponding spectral line of the host machine as output;
recording RMSE of the spectrum lines of the satellite machine and the host machine after the correction of the verification set, wherein the window width is = width +1, and the width is = WLmax;
selecting an optimal window width according to the RMSE of the verification set to establish a conversion matrix;
converting the spectral data of the satellite machine through a conversion matrix;
and taking the converted satellite machine spectrum data as the input of a host machine model, and predicting by using the host machine model.
The host and the satellite set establish a functional relation of the spectral data in a segmented mode, and the spectral data of the satellite set are corrected, and the formula is as follows:
in the formula:Riis the first of the spectral data of the hostiLine intensity values of the lines, window widths ofd，DIs at the width of the windowd(ii) the spectrum of (a);f i to correspond toiThe transfer function of the line.
The model migration system based on the elastic segmentation standardization algorithm provided by the invention performs the migration of the quantitative analysis model of the LIBS based on the elastic segmentation standardization algorithm, reduces the RMSE between the spectral lines of the peak positions and improves the accuracy of the model migration.
For the flexible segment normalization algorithm based model migration method and system provided by the present invention, the units and algorithm steps of each example described in connection with the embodiments disclosed herein can be implemented as electronic hardware, computer software, or a combination of both, and in the above description, the components and steps of each example have been generally described in terms of functions in order to clearly illustrate the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (8)
1. A model migration method based on an elastic segmentation standardization algorithm is characterized by comprising the following steps:
step 1: acquiring spectral data acquired by a host and a satellite as samples, and determining the wavelength range of the spectral data;
step 2: carrying out full spectrum analysis and normalization operation on the spectral data;
and step 3: dividing each refraction spectrum data into a training data set and a verification data set;
and 4, step 4: establishing a quantitative analysis model by using the host spectral data;
and 5: setting the initial width of a window as WL and the maximum width as WLmax;
step 6: training a multivariate regression model by taking the spectrum of the training set of the satellite machine as input and the corresponding spectral line of the host machine as output;
and 7: recording RMSE of the satellite spectral line and the host spectral line after the verification set is corrected, wherein the window width is = width + 1;
and 8: repeating the step 5 to the step 6 until the width = WLmax;
and step 9: selecting an optimal window width according to the RMSE of the verification set to establish a conversion matrix;
step 10: converting the spectral data of the satellite machine through a conversion matrix;
step 11: and taking the converted satellite machine spectrum data as the input of a host machine model, and predicting by using the host machine model.
2. The model migration method based on the elastic segmentation normalization algorithm according to claim 1,
and acquiring the spectral data of each sample for multiple times, and calculating the average value of the spectral data in the samples.
3. The model migration method based on the elastic segmentation normalization algorithm according to claim 1,
and setting the initial width to WL to realize the conversion of the satellite spectrum data.
4. The model migration method based on the elastic segmentation normalization algorithm according to claim 1,
a relationship between the satellite spectral data and the host spectral data is represented using multiple linear regression.
5. The model migration method based on the elastic segmentation normalization algorithm according to claim 1,
the window width is automatically adjusted according to the position change of the spectral line.
6. The model migration method based on the elastic segmentation normalization algorithm according to claim 1,
the host computer and the satellite machine establish the functional relation of the spectral data in a sectional mode, and correct the spectral data of the satellite machine, and the formula is as follows:
in the formula: ri is the spectral line intensity value of the ith spectral line of the host spectral data, the window width is D, and D is the spectrum under the window width D; f. of i Is the transfer function corresponding to the ith spectral line.
7. A model migration system based on an elastic segmentation standardization algorithm is characterized in that the method adopts the model migration method based on the elastic segmentation standardization algorithm according to any one of claims 1 to 6;
the system comprises: a migration terminal, a host and a satellite;
the host computer is used for establishing a quantitative analysis model by using the spectral data;
the migration terminal is used for acquiring spectral data acquired by the host and the satellite and determining the wavelength range of the spectral data;
carrying out full spectrum analysis and normalization operation on the spectral data; dividing each refraction spectrum data into a training data set and a verification data set;
setting the initial width of a window as WL and the maximum width as WLmax;
training a multivariate regression model by taking the spectrum of the training set of the satellite machine as input and the corresponding spectral line of the host machine as output;
recording RMSE of the spectrum lines of the satellite machine and the host machine after the correction of the verification set, wherein the window width is = width +1, and the width is = WLmax;
selecting an optimal window width according to the RMSE of the verification set to establish a conversion matrix;
converting the spectral data of the satellite machine through a conversion matrix;
and taking the converted satellite machine spectrum data as the input of a host machine model, and predicting by using the host machine model.
8. The model migration system based on the elastic segmentation normalization algorithm according to claim 7,
the host computer and the satellite machine establish the functional relation of the spectral data in a sectional mode, and correct the spectral data of the satellite machine, and the formula is as follows:
in the formula: ri is the spectral line intensity value of the ith spectral line of the host spectral data, the window width is D, and D is the spectrum under the window width D; f. of i Is the transfer function corresponding to the ith spectral line.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210977891.0A CN115049025B (en) | 2022-08-16 | 2022-08-16 | Model migration method and system based on elastic segmentation standardization algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210977891.0A CN115049025B (en) | 2022-08-16 | 2022-08-16 | Model migration method and system based on elastic segmentation standardization algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115049025A true CN115049025A (en) | 2022-09-13 |
| CN115049025B CN115049025B (en) | 2022-11-04 |
Family
ID=83166891
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210977891.0A Active CN115049025B (en) | 2022-08-16 | 2022-08-16 | Model migration method and system based on elastic segmentation standardization algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115049025B (en) |
Citations (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6898533B1 (en) * | 2000-02-01 | 2005-05-24 | The United States Of America As Represented By The Department Of Health And Human Services | Methods for predicting the biological, chemical, and physical properties of molecules from their spectral properties |
| CN101865828A (en) * | 2010-05-31 | 2010-10-20 | 湖南大学 | Method for maintaining predication capability of spectrum correction model of complex system |
| CN103267739A (en) * | 2013-05-07 | 2013-08-28 | 浙江大学 | Aquaculture wastewater organic matter concentration measurement method based on spectrum technology |
| CN106815643A (en) * | 2017-01-18 | 2017-06-09 | 中北大学 | Infrared spectrum Model Transfer method based on random forest transfer learning |
| WO2018010352A1 (en) * | 2016-07-11 | 2018-01-18 | 上海创和亿电子科技发展有限公司 | Qualitative and quantitative combined method for constructing near infrared quantitative model |
| CN109060716A (en) * | 2018-08-31 | 2018-12-21 | 湖南农业大学 | Near-infrared characteristic spectrum Variable Selection based on the competitive adaptive weight weight sampling strategy of window |
| CN109991174A (en) * | 2019-03-26 | 2019-07-09 | 山东省科学院海洋仪器仪表研究所 | A kind of Model transfer method between EO-1 hyperion camera and conventional spectrograph |
| CN111563436A (en) * | 2020-04-28 | 2020-08-21 | 东北大学秦皇岛分校 | Infrared spectrum measuring instrument calibration migration method based on CT-CDD |
| CN111695088A (en) * | 2020-06-12 | 2020-09-22 | 水利部交通运输部国家能源局南京水利科学研究院 | Method and device for screening optimal regression window of satellite precipitation space downscaling |
| CN111999258A (en) * | 2020-07-03 | 2020-11-27 | 桂林理工大学 | Spectral baseline correction-oriented weighting modeling local optimization method |
| CN112414966A (en) * | 2019-08-21 | 2021-02-26 | 东北大学秦皇岛分校 | Near infrared spectrum multi-target calibration migration method based on affine change |
| CN113340874A (en) * | 2020-03-02 | 2021-09-03 | 中国科学院沈阳自动化研究所 | Quantitative analysis method based on combined ridge regression and recursive feature elimination |
| CN113970531A (en) * | 2020-07-24 | 2022-01-25 | 中国石油化工股份有限公司 | Method for correcting spectrum |
| CN114062306A (en) * | 2021-10-25 | 2022-02-18 | 华东理工大学 | Near infrared spectrum data segmentation preprocessing method |
| CN114252430A (en) * | 2020-09-21 | 2022-03-29 | 中国科学院沈阳自动化研究所 | Online ore pulp grade detection method based on laser-induced breakdown spectroscopy technology |
| CN114577671A (en) * | 2022-03-17 | 2022-06-03 | 东北林业大学 | Near-infrared wood density detection method based on parameter correction and transfer learning |
-
2022
- 2022-08-16 CN CN202210977891.0A patent/CN115049025B/en active Active
Patent Citations (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6898533B1 (en) * | 2000-02-01 | 2005-05-24 | The United States Of America As Represented By The Department Of Health And Human Services | Methods for predicting the biological, chemical, and physical properties of molecules from their spectral properties |
| CN101865828A (en) * | 2010-05-31 | 2010-10-20 | 湖南大学 | Method for maintaining predication capability of spectrum correction model of complex system |
| CN103267739A (en) * | 2013-05-07 | 2013-08-28 | 浙江大学 | Aquaculture wastewater organic matter concentration measurement method based on spectrum technology |
| WO2018010352A1 (en) * | 2016-07-11 | 2018-01-18 | 上海创和亿电子科技发展有限公司 | Qualitative and quantitative combined method for constructing near infrared quantitative model |
| CN106815643A (en) * | 2017-01-18 | 2017-06-09 | 中北大学 | Infrared spectrum Model Transfer method based on random forest transfer learning |
| CN109060716A (en) * | 2018-08-31 | 2018-12-21 | 湖南农业大学 | Near-infrared characteristic spectrum Variable Selection based on the competitive adaptive weight weight sampling strategy of window |
| CN109991174A (en) * | 2019-03-26 | 2019-07-09 | 山东省科学院海洋仪器仪表研究所 | A kind of Model transfer method between EO-1 hyperion camera and conventional spectrograph |
| CN112414966A (en) * | 2019-08-21 | 2021-02-26 | 东北大学秦皇岛分校 | Near infrared spectrum multi-target calibration migration method based on affine change |
| CN113340874A (en) * | 2020-03-02 | 2021-09-03 | 中国科学院沈阳自动化研究所 | Quantitative analysis method based on combined ridge regression and recursive feature elimination |
| CN111563436A (en) * | 2020-04-28 | 2020-08-21 | 东北大学秦皇岛分校 | Infrared spectrum measuring instrument calibration migration method based on CT-CDD |
| CN111695088A (en) * | 2020-06-12 | 2020-09-22 | 水利部交通运输部国家能源局南京水利科学研究院 | Method and device for screening optimal regression window of satellite precipitation space downscaling |
| CN111999258A (en) * | 2020-07-03 | 2020-11-27 | 桂林理工大学 | Spectral baseline correction-oriented weighting modeling local optimization method |
| CN113970531A (en) * | 2020-07-24 | 2022-01-25 | 中国石油化工股份有限公司 | Method for correcting spectrum |
| CN114252430A (en) * | 2020-09-21 | 2022-03-29 | 中国科学院沈阳自动化研究所 | Online ore pulp grade detection method based on laser-induced breakdown spectroscopy technology |
| CN114062306A (en) * | 2021-10-25 | 2022-02-18 | 华东理工大学 | Near infrared spectrum data segmentation preprocessing method |
| CN114577671A (en) * | 2022-03-17 | 2022-06-03 | 东北林业大学 | Near-infrared wood density detection method based on parameter correction and transfer learning |
Non-Patent Citations (1)
| Title |
|---|
| 林振兴等: "PDS用于不同温度下的近红外光谱模型传递研究", 《分析测试学报》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115049025B (en) | 2022-11-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104089911A (en) | Spectral model transmission method based on unary linear regression | |
| CN106934416B (en) | Big data-based model matching method | |
| US20030122080A1 (en) | Method of processing data from a spectrophotometer | |
| CN111563436A (en) | Infrared spectrum measuring instrument calibration migration method based on CT-CDD | |
| CN115049025B (en) | Model migration method and system based on elastic segmentation standardization algorithm | |
| CN111157484A (en) | Near infrared spectrum model transfer method for fruit sugar degree detection equipment | |
| CN114739980B (en) | Element information prediction method, device, equipment and medium | |
| CN111337452A (en) | Method for verifying feasibility of spectral data model transfer algorithm | |
| CN113916817B (en) | Spectrum method chromaticity online measurement method for urban living drinking water | |
| CN104181125A (en) | Method for rapidly determining Kol-bach value of beer malt | |
| Kunze et al. | Contactless inline IV measurement of solar cells using an empirical model | |
| CN112949169B (en) | Coal sample test value prediction method based on spectral analysis | |
| CN102057261B (en) | Method and apparatus for automatic calibration of spectrometers in chemometry by means of a bayes iterative estimation method | |
| CN109829513B (en) | Sequential wavelength dispersion X-ray fluorescence spectrum intelligent analysis method | |
| CN114067169A (en) | Raman spectrum analysis method based on convolutional neural network | |
| CN111220565A (en) | CPLS-based infrared spectrum measuring instrument calibration migration method | |
| CN116166969A (en) | Spectral line intensity correlation matching-based spectrum standardization method | |
| CN113155775A (en) | Near infrared spectrum calibration method and fruit quality detection system applying same | |
| CN110286094B (en) | Spectrum model transfer method based on cluster analysis | |
| CN113376107B (en) | Water quality monitoring system and method based on cloud platform | |
| CN117250784B (en) | Liquid crystal display testing method and system | |
| CN117031218A (en) | Aging grade detection method for hyperspectral image of composite insulator | |
| CN117214108B (en) | Batch automatic calibration method based on spectrum detection equipment | |
| CN114036456A (en) | Spectrum model transfer method and system for quantitatively evaluating polymerization degree of insulating paper | |
| Li et al. | Robust Multi-task Learning for Calibration Transfer in DP Detection by NIRS of Insulating Paper |
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 |