CN111044460A - Calibration method of artificial intelligence instrument - Google Patents

Calibration method of artificial intelligence instrument Download PDF

Info

Publication number
CN111044460A
CN111044460A CN201911421708.3A CN201911421708A CN111044460A CN 111044460 A CN111044460 A CN 111044460A CN 201911421708 A CN201911421708 A CN 201911421708A CN 111044460 A CN111044460 A CN 111044460A
Authority
CN
China
Prior art keywords
target data
predicted value
model
artificial intelligence
parameters
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
Application number
CN201911421708.3A
Other languages
Chinese (zh)
Inventor
余水保
肖才斌
沈杰
刘瑞杰
徐力
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Changzhou Luopanxing Detection Technology Co ltd
Original Assignee
Changzhou Luopanxing Detection Technology Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Changzhou Luopanxing Detection Technology Co ltd filed Critical Changzhou Luopanxing Detection Technology Co ltd
Priority to CN201911421708.3A priority Critical patent/CN111044460A/en
Publication of CN111044460A publication Critical patent/CN111044460A/en
Pending legal-status Critical Current

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
    • G01N21/27—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands using photo-electric detection ; circuits for computing concentration
    • G01N21/274—Calibration, base line adjustment, drift correction
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component 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
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Spectroscopy & Molecular Physics (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Mathematical Physics (AREA)
  • Computational Linguistics (AREA)
  • Chemical & Material Sciences (AREA)
  • Molecular Biology (AREA)
  • Software Systems (AREA)
  • Biophysics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Pathology (AREA)
  • Immunology (AREA)
  • Biochemistry (AREA)
  • Biomedical Technology (AREA)
  • Computing Systems (AREA)
  • Analytical Chemistry (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a calibration method of an artificial intelligence instrument, which comprises the following steps: under the laboratory environment, obtaining parameters of an AI model used for predicting target data in the instrument by using a full sample; and re-learning according to the field samples to realize field calibration of the predicted value of the target data. The invention solves the problem of interference of the field environment to the instrument, and can realize field calibration with less calculation amount, so that the instrument has anti-interference capability, and the accuracy of instrument prediction is improved.

Description

Calibration method of artificial intelligence instrument
Technical Field
The invention relates to the technical field of spectral analysis, in particular to a calibration method of an artificial intelligence instrument.
Background
Spectroscopic analysis is an important means of analyzing the composition of a substance. When a beam of light having a continuous wavelength passes through a substance, the intensity of some wavelengths in the beam is reduced, and the absorbance of a particular wavelength is proportional to the concentration of the substance according to the lambert beer's law, and almost all substances have their unique absorption spectra, so that the substance can be spectrally analyzed by obtaining the absorption spectra of the substance by a spectrometer.
The spectral analysis has the following problems in the production process:
1. the complicated environment of the industrial field causes serious interference to the spectrum of the sample, and the concentration of the sample in the spectrogram can not be identified by depending on experience;
2. the artificial intelligence instrument is interfered by the spectrum of chemical substances in the complex environment of an industrial field, the measured value can deviate from the actual value greatly, and the calibration must be carried out according to a field sample;
3. the small instrument adopts a singlechip embedded system, has limited memory and CPU resources, and cannot store more training samples and perform complex operation.
Disclosure of Invention
The present invention is directed to solving, at least to some extent, one of the technical problems in the art described above. Therefore, the invention aims to provide a calibration method of an artificial intelligence instrument, which solves the problem of interference of a field environment on the instrument, can finish field calibration with less calculation amount, enables instrument equipment to have anti-interference capability and improves the accuracy of instrument prediction.
In order to achieve the above object, a calibration method for an artificial intelligence instrument according to an embodiment of the present invention includes the following steps: under the laboratory environment, obtaining parameters of an AI model used for predicting target data in the instrument by using a full sample; and re-learning according to the field samples to realize field calibration of the predicted value of the target data.
According to the calibration method of the artificial intelligence instrument, the parameters of the AI model used by the instrument are obtained in a laboratory environment, and the on-site calibration is completed by relearning according to a plurality of on-site samples, so that the problem of interference of the on-site environment to the instrument is solved, the on-site calibration can be completed with less calculation amount, the instrument equipment has anti-interference capability, and the accuracy of instrument prediction is improved.
In addition, the calibration method of the artificial intelligence instrument according to the above embodiment of the present invention may further have the following additional technical features:
the AI model is obtained by carrying out neural network training after PCA dimension reduction.
According to an embodiment of the invention, the AI model is an AI model obtained by PLS.
Wherein the relearning is incremental relearning or small network relearning.
According to one embodiment of the invention, the relearning is polynomial relearning.
Specifically, the in-situ calibration of the predicted value of the target data by the incremental relearning specifically includes: and calculating a loss function on a lower computer by using the parameters of the AI model obtained by carrying out neural network training after dimensionality reduction through PCA (principal component analysis), and updating the parameters of the AI model by iterating for a plurality of times in the fastest gradient descent direction to realize the field calibration of the predicted value of the target data.
Specifically, separating the last layer of the neural network in the AI model as a small network, and implementing the field calibration of the predicted value of the target data by the small network learning specifically includes: and training parameters of the small network by using the plurality of field samples, synchronizing the trained parameters of the small network into the whole neural network, updating the parameters of the AI model obtained by training the neural network after the dimensionality reduction of the PCA, and realizing the field calibration of the predicted value of the target data.
Specifically, the in-situ calibration of the predicted value of the target data by the polynomial relearning specifically includes: and obtaining target data actual values of the plurality of field samples in a manual instrument inspection mode, mapping a predicted value of target data of an AI model obtained by carrying out neural network training or PLS after dimensionality reduction through PCA (principal component analysis), to the target data actual value through a polynomial on a lower computer, calculating a polynomial coefficient, obtaining a predicted value of the target data according to the polynomial, and realizing field calibration of the predicted value of the target data.
Drawings
FIG. 1 is a flow chart of a method of calibrating an artificial intelligence instrument according to an embodiment of the invention;
FIG. 2 is a schematic diagram of a field calibration flow of incremental relearning and small network relearning of a calibration method for an artificial intelligence instrument, in accordance with an embodiment of the present invention;
FIG. 3 is a graph illustrating the effect of an incremental relearning field calibration of a calibration method for an artificial intelligence instrument, in accordance with one embodiment of the present invention;
FIG. 4 is a diagram illustrating the effect of a small network relearning field calibration of the calibration method of an artificial intelligence instrument in accordance with one embodiment of the present invention;
FIG. 5 is a schematic diagram of a field calibration flow for polynomial relearning of a calibration method for an artificial intelligence instrument in accordance with an embodiment of the present invention;
FIG. 6 is a diagram illustrating the effects of a polynomial relearning field calibration of a calibration method for an artificial intelligence instrument, in accordance with an embodiment of the present invention.
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.
As shown in fig. 1, the calibration method of the artificial intelligence instrument according to the embodiment of the present invention includes the following steps:
and S1, acquiring parameters of an AI model used for predicting target data in the instrument by using the full sample under the laboratory environment.
In one embodiment of the present invention, full samples refer to a large number of samples, and the reliability of data obtained from a large number of samples is high. The AI model is a trained neural network model, and when parameters of the AI model are obtained, the parameters of the AI model can be downloaded to a single chip microcomputer of a relevant instrument so as to predict data.
And S2, re-learning is carried out according to the field samples, and field calibration of the predicted value of the target data is realized.
In one embodiment of the present invention, the plurality of samples refers to a small number of sampled samples, and the data obtained from the small number of sampled samples is representative of the data obtained from the large number of samples under certain conditions.
In an embodiment of the present invention, as shown in fig. 2, when the AI model is obtained by neural network training after dimensionality reduction by PCA, the field calibration of the predicted value of the target data can be achieved by incremental relearning or small network relearning.
The field calibration of the predicted value of the target data is realized by incremental relearning, and the field calibration method specifically comprises the following steps: and (3) taking a plurality of field samples as incremental samples, calculating a loss function on a lower computer by utilizing parameters of the AI model obtained by carrying out neural network training after dimensionality reduction through PCA, and updating the parameters of the AI model by iterating for a plurality of times in the fastest gradient descent direction to realize field calibration of the predicted value of the target data.
In addition, the last layer of network of the neural network in the AI model is separated to be used as a small network, and the field calibration of the predicted value of the target data is realized by small network learning, which specifically comprises the following steps: training parameters of the small network by using a plurality of field samples, synchronizing the trained parameters of the small network into the whole neural network, updating parameters of an AI model obtained by training the neural network after dimensionality reduction by PCA, and realizing field calibration of a predicted value of target data.
In one embodiment of the invention, the target data can be the nitrate solubility of a field water sample, a laboratory can configure water samples with different nitrate solubilities, various interferents are added according to industry experience, N samples are randomly taken, the nitrate solubility of the ith sample is yiThe absorbance of the ith sample obtained by the spectrometer was taken as the absorbance at 510 points from 196nm to 630nm and recorded as a vector of 510 dimensions
Figure BDA0002352552930000041
In a specific embodiment of the invention, a proper target dimension is found by observing variance value reaction information energy of PCA, a Sciki-learn module is used for reducing the dimension of a spectrum vector with 510 dimensions to 13 dimensions, the output result of the PCA is trained by a neural network to obtain parameters of an AI model, and the AI model is used for predicting the solubility of nitrate to obtain a predicted value with the fitting degree of 0.99. The equipment is transported to the site, the water sample of the site is directly predicted according to the parameters of the AI model and the AI model under the site environment, the fitness of the obtained predicted value is 0.601, and the prediction result is known to have larger deviation due to site interference.
The method comprises the steps of taking a plurality of water samples on site, for example 3-5 water samples on site as incremental samples, calculating a loss function on a lower computer according to a 'fastest gradient descent method' for parameters of an AI model, iterating for a plurality of times along the loss function in the direction of fastest gradient descent to update part of inter-node coefficients and intercept parameters in the parameters of the AI model obtained by carrying out neural network training after dimensionality reduction through PCA, predicting the water samples on site by using the AI model after updating the parameters to obtain a predicted value fitting degree of 0.988, wherein the iterative computation amount of the incremental samples is lower than that of the full samples by a plurality of orders of magnitude, and the lower computer can easily complete computation. Referring to fig. 3, the predicted value fitting degree of the nitrate solubility predicted by the AI model in the laboratory is as high as 0.99, while the predicted value fitting degree obtained by the prediction with the AI model without updated parameters on site is 0.601, which has almost no use value, and the predicted value fitting degree after the incremental relearning is 0.988, which achieves good calibration effect.
In addition, the last layer of the neural network in the AI model is separated to be used as a small network, and a plurality of water samples in the field, for example 3-5 water samples in the field, are used as samples to train parameters of the small network, specifically, the solubility of the water samples in the field is used
Figure BDA0002352552930000051
And recording the input quantity of the last layer in the neural network
Figure BDA0002352552930000052
The two vectors train parameters of the small network, the weight parameters and the intercept parameters are adjusted, finally the trained parameters of the small network are synchronized to the whole neural network, so that parameters of an AI model obtained by neural network training after PCA dimension reduction are updated, the AI model with the updated parameters is used for predicting a field water sample, the fitting degree of a predicted value is 0.964, the calculation amount of parameter training of the small network is small, and a lower computer can easily complete calculation. Referring to fig. 4, the predicted value fitting degree of the nitrate solubility predicted by the AI model in the laboratory is 0.99, while the predicted value fitting degree obtained by the AI model without updated parameters in the field is 0.601, which has almost no use value, and the predicted value fitting degree after relearning through the small network is 0.964, thereby achieving a good calibration effect.
In one embodiment of the present invention, as shown in fig. 5, when the AI model is obtained by neural network training after dimensionality reduction by PCA or is obtained by PLS, the in-situ calibration of the predicted values of the target data can be achieved by polynomial relearning.
The field calibration of the predicted value of the target data is realized by relearning a polynomial, and specifically comprises the following steps: obtaining target data actual values of a plurality of field samples in a manual instrument inspection mode, mapping a predicted value of target data of an AI model obtained by carrying out neural network training or PLS after dimensionality reduction through PCA to the target data actual value through a polynomial on a lower computer, calculating a polynomial coefficient, obtaining the predicted value of the target data according to the polynomial, and realizing field calibration of the predicted value of the target data.
In one embodiment of the invention, in a laboratory environment, 510 dimensions are input from a spectrometer
Figure BDA0002352552930000061
Vector-derived approximation function of AI model
Figure BDA0002352552930000062
To derive parameters of an AI model or by PLSAnd (3) obtaining parameters of the AI model, predicting the nitrate solubility to obtain a predicted value fitting degree of 0.99, directly predicting a water sample on site by using the parameters of the AI model and the AI model under the site environment when the equipment moves on the site to obtain the predicted value fitting degree of 0.601, and showing that the prediction result has larger deviation due to site interference.
In one embodiment of the present invention, the actual solubility of 3-5 field water samples is obtained by manual instrumental examination, for example, by wet chemistry. Because the interferent with specific interference action exists on site, according to the correlation of the influence of the interferent on the measurement result of the nitrate solubility, the correlation is characterized by a polynomial relation in the application range, and finally the measured value is compensated, namely the approximation of a polynomial function g can be written as: y islay2＝gp(ylay1) According to the measured actual solubility, polynomial fitting is carried out by means of a polynomial fitting function, specifically, polynomial fitting can be carried out by means of a polyfit function of a Numpy library, a polynomial coefficient vector p is determined, further, the predicted value fitting degree of the nitrate solubility is 0.942, the second layer polynomial coefficient is usually smaller than 3, fitting calculation amount on a lower computer is small, referring to fig. 6, the predicted value fitting degree of the nitrate solubility predicted by an AI model in a laboratory is 0.99, the predicted value fitting degree of the AI model obtained by neural network training after dimensionality reduction through PCA is 0.601 without calibration, the use value is almost zero, and the predicted value fitting degree after relearning through the polynomial is 0.942, so that a good calibration effect is achieved.
Most of the calculated amount is concentrated on PLS fitting regression or neural network training, calculation is completed on the upper computer, and the lower computer only needs to directly download the trained parameters, so that the calculated amount is small.
According to the calibration method of the artificial intelligence instrument, the parameters of the AI model used by the instrument are obtained in the laboratory environment, and the on-site calibration is completed by relearning according to the field samples, so that the problem of interference of the on-site environment to the instrument is solved, the on-site calibration can be completed with less calculation amount, the instrument equipment has high anti-interference capability, and the accuracy of instrument prediction is improved.
In the present invention, unless otherwise expressly stated or limited, the terms "mounted," "connected," "secured," and the like are to be construed broadly and can, for example, be fixedly connected, detachably connected, or integrally formed; can be mechanically or electrically connected; either directly or indirectly through intervening media, either internally or in any other relationship. The specific meanings of the above terms in the present invention can be understood by those skilled in the art according to specific situations.
In the present invention, unless otherwise expressly stated or limited, the first feature "on" or "under" the second feature may be directly contacting the first and second features or indirectly contacting the first and second features through an intermediate. Also, a first feature "on," "over," and "above" a second feature may be directly or diagonally above the second feature, or may simply indicate that the first feature is at a higher level than the second feature. A first feature being "under," "below," and "beneath" a second feature may be directly under or obliquely under the first feature, or may simply mean that the first feature is at a lesser elevation than the second feature.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.

Claims (8)

1. A calibration method of an artificial intelligence instrument is characterized by comprising the following steps:
under the laboratory environment, obtaining parameters of an AI model used for predicting target data in the instrument by using a full sample;
and re-learning according to the field samples to realize field calibration of the predicted value of the target data.
2. The method of calibrating an artificial intelligence instrument of claim 1 wherein the AI model is an AI model trained on a neural network after dimensionality reduction by PCA.
3. The method of calibrating an artificial intelligence instrument of claim 1, wherein the AI model is an AI model obtained by PLS.
4. The method of calibrating an artificial intelligence instrument of claim 2 wherein the relearning is incremental relearning or small network relearning.
5. The method of calibrating an artificial intelligence instrument of claim 2 or 3 wherein the relearning is polynomial relearning.
6. The method for calibrating an artificial intelligence instrument of claim 4, wherein the in-situ calibration of the predicted value of the target data is achieved by the incremental relearning, specifically comprising:
and calculating a loss function on a lower computer by using the parameters of the AI model obtained by carrying out neural network training after dimensionality reduction through PCA (principal component analysis), and updating the parameters of the AI model by iterating for a plurality of times in the fastest gradient descent direction to realize the field calibration of the predicted value of the target data.
7. The method for calibrating an artificial intelligence instrument according to claim 4, wherein a last layer of a neural network in the AI model is separated as a small network, and the small network learning is used to achieve field calibration of the predicted value of the target data, specifically comprising:
and training parameters of the small network by using the plurality of field samples, synchronizing the trained parameters of the small network into the whole neural network, updating the parameters of the AI model obtained by training the neural network after the dimensionality reduction of the PCA, and realizing the field calibration of the predicted value of the target data.
8. The method for calibrating an artificial intelligence instrument according to claim 5, wherein the relearning with the polynomial to achieve the field calibration of the predicted value of the target data specifically comprises:
and obtaining target data actual values of the plurality of field samples in a manual instrument inspection mode, mapping a predicted value of target data of an AI model obtained by carrying out neural network training or PLS after dimensionality reduction through PCA (principal component analysis), to the target data actual value through a polynomial on a lower computer, calculating a polynomial coefficient, obtaining a predicted value of the target data according to the polynomial, and realizing field calibration of the predicted value of the target data.
CN201911421708.3A 2019-12-31 2019-12-31 Calibration method of artificial intelligence instrument Pending CN111044460A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911421708.3A CN111044460A (en) 2019-12-31 2019-12-31 Calibration method of artificial intelligence instrument

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911421708.3A CN111044460A (en) 2019-12-31 2019-12-31 Calibration method of artificial intelligence instrument

Publications (1)

Publication Number Publication Date
CN111044460A true CN111044460A (en) 2020-04-21

Family

ID=70243363

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911421708.3A Pending CN111044460A (en) 2019-12-31 2019-12-31 Calibration method of artificial intelligence instrument

Country Status (1)

Country Link
CN (1) CN111044460A (en)

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5402240A (en) * 1994-02-17 1995-03-28 Dupree, Inc. Sperm densimeter
US6281499B1 (en) * 1998-05-12 2001-08-28 Nireco Corporation Method for analyzing measured value by on-line spectral analyzer
US20040033617A1 (en) * 2002-08-13 2004-02-19 Sonbul Yaser R. Topological near infrared analysis modeling of petroleum refinery products
US20040033618A1 (en) * 1998-10-13 2004-02-19 Haass Michael J. Accommodating subject and instrument variations in spectroscopic determinations
WO2007106942A1 (en) * 2006-03-21 2007-09-27 Commonwealth Scientific And Industrial Research Organisation Analysis of grape quality using neural network
CN103226095A (en) * 2013-04-08 2013-07-31 南京国电环保科技有限公司 Fast calibration method of spectrometer wavelength
US20140309959A1 (en) * 2011-12-16 2014-10-16 Halliburton Energy Services, Inc. Methods of calibration transfer for a testing instrument
CN105825229A (en) * 2016-03-15 2016-08-03 浙江工业大学 Classification method for spectral features of artificial satellite based on deep neural network learning algorithm
CN110084379A (en) * 2019-05-08 2019-08-02 东莞德福得精密五金制品有限公司 The method that calibration instrument is calibrated is treated using artificial intelligence cloud computing
CN110530816A (en) * 2019-05-05 2019-12-03 中国科学院南京土壤研究所 The method that rice blast is early diagnosed using infrared photoacoustic spectra

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5402240A (en) * 1994-02-17 1995-03-28 Dupree, Inc. Sperm densimeter
US6281499B1 (en) * 1998-05-12 2001-08-28 Nireco Corporation Method for analyzing measured value by on-line spectral analyzer
US20040033618A1 (en) * 1998-10-13 2004-02-19 Haass Michael J. Accommodating subject and instrument variations in spectroscopic determinations
US20040033617A1 (en) * 2002-08-13 2004-02-19 Sonbul Yaser R. Topological near infrared analysis modeling of petroleum refinery products
WO2007106942A1 (en) * 2006-03-21 2007-09-27 Commonwealth Scientific And Industrial Research Organisation Analysis of grape quality using neural network
US20140309959A1 (en) * 2011-12-16 2014-10-16 Halliburton Energy Services, Inc. Methods of calibration transfer for a testing instrument
CN103226095A (en) * 2013-04-08 2013-07-31 南京国电环保科技有限公司 Fast calibration method of spectrometer wavelength
CN105825229A (en) * 2016-03-15 2016-08-03 浙江工业大学 Classification method for spectral features of artificial satellite based on deep neural network learning algorithm
CN110530816A (en) * 2019-05-05 2019-12-03 中国科学院南京土壤研究所 The method that rice blast is early diagnosed using infrared photoacoustic spectra
CN110084379A (en) * 2019-05-08 2019-08-02 东莞德福得精密五金制品有限公司 The method that calibration instrument is calibrated is treated using artificial intelligence cloud computing

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
俞斌峰: "多任务学习及其在光谱数据分析中的应用", 《中国优秀博硕士学位论文全文数据库(博士) 基础科学辑》 *
李志良 等: "人工神经网络应用于光谱分析：校正与辨识", 《应用科学学报》 *

Similar Documents

Publication Publication Date Title
Du et al. Spectral regions selection to improve prediction ability of PLS models by changeable size moving window partial least squares and searching combination moving window partial least squares
US6341257B1 (en) Hybrid least squares multivariate spectral analysis methods
US5459677A (en) Calibration transfer for analytical instruments
KR101727889B1 (en) Monitoring, detecting and quantifying chemical compounds in a gas sample stream
Ferré et al. Net analyte signal calculation for multivariate calibration
US6415233B1 (en) Classical least squares multivariate spectral analysis
CN101825567A (en) Screening method for near infrared spectrum wavelength and Raman spectrum wavelength
UA86820C2 (en) Method for development of independent multi-dimensional calibration models
Panchuk et al. Calibration transfer between different analytical methods
CN106650926B (en) A kind of steady boosting extreme learning machine integrated modelling approach
CN109060771B (en) Consensus model construction method based on different characteristic sets of spectrum
Yu et al. Prediction of soil properties based on characteristic wavelengths with optimal spectral resolution by using Vis-NIR spectroscopy
Wang et al. A deep learning-based method for screening soil total nitrogen characteristic wavelengths
CN111999258A (en) Spectral baseline correction-oriented weighting modeling local optimization method
CN113496100B (en) System and computer-implemented method for extrapolating a calibration spectrum
CN112630180B (en) Ultraviolet/visible light absorption spectrum model for detecting concentration of organophosphorus pesticide in water body
Liu et al. Detection of Apple Taste Information Using Model Based on Hyperspectral Imaging and Electronic Tongue Data.
Wehlburg et al. New hybrid algorithm for maintaining multivariate quantitative calibrations of a near-infrared spectrometer
Xu et al. Optimizing machine learning models for predicting soil pH and total P in intact soil profiles with visible and near-infrared reflectance (VNIR) spectroscopy
KR20050080818A (en) Methods for deciding weighted regression model and predicting concentration of component of mixture using the same
CN110887798B (en) Nonlinear full-spectrum water turbidity quantitative analysis method based on extreme random tree
Pessoa et al. Development of ant colony optimization (aco) algorithms based on statistical analysis and hypothesis testing for variable selection
CN117191739A (en) Near infrared spectrum quantitative detection method and system for correcting moisture interference
CN111044460A (en) Calibration method of artificial intelligence instrument
Kompany-Zareh et al. Multi-way based calibration transfer between two Raman spectrometers

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
RJ01 Rejection of invention patent application after publication

Application publication date: 20200421

RJ01 Rejection of invention patent application after publication