CN115184281A - Method and system for determining concentration of solution components based on two-dimensional spectrum - Google Patents
Method and system for determining concentration of solution components based on two-dimensional spectrum Download PDFInfo
- Publication number
- CN115184281A CN115184281A CN202211075417.5A CN202211075417A CN115184281A CN 115184281 A CN115184281 A CN 115184281A CN 202211075417 A CN202211075417 A CN 202211075417A CN 115184281 A CN115184281 A CN 115184281A
- Authority
- CN
- China
- Prior art keywords
- solution
- dimensional
- spectral image
- image
- concentration
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/766—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using regression, e.g. by projecting features on hyperplanes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Medical Informatics (AREA)
- Databases & Information Systems (AREA)
- Mathematical Physics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention provides a method and a system for determining the concentration of a solution component based on a two-dimensional spectrum, which relate to the technical field of solution component determination and comprise the following steps: acquiring spectral data of a solution to be detected; constructing a two-dimensional spectral image of the solution to be detected according to the spectral data of the solution to be detected; extracting the characteristics of the two-dimensional spectral image; inputting the characteristics of the two-dimensional spectral image into a solution component concentration prediction model to obtain the components and the concentration of the solution to be measured; the solution component concentration prediction model is obtained by training an initial prediction model according to the components and the concentration of the historical solution and the characteristics of the two-dimensional spectral image; the initial prediction model is a partial least squares regression model, or a stacked self-coding network model, or a convolutional neural network model. According to the invention, the determination precision of the component concentration of the solution can be improved by performing feature extraction on the two-dimensional spectral image of the solution to be detected and establishing the solution component concentration prediction model.
Description
Technical Field
The invention relates to the technical field of solution component determination, in particular to a method and a system for determining solution component concentration based on a two-dimensional spectrum.
Background
The visible near infrared spectrum technology is widely applied to quantitative analysis of complex solutions such as food, bacteria liquid, blood and the like due to the advantages of no wound, low cost, rapidness, no pollution and the like. The complex solution has both absorption effect and scattering effect, wherein the nonlinear information caused by the scattering effect makes the detection precision of the traditional detection mode not high.
In order to improve the accuracy of detecting the content of a complex solution component, many researchers try to establish a correlation between scattering information in a complex solution and a substance component concentration by using scattering information of a medium, and a multi-optical-path spectrum can obtain the scattering information of a turbid medium by constructing spectral information under different optical paths, thereby effectively improving the analysis accuracy of a strong scattering substance. The spectrum acquisition system and the platform based on the wedge-shaped sample vessel change the optical path length by simply adjusting the angle and the position of an incident point, and realize the simultaneous expression of absorption effect information and scattering effect information. According to the platform, researchers provide a multi-dimensional spectrum fusion method and a multi-dimensional radial distance method, and the multi-dimensional spectrum fusion method ignores inconsistency of information carried by each pixel point in an image during modeling. The multi-dimensional radial distance method extracts the light spot profile parameters, does not utilize light intensity, and is still low in detection precision. In addition, the existing solution component analysis method has more redundant data, so that the overfitting phenomenon is easy to occur in the training model, the calculation complexity of the model is improved, and the prediction precision of the model is reduced.
Disclosure of Invention
The invention aims to provide a method and a system for determining the concentration of a solution component based on a two-dimensional spectrum, which can improve the determination accuracy of the concentration of the solution component.
In order to achieve the purpose, the invention provides the following scheme:
a solution component concentration determination method based on two-dimensional spectroscopy comprises the following steps:
acquiring spectral data of a solution to be detected;
constructing a two-dimensional spectral image of the solution to be detected according to the spectral data of the solution to be detected;
extracting features of the two-dimensional spectral image;
inputting the characteristics of the two-dimensional spectral image into a solution component concentration prediction model to obtain the components and the concentration of the solution to be detected; the solution component concentration prediction model is obtained by training an initial prediction model according to the components and the concentration of a historical solution and the characteristics of a two-dimensional spectral image; the initial prediction model is a partial least squares regression model, or a stacked self-coding network model, or a convolutional neural network model.
Optionally, after the two-dimensional spectral image of the solution to be measured is constructed according to the spectral data of the solution to be measured, the method further includes:
and denoising the two-dimensional spectral image.
Optionally, extracting the feature of the two-dimensional spectral image includes:
extracting spectral data of a preset waveband from the two-dimensional spectral image to obtain a two-dimensional spectral image of the preset waveband;
determining a peak point of the preset waveband two-dimensional spectrum image;
extracting a longitudinal light intensity distribution curve passing through the peak point on the preset waveband two-dimensional spectrum image;
extracting a transverse light intensity distribution curve passing through the peak point on the preset waveband two-dimensional spectrum image;
determining a region larger than a preset threshold value on a longitudinal light intensity distribution curve as a first characteristic region;
determining a region larger than a preset threshold on the transverse light intensity distribution curve as a second characteristic region;
determining the intersection of the first characteristic region and the second characteristic region as a total characteristic region; the characteristics of the two-dimensional spectral image include a longitudinal light intensity distribution curve, a transverse light intensity distribution curve, and a total characteristic region.
Optionally, before the acquiring the spectral data of the solution to be tested, the method further includes:
acquiring spectral data of a plurality of historical solutions;
constructing a two-dimensional spectrum image of the historical solution according to the spectrum data of each historical solution;
extracting the characteristics of the two-dimensional spectral image of each historical solution;
and training an initial prediction model by taking the characteristics of the two-dimensional spectral image of the historical solution as input and the components and the concentration of the historical solution as expected output to obtain a solution component concentration prediction model.
A two-dimensional spectroscopy-based solution component concentration determination system, comprising:
the spectral data acquisition module is used for acquiring spectral data of the solution to be detected;
the two-dimensional spectral image construction module is used for constructing a two-dimensional spectral image of the solution to be detected according to the spectral data of the solution to be detected;
the characteristic extraction module is used for extracting the characteristics of the two-dimensional spectral image;
the prediction module is used for inputting the characteristics of the two-dimensional spectral image into a solution component concentration prediction model to obtain the components and the concentration of the solution to be detected; the solution component concentration prediction model is obtained by training an initial prediction model according to the components and the concentration of a historical solution and the characteristics of a two-dimensional spectral image; the initial prediction model is a partial least square regression model, or a stacked self-coding network model, or a convolutional neural network model.
Optionally, the system includes:
and the denoising module is used for denoising the two-dimensional spectral image.
Optionally, the feature extraction module includes:
the device comprises a preset waveband two-dimensional spectral image determining unit, a spectrum analyzing unit and a spectrum analyzing unit, wherein the preset waveband two-dimensional spectral image determining unit is used for extracting spectral data of a preset waveband from the two-dimensional spectral image to obtain a preset waveband two-dimensional spectral image;
the peak point determining unit is used for determining a peak point of the preset waveband two-dimensional spectrum image;
the longitudinal light intensity distribution curve determining unit is used for extracting a longitudinal light intensity distribution curve passing through the peak point on the preset waveband two-dimensional spectrum image;
the transverse light intensity distribution curve determining unit is used for extracting a transverse light intensity distribution curve passing through the peak point on the preset waveband two-dimensional spectrum image;
the first characteristic region determining unit is used for determining a region larger than a preset threshold on the longitudinal light intensity distribution curve as a first characteristic region;
the first characteristic region determining unit is used for determining a region larger than a preset threshold value on the transverse light intensity distribution curve as a second characteristic region;
a total feature region determining unit, configured to determine an intersection of the first feature region and the second feature region as a total feature region; the characteristics of the two-dimensional spectral image include a longitudinal light intensity distribution curve, a transverse light intensity distribution curve, and a total characteristic region.
Optionally, the system further includes:
the historical solution spectral data acquisition module is used for acquiring the spectral data of a plurality of historical solutions;
the historical solution two-dimensional spectral image construction module is used for constructing a historical solution two-dimensional spectral image according to the spectral data of each historical solution;
the historical solution two-dimensional spectral image feature extraction module is used for extracting the features of each historical solution two-dimensional spectral image;
and the training module is used for training the initial prediction model by taking the characteristics of the two-dimensional spectral image of the historical solution as input and taking the components and the concentration of the historical solution as expected output to obtain a solution component concentration prediction model.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
the invention provides a solution component concentration determination method based on a two-dimensional spectrum, which comprises the following steps: acquiring spectral data of a solution to be detected; constructing a two-dimensional spectral image of the solution to be detected according to the spectral data of the solution to be detected; extracting the characteristics of the two-dimensional spectral image; inputting the characteristics of the two-dimensional spectral image into a solution component concentration prediction model to obtain the components and the concentration of the solution to be detected; the solution component concentration prediction model is obtained by training an initial prediction model according to the components and the concentration of a historical solution and the characteristics of a two-dimensional spectral image; the initial prediction model is a partial least squares regression model, or a stacked self-coding network model, or a convolutional neural network model. The invention can improve the determination precision of the component concentration of the solution by extracting the characteristics of the two-dimensional spectral image of the solution to be detected and establishing the solution component concentration prediction model.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required in the embodiments will be briefly described below, 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 according to these drawings without creative efforts.
FIG. 1 is a flow chart of a method for determining concentration of a component in a solution based on two-dimensional spectroscopy according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a feature region extraction method according to an embodiment of the present disclosure;
FIG. 3 is a diagram illustrating the predicted results of a method for determining the concentration of a component in a solution based on two-dimensional spectroscopy according to an embodiment of the present invention;
FIG. 4 is a diagram illustrating the result of prediction based on the road survey picture according to the embodiment of the 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.
The invention aims to provide a method and a system for determining the concentration of a solution component based on a two-dimensional spectrum, which can improve the determination accuracy of the concentration of the solution component.
In order to make the aforementioned objects, features and advantages of the present invention more comprehensible, the present invention is described in detail with reference to the accompanying drawings and the detailed description thereof.
The invention provides a solution component concentration determination method based on a two-dimensional spectrum, which comprises the following steps: acquiring spectral data of a solution to be detected; constructing a two-dimensional spectral image of the solution to be detected according to the spectral data of the solution to be detected; extracting the characteristics of the two-dimensional spectral image; inputting the characteristics of the two-dimensional spectral image into a solution component concentration prediction model to obtain the components and the concentration of the solution to be detected; the solution component concentration prediction model is obtained by training an initial prediction model according to the components and the concentration of the historical solution and the characteristics of the two-dimensional spectral image; the initial prediction model is a partial least squares regression model, or a stacked self-coding network model, or a convolutional neural network model. After the two-dimensional spectral image of the solution to be measured is constructed, the method further comprises the following steps: and denoising the two-dimensional spectral image. Wherein, extracting the characteristics of the two-dimensional spectral image comprises: extracting spectral data of a preset waveband from the two-dimensional spectral image to obtain a two-dimensional spectral image of the preset waveband; determining a peak point of a two-dimensional spectrum image of a preset waveband; extracting a longitudinal light intensity distribution curve passing through a peak point on a preset waveband two-dimensional spectrum image; extracting a transverse light intensity distribution curve passing through a peak point on a preset waveband two-dimensional spectrum image; determining a region larger than a preset threshold value on a longitudinal light intensity distribution curve as a first characteristic region; determining a region larger than a preset threshold on the transverse light intensity distribution curve as a second characteristic region; determining the intersection of the first characteristic region and the second characteristic region as a total characteristic region; the features of the two-dimensional spectral image include a longitudinal light intensity distribution curve, a transverse light intensity distribution curve, and a total feature region. Before acquiring the spectral data of the solution to be measured, the method further comprises the following steps: acquiring spectral data of a plurality of historical solutions; according to the spectral data of each historical solution, a two-dimensional spectral image of the historical solution is constructed; extracting the characteristics of the two-dimensional spectral image of each historical solution; and training the initial prediction model by taking the characteristics of the two-dimensional spectral image of the historical solution as input and taking the components and the concentration of the historical solution as expected output to obtain a solution component concentration prediction model.
Specifically, as shown in fig. 1, the present invention comprises the following steps:
s1, acquiring hyperspectral data of a sample to be detected;
and S2, restoring the hyperspectral data into a hyperspectral image and preprocessing (removing noise). The preprocessing adopts at least one of two-dimensional convolution smoothing filtering, standard normal variable transformation, savitzky-Golay smoothing (convolution smoothing filtering), multivariate scattering correction, wavelet transformation and Orthogonal Signal Correction (OSC).
S3, extracting a hyperspectral image of a specific wave band in a visible near-infrared light range; the specific wave band extraction method is one of a continuous projection algorithm, a competitive self-adaptive re-weighted sampling method, a genetic algorithm and a random frog-jumping algorithm.
And S4, extracting the transverse (X direction) and longitudinal (Y direction) light intensity distribution curves of the peak points in the hyperspectral image.
And S5, extracting a characteristic region (namely a total characteristic region) in the hyperspectral image. Selecting areas with light intensity values of more than 60% of peak values in the transverse X direction and the longitudinal Y direction of the hyperspectral images of the specific wave bands, and using the intersection of the areas defined in the X direction and the Y direction as a characteristic area;
and S6, establishing a component content prediction model by using all the extracted spectral information. And establishing an analysis and prediction model for the transverse light intensity distribution curve, the longitudinal light intensity distribution curve and the spectral information of the characteristic region extracted from the hyperspectral image and the component concentration of the complex solution, thereby predicting the component concentration in the complex solution to be detected. One of a concentration component analysis model partial least squares regression model, a stacked self-coding network model, and a convolutional neural network model.
The specific implementation mode is as follows:
firstly, preparing a complex solution sample, and acquiring a hyperspectral image of the complex solution sample by using a point-array sliding acquisition hyperspectral platform, wherein the integration time of an optical fiber spectrometer of the hyperspectral image acquisition system is 600ms, the acquisition mode is triggered by level, the acquisition wavelength range is 195.4-991.6nm, the spectral resolution is 0.77nm, and the total is 1044 wavelength spectral data information. The experiment used a mixed solution of indian ink and fat emulsion solution as the sample of the complex solution to be tested. The concentration of indian ink was fixed at 0.023% and the concentration of fat emulsion was graded from 1% to 4.9% at 0.1%, for a total of 40 sets of samples, the hyperspectral platform was used for the collection of hyperspectral images of 40 sets of complex solution samples using a point-array type sliding collection, the size of the collected images was 50pixels by 49pixels, and each set of samples took about 30 minutes. If 1 bad sample is generated due to manual operation errors, the bad samples are kicked, and 39 groups of effective hyperspectral data are calculated. And then splicing the hyperspectral data to restore the hyperspectral data into a two-dimensional spectrum of a spatial domain, and performing preprocessing operation to remove noise.
In order to reduce the interference of irrelevant information, a waveband selection operation is carried out, and 340 wavelengths in the range of 590nm to 847nm are selected to obtain a hyperspectral image of a specific waveband. To further reduce redundant data, information highly related to the component content is extracted, and as shown in fig. 2, a longitudinal direction Y light intensity distribution curve and a transverse direction X positive semi-axis light intensity distribution curve passing through a peak point are extracted for a hyperspectral image of a specific waveband. In fig. 2, pixel represents Pixel, india Ink represents indian Ink, and Intralipid represents fat milk. Gradients are respectively obtained in the transverse X direction and the longitudinal Y direction of the hyperspectral images of the specific wave bands, and areas with the peak values of 60% or more are selected to obtain an intersection as a characteristic area. The extracted spectral information of each point in the hyperspectral image characteristic region, a longitudinal direction Y light intensity distribution curve of the peak point and a transverse direction X positive half-axis light intensity distribution curve are used as input information of modeling to totally 274 characteristic points, one of selectable models is built, and the component content of a complex solution is predicted.
In fig. 3-4, prediction Set represents the Prediction Set, and Linear Fitting Results of Prediction represents the Linear Fitting result of the Prediction Set. As shown in fig. 3 to 4, the modeling using the extracted spectral information can obtain better prediction effects than the modeling using the spectral information of the whole region, and the points marked by arrows in the figures are the points of the feature region. The invention reduces the mean square error by 6.12% under the condition that the data dimension is reduced to 1/9 of the original dimension. By extracting the information highly related to the component content in the hyperspectral image, the complexity of the model is greatly reduced, and the prediction precision of the component content of the complex solution is improved. The computational complexity of extracting the feature region for prediction is compared with that of directly using the spectral picture as shown in table 1.
TABLE 1 comparison table of computational complexity based on feature prediction and spectral picture prediction
In the table, rc represents correction set correlation, and RMSEC represents correction root mean square error; rp represents prediction set correlation; RMSEP denotes the prediction set root mean square error.
In addition, the invention also provides a solution component concentration determination system based on two-dimensional spectrum, which comprises: and the spectral data acquisition module is used for acquiring the spectral data of the solution to be detected. And the two-dimensional spectral image construction module is used for constructing a two-dimensional spectral image of the solution to be detected according to the spectral data of the solution to be detected. And the characteristic extraction module is used for extracting the characteristics of the two-dimensional spectral image. The prediction module is used for inputting the characteristics of the two-dimensional spectral image into the solution component concentration prediction model to obtain the components and the concentration of the solution to be measured; the solution component concentration prediction model is obtained by training an initial prediction model according to the components and the concentration of the historical solution and the characteristics of the two-dimensional spectral image; the initial prediction model is a partial least squares regression model, or a stacked self-coding network model, or a convolutional neural network model. And the denoising module is used for denoising the two-dimensional spectral image.
Wherein, the feature extraction module includes: and the preset waveband two-dimensional spectrum image determining unit is used for extracting the spectrum data of the preset waveband from the two-dimensional spectrum image to obtain the preset waveband two-dimensional spectrum image. And the peak point determining unit is used for determining a peak point of the two-dimensional spectrum image with the preset waveband. And the longitudinal light intensity distribution curve determining unit is used for extracting a longitudinal light intensity distribution curve passing through the peak point on the two-dimensional spectrum image of the preset waveband. And the transverse light intensity distribution curve determining unit is used for extracting a transverse light intensity distribution curve passing through the peak point on the preset waveband two-dimensional spectrum image. And the first characteristic region determining unit is used for determining a region larger than a preset threshold value on the longitudinal light intensity distribution curve as a first characteristic region. And the first characteristic region determining unit is used for determining a region larger than a preset threshold value on the transverse light intensity distribution curve as a second characteristic region. The total characteristic region determining unit is used for determining the intersection of the first characteristic region and the second characteristic region as a total characteristic region; the features of the two-dimensional spectral image include a longitudinal light intensity distribution curve, a transverse light intensity distribution curve, and a total feature region.
The invention generally provides a solution component concentration determination system based on two-dimensional spectrum, which further comprises: the historical solution spectral data acquisition module is used for acquiring the spectral data of a plurality of historical solutions. And the historical solution two-dimensional spectral image construction module is used for constructing a historical solution two-dimensional spectral image according to the spectral data of each historical solution. And the historical solution two-dimensional spectral image feature extraction module is used for extracting the features of each historical solution two-dimensional spectral image. And the training module is used for training the initial prediction model by taking the characteristics of the two-dimensional spectral image of the historical solution as input and taking the components and the concentration of the historical solution as expected output to obtain a solution component concentration prediction model.
In the present specification, the embodiments are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the method part for description.
The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.
Claims (8)
1. A method for determining the concentration of a component in a solution based on two-dimensional spectroscopy is characterized by comprising the following steps:
acquiring spectral data of a solution to be detected;
constructing a two-dimensional spectral image of the solution to be detected according to the spectral data of the solution to be detected;
extracting features of the two-dimensional spectral image;
inputting the characteristics of the two-dimensional spectral image into a solution component concentration prediction model to obtain the components and the concentration of the solution to be detected; the solution component concentration prediction model is obtained by training an initial prediction model according to the components and the concentration of a historical solution and the characteristics of a two-dimensional spectral image; the initial prediction model is a partial least squares regression model, or a stacked self-coding network model, or a convolutional neural network model.
2. The method for determining the concentration of a component in a solution according to claim 1, wherein after the constructing a two-dimensional spectrum image of the solution to be measured according to the spectral data of the solution to be measured, the method further comprises:
and denoising the two-dimensional spectral image.
3. The method for determining the concentration of a component in a solution based on two-dimensional spectrum according to claim 1, wherein the extracting the feature of the two-dimensional spectrum image comprises:
extracting spectral data of a preset waveband from the two-dimensional spectral image to obtain a two-dimensional spectral image of the preset waveband;
determining a peak point of the preset waveband two-dimensional spectrum image;
extracting a longitudinal light intensity distribution curve passing through the peak point on the preset waveband two-dimensional spectrum image;
extracting a transverse light intensity distribution curve passing through the peak point on the preset waveband two-dimensional spectrum image;
determining a region larger than a preset threshold value on a longitudinal light intensity distribution curve as a first characteristic region;
determining a region larger than a preset threshold on the transverse light intensity distribution curve as a second characteristic region;
determining the intersection of the first characteristic region and the second characteristic region as a total characteristic region; the characteristics of the two-dimensional spectral image include a longitudinal light intensity distribution curve, a transverse light intensity distribution curve, and a total characteristic region.
4. The method for determining the concentration of a component in a solution based on two-dimensional spectroscopy according to claim 1, further comprising, before the acquiring the spectroscopic data of the solution to be measured:
acquiring spectral data of a plurality of historical solutions;
according to the spectral data of each historical solution, constructing a two-dimensional spectral image of the historical solution;
extracting the characteristics of the two-dimensional spectral image of each historical solution;
and training an initial prediction model by taking the characteristics of the two-dimensional spectral image of the historical solution as input and the components and the concentration of the historical solution as expected output to obtain a solution component concentration prediction model.
5. A two-dimensional spectroscopy-based solution component concentration determination system, comprising:
the spectral data acquisition module is used for acquiring spectral data of the solution to be detected;
the two-dimensional spectral image construction module is used for constructing a two-dimensional spectral image of the solution to be detected according to the spectral data of the solution to be detected;
the characteristic extraction module is used for extracting the characteristics of the two-dimensional spectral image;
the prediction module is used for inputting the characteristics of the two-dimensional spectral image into a solution component concentration prediction model to obtain the components and the concentration of the solution to be detected; the solution component concentration prediction model is obtained by training an initial prediction model according to the components and the concentration of a historical solution and the characteristics of a two-dimensional spectral image; the initial prediction model is a partial least squares regression model, or a stacked self-coding network model, or a convolutional neural network model.
6. The two-dimensional spectroscopy-based solution component concentration determination system of claim 5, comprising:
and the denoising module is used for denoising the two-dimensional spectral image.
7. The two-dimensional spectrum based solution component concentration determination system of claim 5, wherein the feature extraction module comprises:
the device comprises a preset waveband two-dimensional spectral image determining unit, a spectrum analyzing unit and a spectrum analyzing unit, wherein the preset waveband two-dimensional spectral image determining unit is used for extracting spectral data of a preset waveband from the two-dimensional spectral image to obtain a preset waveband two-dimensional spectral image;
the peak point determining unit is used for determining a peak point of the preset waveband two-dimensional spectrum image;
the longitudinal light intensity distribution curve determining unit is used for extracting a longitudinal light intensity distribution curve passing through the peak point on the preset waveband two-dimensional spectrum image;
a transverse light intensity distribution curve determining unit, configured to extract a transverse light intensity distribution curve passing through the peak point on the preset-waveband two-dimensional spectrum image;
the first characteristic region determining unit is used for determining a region larger than a preset threshold on the longitudinal light intensity distribution curve as a first characteristic region;
the first characteristic region determining unit is used for determining a region larger than a preset threshold value on the transverse light intensity distribution curve as a second characteristic region;
a total feature region determining unit, configured to determine an intersection of the first feature region and the second feature region as a total feature region; the characteristics of the two-dimensional spectral image include a longitudinal light intensity distribution curve, a transverse light intensity distribution curve, and a total characteristic region.
8. The two-dimensional spectroscopy-based solution component concentration determination system of claim 5, further comprising:
the historical solution spectral data acquisition module is used for acquiring the spectral data of a plurality of historical solutions;
the historical solution two-dimensional spectral image construction module is used for constructing a historical solution two-dimensional spectral image according to the spectral data of each historical solution;
the historical solution two-dimensional spectral image feature extraction module is used for extracting the features of each historical solution two-dimensional spectral image;
and the training module is used for training the initial prediction model by taking the characteristics of the two-dimensional spectral image of the historical solution as input and taking the components and the concentration of the historical solution as expected output to obtain a solution component concentration prediction model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211075417.5A CN115184281B (en) | 2022-09-05 | 2022-09-05 | Method and system for determining concentration of solution components based on two-dimensional spectrum |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211075417.5A CN115184281B (en) | 2022-09-05 | 2022-09-05 | Method and system for determining concentration of solution components based on two-dimensional spectrum |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115184281A true CN115184281A (en) | 2022-10-14 |
| CN115184281B CN115184281B (en) | 2022-12-09 |
Family
ID=83524100
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211075417.5A Active CN115184281B (en) | 2022-09-05 | 2022-09-05 | Method and system for determining concentration of solution components based on two-dimensional spectrum |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115184281B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117074360A (en) * | 2023-08-29 | 2023-11-17 | 无锡迅杰光远科技有限公司 | Modeling method, detection method, device and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102609703A (en) * | 2012-03-05 | 2012-07-25 | 中国科学院对地观测与数字地球科学中心 | Method and device for detecting target ground object in hyperspectral image |
| FR2989497A1 (en) * | 2012-04-16 | 2013-10-18 | Green Vision Systems Ltd | Imaging and analyzing sample for identifying object of interest by generating and collecting hyper-spectral image data and information of prepared test solution, processing and analyzing hyper-spectral image data and information |
| CN107505268A (en) * | 2017-08-04 | 2017-12-22 | 中国科学院半导体研究所 | Blood sugar detecting method and system |
| CN109900645A (en) * | 2019-04-17 | 2019-06-18 | 岭南师范学院 | A kind of oyster Measuring Method of Heavy Metal based on hyper-spectral image technique |
| CN111289446A (en) * | 2020-03-30 | 2020-06-16 | 天津工业大学 | Method and system for detecting component concentration of complex solution |
| US20220008157A1 (en) * | 2018-07-31 | 2022-01-13 | Deutsches Krebsforschungszentrum Stiftung des öffentlichen Rechts | Method and system for augmented imaging in open treatment using multispectral information |
-
2022
- 2022-09-05 CN CN202211075417.5A patent/CN115184281B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102609703A (en) * | 2012-03-05 | 2012-07-25 | 中国科学院对地观测与数字地球科学中心 | Method and device for detecting target ground object in hyperspectral image |
| FR2989497A1 (en) * | 2012-04-16 | 2013-10-18 | Green Vision Systems Ltd | Imaging and analyzing sample for identifying object of interest by generating and collecting hyper-spectral image data and information of prepared test solution, processing and analyzing hyper-spectral image data and information |
| CN107505268A (en) * | 2017-08-04 | 2017-12-22 | 中国科学院半导体研究所 | Blood sugar detecting method and system |
| US20220008157A1 (en) * | 2018-07-31 | 2022-01-13 | Deutsches Krebsforschungszentrum Stiftung des öffentlichen Rechts | Method and system for augmented imaging in open treatment using multispectral information |
| CN109900645A (en) * | 2019-04-17 | 2019-06-18 | 岭南师范学院 | A kind of oyster Measuring Method of Heavy Metal based on hyper-spectral image technique |
| CN111289446A (en) * | 2020-03-30 | 2020-06-16 | 天津工业大学 | Method and system for detecting component concentration of complex solution |
Non-Patent Citations (2)
| Title |
|---|
| 张梅等: "基于快速蒙特卡罗的散射介质光学参量干涉测量方法研究", 《光子学报》 * |
| 赵紫竹等: "基于高光谱的牛奶脂肪质量浓度预测模型建立与评价", 《中国乳品工业》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117074360A (en) * | 2023-08-29 | 2023-11-17 | 无锡迅杰光远科技有限公司 | Modeling method, detection method, device and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115184281B (en) | 2022-12-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103411973B (en) | A kind of based on anthocyanin content method for measuring in the vinifera pericarp of EO-1 hyperion | |
| Tsai et al. | Derivative analysis of hyperspectral data | |
| CN111289446B (en) | Method and system for detecting component concentration of complex solution | |
| CN108956505B (en) | Method and device for detecting concentration of chlorophyll a in small water body based on Sentinel-2 image | |
| CN103900972B (en) | Multi-feature fusion-based meat freshness hyperspectral image visual detection | |
| Reddy et al. | Accurate histopathology from low signal-to-noise ratio spectroscopic imaging data | |
| Song et al. | Chlorophyll content estimation based on cascade spectral optimizations of interval and wavelength characteristics | |
| CN105548070B (en) | A kind of apple soluble solid near infrared detection position compensation method and system | |
| CN103472031A (en) | Navel orange sugar degree detection method based on hyper-spectral imaging technology | |
| CN103293118A (en) | Hogwash oil identification method based on near infrared reflectance spectroscopy | |
| Yuan et al. | Non-invasive measurements of ‘Yunhe’pears by vis-NIRS technology coupled with deviation fusion modeling approach | |
| CN115184281B (en) | Method and system for determining concentration of solution components based on two-dimensional spectrum | |
| CN106570862A (en) | Super-resolution reconstruction quality evaluation method and apparatus thereof | |
| ElMasry et al. | Prediction of meat spectral patterns based on optical properties and concentrations of the major constituents | |
| CN105181606A (en) | Hyperspectral imaging technology-based method for detecting sucrose content distribution of peanut | |
| Chen et al. | Identification of various food residuals on denim based on hyperspectral imaging system and combination optimal strategy | |
| CN110779875B (en) | Method for detecting moisture content of winter wheat ear based on hyperspectral technology | |
| CN107247033B (en) | Identify the method for Huanghua Pear maturity based on rapid decay formula life cycle algorithm and PLSDA | |
| Liu et al. | Estimation of chlorophyll content in maize canopy using wavelet denoising and SVR method | |
| Yang et al. | A model for yellow tea polyphenols content estimation based on multi-feature fusion | |
| CN113496218B (en) | Evaluation method and system for hyperspectral remote sensing sensitive wave band selection mode | |
| CN115575345A (en) | Hyperspectral combination data fusion-based tilapia fillet freshness determination method | |
| O'Rourke et al. | The effect of region of interest size on model calibration for soil organic carbon prediction from hyperspectral images of prepared soils | |
| Li et al. | Comparison of two methods for estimation of soil water content from measured reflectance | |
| CN114778457A (en) | Method and device for detecting aflatoxin B1 content in grains |
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 | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20230111 Address after: 1518-027, Unit 2, Floor 15-16, Building 10, Baiqiang Avenue, Fengtai District, Beijing, 100070 Patentee after: Zhilin Technology Co.,Ltd. Patentee after: Beijing Zhilin Technology Co.,Ltd. Address before: Room 2007, Floor 17, Floor 301, Floor 301, Building 4, Building 4, No. 1, Hangfeng Road, Fengtai District, Beijing 100070 Patentee before: Beijing Zhilin Technology Co.,Ltd. |
|
| TR01 | Transfer of patent right |