CN113916817B - Spectrum method chromaticity online measurement method for urban living drinking water - Google Patents
Spectrum method chromaticity online measurement method for urban living drinking water Download PDFInfo
- Publication number
- CN113916817B CN113916817B CN202111227619.2A CN202111227619A CN113916817B CN 113916817 B CN113916817 B CN 113916817B CN 202111227619 A CN202111227619 A CN 202111227619A CN 113916817 B CN113916817 B CN 113916817B
- Authority
- CN
- China
- Prior art keywords
- dividing
- optimal
- spectrum
- elements
- chromaticity
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 40
- 238000001228 spectrum Methods 0.000 title claims abstract description 30
- 239000003651 drinking water Substances 0.000 title claims abstract description 19
- 235000020188 drinking water Nutrition 0.000 title claims abstract description 19
- 238000000691 measurement method Methods 0.000 title claims abstract description 13
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims abstract description 53
- 230000011218 segmentation Effects 0.000 claims abstract description 32
- 238000012549 training Methods 0.000 claims abstract description 24
- 238000000862 absorption spectrum Methods 0.000 claims abstract description 21
- 238000004445 quantitative analysis Methods 0.000 claims abstract description 19
- 238000012360 testing method Methods 0.000 claims abstract description 12
- 238000000411 transmission spectrum Methods 0.000 claims abstract description 10
- 239000008367 deionised water Substances 0.000 claims abstract description 7
- 229910021641 deionized water Inorganic materials 0.000 claims abstract description 7
- 239000011159 matrix material Substances 0.000 claims abstract description 7
- 239000013598 vector Substances 0.000 claims description 39
- 238000012795 verification Methods 0.000 claims description 12
- 238000004433 infrared transmission spectrum Methods 0.000 claims description 8
- 238000005192 partition Methods 0.000 claims description 6
- 230000003595 spectral effect Effects 0.000 claims description 6
- 238000002835 absorbance Methods 0.000 claims description 4
- 238000010200 validation analysis Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000004737 colorimetric analysis Methods 0.000 claims description 3
- 238000000926 separation method Methods 0.000 claims description 3
- 238000010183 spectrum analysis Methods 0.000 claims description 3
- 238000005259 measurement Methods 0.000 abstract description 11
- 238000004611 spectroscopical analysis Methods 0.000 abstract description 4
- 238000004458 analytical method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- BASFCYQUMIYNBI-UHFFFAOYSA-N platinum Chemical compound [Pt] BASFCYQUMIYNBI-UHFFFAOYSA-N 0.000 description 3
- ZLMJMSJWJFRBEC-UHFFFAOYSA-N Potassium Chemical compound [K] ZLMJMSJWJFRBEC-UHFFFAOYSA-N 0.000 description 2
- 239000003153 chemical reaction reagent Substances 0.000 description 2
- GVPFVAHMJGGAJG-UHFFFAOYSA-L cobalt dichloride Chemical compound [Cl-].[Cl-].[Co+2] GVPFVAHMJGGAJG-UHFFFAOYSA-L 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000012625 in-situ measurement Methods 0.000 description 2
- 229910052700 potassium Inorganic materials 0.000 description 2
- 239000011591 potassium Substances 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- CLBRCZAHAHECKY-UHFFFAOYSA-N [Co].[Pt] Chemical compound [Co].[Pt] CLBRCZAHAHECKY-UHFFFAOYSA-N 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 229910001410 inorganic ion Inorganic materials 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 238000001819 mass spectrum Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000008239 natural water Substances 0.000 description 1
- 230000009965 odorless effect Effects 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 229910052697 platinum Inorganic materials 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 239000000243 solution Substances 0.000 description 1
- 239000012086 standard solution Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
-
- 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
- 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
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
Abstract
The invention provides a spectroscopic method chromaticity online measurement method for urban living drinking water, which solves the problems that the existing water chromaticity measurement is easy to cause secondary pollution, has lower prediction precision and cannot capture nonlinear characteristics. The method comprises the following steps: 1) Collecting spectrum data; 2) Standard water references were obtained:I 1 to be measured, the transmission spectrum of the water body is I 0 Is the transmission spectrum of standard deionized water; 3) Converting a standard water reference into an absorbance spectrum; 4) Dividing all absorbance spectra into a training set and a test set, and recording absorbance spectrum A of the training set mn Is a matrix with row m being n; 5) According to the optimal segmentation points of all columns of the training set, determining an optimal segmentation variable j and an optimal segmentation point s; 6) Dividing the optimal dividing variable into n stages according to the optimal dividing variable j and the optimal dividing point s until the number q of all the subareas reaches a set value, and completing the establishment of the nonlinear full-spectrum chromaticity quantitative analysis model; 7) And (5) testing.
Description
Technical Field
The invention relates to a chromaticity measurement technology of town life drinking water, in particular to a spectroscopic chromaticity online measurement method for town life drinking water.
Background
The urban living drinking water refers to drinking water and domestic water for urban resident living, and is particularly critical for online measurement of urban living drinking water. The water quality chromaticity is an index for measuring the color of natural water or various water after treatment, and is one of organoleptic indexes of water. Water is a colorless, odorless transparent liquid, and when certain substances (such as some soluble organic substances, some inorganic ions, colored suspended particles, etc.) exist in the water, the water may become colored, that is, the water may have a certain color, that is, chromaticity.
The conventional measurement of water quality chromaticity is carried out by a platinum cobalt standard colorimetric method, namely, a standard solution for chromaticity measurement is prepared by using potassium chloroplatinate (K2 PtCl 6) and cobalt chloride (CoCl2.6H2O), and when 1L of water contains 2.419mg of potassium chloroplatinate and 2.00mg of cobalt chloride, the color shade generated when the concentration of platinum (Pt) is 1mg per liter is defined as 1 degree (1 degree). The method needs to use chemical reagents, is complex in operation, is easy to cause secondary pollution, and cannot complete automatic, rapid and in-situ measurement of water quality and chromaticity.
The traditional spectrometry water quality detection technology mainly comprises a single-wavelength and multi-wavelength combination method. The absorbance of the water body at 550nm is measured by a single wavelength method, the chromaticity of the water body is measured by a linear regression method, and the defects of low model prediction accuracy caused by very small available data volume exist; the common multi-wavelength combination method is a PLS method, the PLS method simplifies data by continuously extracting main components, and a regression model is built.
Disclosure of Invention
The invention provides a spectroscopic method chromaticity online measurement method for urban living drinking water, which aims to solve the technical problems that the existing water quality chromaticity measurement needs chemical reagents, is complex in operation and is easy to cause secondary pollution, the automatic, quick and in-situ measurement of the water quality chromaticity cannot be realized, and the spectroscopic method water quality detection technology has low prediction precision or cannot capture nonlinear characteristics due to the fact that the available data quantity is very small.
In order to achieve the above purpose, the technical scheme provided by the invention is as follows:
the online measurement method for the colorimetry of the urban living drinking water is characterized by comprising the following steps of:
1) Acquisition of spectral data
Measuring ultraviolet-visible-near infrared transmission spectrum curves of a plurality of measured water bodies and ultraviolet-visible-near infrared transmission spectrum curves of standard deionized water by adopting an ultraviolet-visible-near infrared full spectrum analysis module;
2) Standard water reference
Standard water references were obtained by the following formula:
wherein I is 1 To be measured, the transmission spectrum of the water body is I 0 Is the transmission spectrum of standard deionized water;
3) Absorbance conversion
Standard water references were converted to absorbance spectra by the following formula:
A＝-log(I)
4) Separation of training set and validation set
Dividing all the obtained absorbance spectrums into a training set and a testing set, recording the absorbance spectrums A of the training set, wherein the number of samples of the training set is more than 70% of the total samples of all the absorbance spectrums mn Is a matrix with the behavior m being n, wherein m is the number of samples of the training set, n is the number of pixels of each sample, A mn Each column is the pixel value of all samples at the same spectral position, and the m chromaticity values corresponding to m samples of the training set are recorded as vectors c, c= [ c ] 1 ,c 2 ，…，c m ]；
5) Calculating an optimal segmentation variable j and an optimal segmentation point s
5.1 Determining the first column of optimal partition points
5.1.1 Set a dividing variable j 1 For the first column vector x 1 ,x 1 ＝[A 11 ,A 21 ,…,A m1 ]Dividing point s 1 1 is shown in the specification;
5.1.2 According to the division point s 1 Dividing the first column vector into two sub-regions R 1 And R is 2 ；
Wherein: r is R 1 The number of elements p=s 1 =1, region R 1 The element set is expressed as: r is R 1 (j 1 ,s 1 )＝A 11 ；
R 2 The number of elements q=m-p, region R 2 The element set is expressed as: r is R 2 (j 1 ，s 1 )＝[A 21 ，，A 31 ，，…，A m1 ]；
5.1.3 Calculating a predicted output valueAnd->
Wherein: x is x 1i ∈R 1 (j 1 ，s 1 )＝A 11 ；
x 2i ∈R 2 (j 1 ，s 1 )＝[A 21 ，，A 31 ，，…，A m1 ]；
f (E) is a function for realizing setting, and takes the median of the elements in the vector as the output value of f (E);
5.1.4 Calculation of (c)
5.1.5 Dividing the point s 1 Sequentially replaced by s 2 =2 to s m =m, using the methods of steps 5.1.2) to 5.1.4), M(s) is obtained respectively 2 )…M(s m ) Is a value of (2);
5.1.6 Get M(s) 1 )、M(s 2 )…M(s m ) Minimum value M(s) w )，w∈[1，m]，M(s w ) Corresponding s w For the segmentation variable j 1 Is denoted as M (j) 1 ，s w1 )；
5.2 Determining the optimal partition points of the rest columns
Traversing the segmentation variable j using the method of step 5.1) 2 For the second column vector x 2 To the segmentation variable j n Is the nth column vector x n Is denoted as M (j) 2 ，s w2 )…M(j n ，s wn )；
Wherein x is 2 ＝[A 12 ，A 22 ，…，A m2 ]；
x n ＝[A 1n ，A 2n ，…,A mn ]；
5.3 Determining an optimal segmentation variable j and an optimal segmentation point s
Calculate M (j) 1 ,s w1 )、M(j 2 ,s w2 )…M(j n ,s wn ) Minimum value M (j) a ,s wb ),a∈[1,n],b∈[1,m]Record j a ,s wb The optimal segmentation variable j and the optimal segmentation point s are respectively;
6) Establishing a nonlinear full spectrum chromaticity quantitative analysis model
6.1 Dividing the optimal dividing variable into two sub-areas R1 and R2 according to the optimal dividing variable j and the optimal dividing point s;
wherein: let the column vector determined by the optimal segmentation variable j be x, x i Is an element in x;
the number of elements in R1 p=s, and the set of elements of region R1 is denoted R 1 (j,s)，R 1 (j, s) is all x in x i A fraction of +.s;
the number of elements q=m-p in R2, the set of elements of region R2 being denoted R 2 (j,s)，R 2 (j, s) is all x in x i A moiety of > s;
6.2 Using the method of step 5.1), the sub-regions R are determined separately 1 (j, s) and R 2 An optimal dividing point of (j, s), and a sub-region R according to the optimal dividing point 1 (j, s) dividing into new secondary sub-regions, and dividing sub-region R 2 (j, s) dividing into new secondary subregions;
6.3 Using the method of step 5.1) to divide each secondary subarea respectively;
6.4 Repeating the step 6.3), re-dividing each divided sub-region until the number q of all sub-regions after n-level division reaches a set value, and completing the establishment of a nonlinear full-spectrum colorimetric quantitative analysis model, wherein q=2 n ；
7) Testing
Inputting a test set absorbance spectrum sample B into the nonlinear full spectrum colorimetric quantitative analysis model in the step 6.4), and dividing the sample B into subareask∈{1，…，q}，/>The corresponding chromaticity value of the element in the matrix is c k The chroma value z of sample B is:
z＝f(c k )。
further, in step 6.4), the n-level division is divided into five-level divisions, and the number of the sub-areas is 32.
Further, in step 5.1.3), f (∈) is a function for implementing the setting, and the output value of the element in the vector with the median of f (∈) is specifically:
f (E) is a function for realizing setting, wherein the sorting of all elements in the vector (E) is recorded as a vector (theta), if the number of the elements in the vector (E) is more than 6, three maximum values and three minimum values of the elements in the vector (theta) are removed and then recorded as a vector (beta), the median pi of the elements in the orientation quantity (beta), and pi is an output value of f (E); if the number of the elements in f (E) is less than or equal to 6, the median of the elements in the vector E is the output value of f (E).
Further, in step 4), all absorbance spectrum total samples further include a validation set;
step A) verification is further included between the step 6) and the step 7); and (3) taking the verification set as a sample to be input into a nonlinear full-spectrum colorimetric quantitative analysis model, and verifying the accuracy of the model on the verification set.
Compared with the prior art, the invention has the advantages that:
1. according to the invention, mass spectrum data are obtained by measuring the ultraviolet-visible-near infrared transmission spectrum of the water body, and a nonlinear full-spectrum water body chromaticity quantitative analysis model is established by analyzing the spectrum data, so that water body chromaticity measurement can be completed, and compared with the traditional method, the method has the advantages of larger available information quantity and higher measurement precision.
2. The method disclosed by the invention utilizes full spectrum data to carry out modeling, utilizes the model to carry out quantitative analysis on water quality chromaticity, can more fully utilize spectrum information compared with the traditional modeling method of single spectrum and multi-spectrum, excavates nonlinear characteristics of the spectrum information in a high-dimensional space, and establishes a more accurate quantitative analysis model of chromaticity, so that the measurement accuracy is higher, and the method is more suitable for domestic drinking water with relatively good water quality.
3. According to the invention, the transmission spectrum of the water body is measured, the full-spectrum water body chromaticity quantitative analysis model is established to predict the water quality chromaticity, and the water quality chromaticity is predicted through spectrum data, so that the method has the advantages of high accuracy, high measurement speed, no secondary pollution and the like.
4. Partial least square method correlation coefficient R of traditional linear analysis modeling method 2 = 0.6293, mean square error mse= 2.3469; the method establishes the model correlation coefficient R 2 = 0.9418, the mean square error mse= 0.4474, which can significantly improve the prediction accuracy of chromaticity, so that the measured mean square error of chromaticity is reduced by one order of magnitude.
Drawings
FIG. 1 is a flow chart of the spectroscopic colorimetric on-line measurement method for town domestic drinking water of the present invention;
FIG. 2 is a graph of transmission spectra in the spectroscopic colorimetric on-line measurement method of drinking water for town life of the present invention;
FIG. 3 is a graph of the result of verification set chromaticity prediction in the online measurement method of the spectrocolorimetry chromaticity of the urban living drinking water.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.
As shown in fig. 1, the invention provides a spectroscopic colorimetric on-line measurement method for urban living drinking water, which comprises the following steps:
1) Acquisition of spectral data
Measuring ultraviolet-visible-near infrared transmission spectrum curves of a plurality of measured water bodies and ultraviolet-visible-near infrared transmission spectrum curves of standard deionized water by adopting an ultraviolet-visible-near infrared full spectrum analysis module, wherein the ultraviolet-visible-near infrared transmission spectrum curves are shown in figure 2;
2) Standard water reference
Standard water references were obtained by the following formula:
wherein I is 1 To be measured, the transmission spectrum of the water body is I 0 Is the transmission spectrum of standard deionized water;
3) Absorbance conversion
The standard water reference was converted to absorbance spectrum a by:
A＝-log(I)
4) Separation of training set and validation set
Dividing all the obtained absorbance spectrums into a training set, a verification set and a test set, wherein the number of samples of the training set is more than 70% of the total samples of all the absorbance spectrums, the training set of the embodiment accounts for 80% of the total samples, and the absorbance spectrum A of the training set is recorded mn Is a matrix with the behavior m being n, wherein m is the number of samples of the training set, n is the number of pixels of each sample, A mn Each column is the pixel value of all samples at the same spectral position, and the m chromaticity values corresponding to m samples of the training set are recorded as vectors c, c= [ c ] 1 ，c 2 ，…，c m ]The method comprises the steps of carrying out a first treatment on the surface of the The invention aims to find a method capable of calculating the chromaticity value by n pixels;
5) Training by utilizing absorbance spectrum in training set, and calculating optimal segmentation variable j and optimal segmentation point s
5.1 Determining the first column of optimal partition points
5.1.1 Set a dividing variable j 1 For the first column vector x 1 ,x 1 ＝[A 11 ，A 21 ，…，A m1 ]Dividing point s 1 1 is shown in the specification;
5.1.2 According to the division point s 1 Dividing the first column vector into two sub-regions R 1 And R is 2 ；
Wherein: r is R 1 The number of elements p=s 1 =1, region R 1 The element set is expressed as: r is R 1 (j 1 ，s 1 )＝A 11 ；
R 2 The number of elements q=m-p, region R 2 The element set is expressed as: r is R 2 (j 1 ，s 1 )＝[A 21 ,,A 31 ,,…,A m1 ]；
5.1.3 Calculating a predicted output valueAnd->
Wherein: x is x 1i ∈R 1 (j 1 ,s 1 )＝A 11 ；
x 2i ∈R 2 (j 1 ,s 1 )＝[A 21 ,,A 31 ,,…,A m1 ]；
Wherein f (E) is a function for realizing setting, and has the meaning of sequencing all elements in the vector (E) to be denoted as a vector (theta), removing three maximum values and three minimum values of the elements in the vector (theta) to be denoted as a vector (beta), and orienting the median of the elements in the vector (beta)Pi, i.e., the output value of f (E); when the number of elements in f (epsilon) is less than or equal to 6, f (epsilon) is the median of the elements in the direct vector. E shaped example 1 ＝[1,3,4,7,4,2,8,9,5,3,11]Then f (∈) 1 )＝5；∈ 2 ＝[1,3,4,7,4]Then f (∈) 2 )＝4；
5.1.4 Calculation of (c)
Wherein: x is x 1i ∈R 1 (j 1 ,s 1 )＝A 11 ；
x 2i ∈R 2 (j 1 ,s 1 )＝[A 21 ,,A 31 ,,…,A m1 ]；
5.1.5 Dividing the point s 1 Sequentially replaced by s 2 =2 to s m =m, using the methods of steps 5.1.2) to 5.1.4), respectively, yieldsM(s 2 )…M(s m ) Is a value of (2);
5.1.6 Get M(s) 1 )、M(s 2 )…M(s m ) Minimum value M(s) w ),w∈[1,m]，M(s w ) Corresponding s w For the segmentation variable j 1 Is denoted as M (j) 1 ,s w1 )；
5.2 Determining the optimal partition points of the rest columns
Traversing the segmentation variable j using the method of step 5.1) 2 For the second column vector x 2 To the segmentation variable j n Is the nth column vector x n Is denoted as M (j) 2 ,s w2 )…M(j n ,s wn )；
Wherein x is 2 ＝[A 12 ,A 22 ,…,A m2 ]；
x n ＝[A 1n ,A 2n ,…,A mn ]；
5.3 Determining an optimal segmentation variable J and an optimal segmentation point s
Calculate M (j) 1 ,s w1 )、M(j 2 ,s w2 )…M(j n ,s wn ) Minimum value M (j) a ,s wb ),a∈[1,n],b∈[1,m]Record j a ,s wb The optimal segmentation variable j and the optimal segmentation point s are respectively;
6) Establishing a nonlinear full spectrum chromaticity quantitative analysis model
6.1 Dividing the optimal dividing variable into two sub-areas R1 and R2 according to the optimal dividing variable j and the optimal dividing point s;
wherein: let the column vector determined by the optimal segmentation variable j be x, x i Is an element in x;
the number of elements in R1 p=s, and the set of elements of region R1 is denoted R 1 (j,s)，R 1 (j, s) is all x in x i A fraction of +.s;
the number of elements q=m-p in R2, the set of elements of region R2 being denoted R 2 (j,s)，R 2 (j, s) is all x in x i A moiety of > s;
6.2 Using the method of step 5.1), the sub-regions R are determined separately 1 (j, s) and R 2 An optimal dividing point of (j, s), and a sub-region R according to the optimal dividing point 1 (j, s) dividing into new secondary sub-regions And sub-region R 2 (j, s) division into new secondary subregions +.>Obtaining four secondary subregions
6.3 Using the method of step 5.1) to divide each secondary subarea respectively;
6.4 Repeating step 6.3) forDividing each sub-region after dividing again until the number q of all sub-regions after n-level dividing reaches a set value, wherein q=2 n The method comprises the steps of carrying out a first treatment on the surface of the The present embodiment divides the sample into five sub-regions32, wherein the value of 32 is preset, the effect is optimal, and the establishment of the nonlinear full-spectrum chromaticity quantitative analysis model is completed;
7) Verification
Inputting the verification set as a sample into the nonlinear full-spectrum colorimetric quantitative analysis model in the step 6.4), and verifying the accuracy of the model on the verification set; the result is shown in FIG. 3, the correlation coefficient R 2 = 0.9418, mean square error mse= 0.4474;
8) Testing
Inputting a test set of absorbance spectrum sample B into a nonlinear full spectrum colorimetric quantitative analysis model verified by a verification set, wherein the sample B can be directly divided into subareas by only one vector and n pixelsk∈{1,…,32}，/>The corresponding chromaticity value of the element in the matrix is c k The chroma value z of sample B is:
z＝f(c k )。
according to the invention, full spectrum data is utilized for modeling, and the model is utilized for quantitative analysis of water quality and chromaticity, so that compared with the traditional modeling method of single spectrum and multi-spectrum, the method can more fully utilize spectrum information, excavate nonlinear characteristics of the spectrum information in a high-dimensional space, and establish a more accurate quantitative analysis model of chromaticity. Partial least square method correlation coefficient R of traditional linear analysis modeling method 2 = 0.6293, mean square error mse= 2.3469; and the correlation coefficient R of the method of the invention 2 Because of the fact that the mean square error MSE= 0.4474 is 0.9418, the analysis model based on the method can remarkably improve the prediction accuracy of chromaticity, so that the color is formedThe mean square error of the measurement of the degree is reduced by an order of magnitude, which is only 19% of that of the conventional PLS method.
The above description is only of the preferred embodiments of the present invention, and the technical solution of the present invention is not limited thereto, and any modifications made by those skilled in the art based on the main technical concept of the present invention are included in the technical scope of the present invention.
Claims (4)
1. The online spectroscopic chromaticity measurement method for urban living drinking water is characterized by comprising the following steps of:
1) Acquisition of spectral data
Measuring ultraviolet-visible-near infrared transmission spectrum curves of a plurality of measured water bodies and ultraviolet-visible-near infrared transmission spectrum curves of standard deionized water by adopting an ultraviolet-visible-near infrared full spectrum analysis module;
2) Standard water reference
Standard water references were obtained by the following formula:
wherein I is 1 To be measured, the transmission spectrum of the water body is I 0 Is the transmission spectrum of standard deionized water;
3) Absorbance conversion
Standard water references were converted to absorbance spectra by the following formula:
A＝-log(I)
4) Separation of training set and test set
Dividing all the obtained absorbance spectrums into a training set and a testing set, recording the absorbance spectrums A of the training set, wherein the number of samples of the training set is more than 70% of the total samples of all the absorbance spectrums mn Is a matrix with the behavior m being n, wherein m is the number of samples of the training set, n is the number of pixels of each sample, A mn Each column is the pixel value of all samples at the same spectral position, and m chromaticity values corresponding to m samples of the training set are recorded as vectors c, c＝[c 1 ，c 2 ，…，c m ]；
5) Calculating an optimal segmentation variable j and an optimal segmentation point s
5.1 Determining the first column of optimal partition points
5.1.1 Set a dividing variable j 1 For the first column vector x 1 ，x 1 ＝[A 11 ，A 21 ，...，A m1 ]Dividing point s 1 1 is shown in the specification;
5.1.2 According to the division point s 1 Dividing the first column vector into two sub-regions R 1 And R is 2 ；
Wherein: r is R 1 The number of elements p=s 1 =1, region R 1 The element set is expressed as: r is R 1 (j 1 ，s 1 )＝A 11 ；
R 2 The number of elements q=m-p, region R 2 The element set is expressed as: r is R 2 (j 1 ，s 1 )＝[A 21 ，A 31 ，...，A m1 ]；
5.1.3 Calculating a predicted output valueAnd->
Wherein: x is x 1i ∈R 1 (j 1 ，s 1 )＝A 11 ；
x 2i ∈R 2 (j 1 ，s 1 )＝[A 21 ，A 31 ，...，A m1 ]；
f (v) is a preset function, and takes the median of the elements in the vector as the output value of f (v);
5.1.4 Calculation of (c)
5.1.5 Dividing the point s 1 Sequentially replaced by s 2 =2 to s m =m, using the methods of steps 5.1.2) to 5.1.4), M(s) is obtained respectively 2 )...M(s m ) Is a value of (2);
5.1.6 Get M(s) 1 )、M(s 2 )...M(s m ) Minimum value M(s) w )，w∈[1，m]，M(s w ) Corresponding s w For the segmentation variable j 1 Is denoted as M (j) 1 ，s w1 )；
5.2 Determining the optimal partition points of the rest columns
Traversing the segmentation variable j using the method of step 5.1) 2 For the second column vector x 2 To the segmentation variable j n Is the nth column vector x n Is denoted as M (j) 2 ，s w2 )…M(j n ，s wn )；
Wherein x is 2 ＝[A 12 ，A 22 ，...，A m2 ]；
x n ＝[A 1n ，A 2n ，...，A mn ]；
5.3 Determining an optimal segmentation variable j and an optimal segmentation point s
Calculate M (j) 1 ，s w1 )、M(j 2 ，s w2 )…M(j n ，s wn ) Minimum value M (j) a ，s wb )，a∈[1，n]，b∈[1，m]Record j a ，s wb The optimal segmentation variable j and the optimal segmentation point s are respectively;
6) Establishing a nonlinear full spectrum chromaticity quantitative analysis model
6.1 Dividing the optimal dividing variable into two sub-areas R1 and R2 according to the optimal dividing variable j and the optimal dividing point s;
wherein: set the column direction determined by the optimal dividing variable jThe amount is x, x i Is an element in x;
the number of elements in R1 p=s, and the set of elements of region R1 is denoted R 1 (j，s)，R 1 (j, s) is all x in x i A fraction of +.s;
the number of elements q=m-p in R2, the set of elements of region R2 being denoted R 2 (j，s)，R 2 (j, s) is all x in x i A moiety of > s;
6.2 Using the method of step 5.1), the sub-regions R are determined separately 1 (j, s) and R 2 An optimal dividing point of (j, s), and a sub-region R according to the optimal dividing point 1 (j, s) dividing into new secondary sub-regions, and dividing sub-region R 2 (j, s) dividing into new secondary subregions;
6.3 Using the method of step 5.1) to divide each secondary subarea respectively;
6.4 Repeating the step 6.3), re-dividing each divided sub-region until the number t of all sub-regions after n-level division reaches a set value, and completing the establishment of a nonlinear full-spectrum colorimetric quantitative analysis model, wherein t=2 n ；
7) Testing
Inputting a test set absorbance spectrum sample B into the nonlinear full spectrum colorimetric quantitative analysis model in the step 6.4), and dividing the sample B into subareas The corresponding chromaticity value of the element in the matrix is c k The chroma value z of sample B is:
z＝f(c k )。
2. the spectroscopic colorimetric in-line measurement method for town domestic drinking water as claimed in claim 1, wherein: in step 6.4), the n-level division is a five-level division, and the number of the sub-areas is 32.
3. The online spectroscopic colorimetric method for urban living drinking water according to claim 1, wherein in step 5.1.3), f (v) is a preset function, and the output value of the median of the element in the vector f (v) is specifically:
f (v) is a preset function, wherein the ordering of all elements in the vector v is recorded as a vector theta, if the number of the elements in the f (v) is more than 6, three maximum values and three minimum values of the elements in the vector theta are removed and then recorded as a vector beta, and the median pi of the elements in the orientation quantity beta is the output value of f (v); if the number of the elements in f (v) is less than or equal to 6, the median of the elements in the vector v is the output value of f (v).
4. The spectroscopic colorimetric in-line measurement method for town domestic drinking water as claimed in claim 1, wherein: in step 4), all absorbance spectrum total samples further comprise a validation set;
step A) verification is further included between the step 6) and the step 7); and (3) taking the verification set as a sample to be input into a nonlinear full-spectrum colorimetric quantitative analysis model, and verifying the accuracy of the model on the verification set.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111227619.2A CN113916817B (en) | 2021-10-21 | 2021-10-21 | Spectrum method chromaticity online measurement method for urban living drinking water |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111227619.2A CN113916817B (en) | 2021-10-21 | 2021-10-21 | Spectrum method chromaticity online measurement method for urban living drinking water |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113916817A CN113916817A (en) | 2022-01-11 |
| CN113916817B true CN113916817B (en) | 2024-02-02 |
Family
ID=79242157
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111227619.2A Active CN113916817B (en) | 2021-10-21 | 2021-10-21 | Spectrum method chromaticity online measurement method for urban living drinking water |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113916817B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116735444B (en) * | 2023-06-08 | 2024-04-02 | 深圳市中科云驰环境科技有限公司 | Method and device for detecting concentration of suspended matters in water body based on infrared scattering spectrum |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6246471B1 (en) * | 1998-06-08 | 2001-06-12 | Lj Laboratories, Llc | Apparatus and method for measuring optical characteristics of an object |
| JP2014135649A (en) * | 2013-01-10 | 2014-07-24 | Nippon Hoso Kyokai <Nhk> | Spectroscopy parameter arithmetic apparatus, spectroscopy parameter arithmetic method, and program for optimal metamer |
| CN107067444A (en) * | 2017-04-27 | 2017-08-18 | 上海出版印刷高等专科学校 | A kind of spectral gamut mapping method of optimization |
| CN108663339A (en) * | 2018-05-15 | 2018-10-16 | 南京财经大学 | Corn online test method of going mouldy based on spectrum and image information fusion |
| CN110945561A (en) * | 2017-06-13 | 2020-03-31 | 爱色丽公司 | Hyperspectral imaging spectrophotometer and system |
| CN111096815A (en) * | 2018-10-25 | 2020-05-05 | 伊沃克拉尔维瓦登特股份公司 | Method for determining tooth color |
| KR102148992B1 (en) * | 2020-07-01 | 2020-08-27 | 김선기 | device that tests drinking water using color and turbidity |
| CN112730293A (en) * | 2020-12-28 | 2021-04-30 | 深圳市中科云驰环境科技有限公司 | Ammonia nitrogen water quality monitor and method based on spectrum analysis method |
| CN113418874A (en) * | 2021-03-15 | 2021-09-21 | 安徽工业技术创新研究院六安院 | Hyperspectral image reconstruction method and detection method for fruit quality detection |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7459696B2 (en) * | 2003-04-18 | 2008-12-02 | Schomacker Kevin T | Methods and apparatus for calibrating spectral data |
| US11432706B2 (en) * | 2019-06-20 | 2022-09-06 | Cilag Gmbh International | Hyperspectral imaging with minimal area monolithic image sensor |
-
2021
- 2021-10-21 CN CN202111227619.2A patent/CN113916817B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6246471B1 (en) * | 1998-06-08 | 2001-06-12 | Lj Laboratories, Llc | Apparatus and method for measuring optical characteristics of an object |
| JP2014135649A (en) * | 2013-01-10 | 2014-07-24 | Nippon Hoso Kyokai <Nhk> | Spectroscopy parameter arithmetic apparatus, spectroscopy parameter arithmetic method, and program for optimal metamer |
| CN107067444A (en) * | 2017-04-27 | 2017-08-18 | 上海出版印刷高等专科学校 | A kind of spectral gamut mapping method of optimization |
| CN110945561A (en) * | 2017-06-13 | 2020-03-31 | 爱色丽公司 | Hyperspectral imaging spectrophotometer and system |
| CN108663339A (en) * | 2018-05-15 | 2018-10-16 | 南京财经大学 | Corn online test method of going mouldy based on spectrum and image information fusion |
| CN111096815A (en) * | 2018-10-25 | 2020-05-05 | 伊沃克拉尔维瓦登特股份公司 | Method for determining tooth color |
| KR102148992B1 (en) * | 2020-07-01 | 2020-08-27 | 김선기 | device that tests drinking water using color and turbidity |
| CN112730293A (en) * | 2020-12-28 | 2021-04-30 | 深圳市中科云驰环境科技有限公司 | Ammonia nitrogen water quality monitor and method based on spectrum analysis method |
| CN113418874A (en) * | 2021-03-15 | 2021-09-21 | 安徽工业技术创新研究院六安院 | Hyperspectral image reconstruction method and detection method for fruit quality detection |
Non-Patent Citations (3)
| Title |
|---|
| An improved scheme for multifeature-based foreground detection using challenging conditions;Mohanty, SK 等;Digital Signal Processing;第113卷;第1-22页 * |
| 基于光谱与图像信息的杏鲍菇多糖含量检测;宋镇 等;华南农业大学学报;第40卷(第03期);第104-110页 * |
| 基于多特征融合和软投票的遥感图像河流检测;张庆春 等;光学学报;第38卷(第06期);第320-326页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113916817A (en) | 2022-01-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104089911B (en) | Spectral model transmission method based on one-variable linear regression | |
| CN110726694A (en) | Characteristic wavelength selection method and system of spectral variable gradient integrated genetic algorithm | |
| CN108152235B (en) | Heavy metal content inversion method combining soil indoor and outdoor spectra | |
| CN111965140B (en) | Wavelength point recombination method based on characteristic peak | |
| CN113916817B (en) | Spectrum method chromaticity online measurement method for urban living drinking water | |
| CN105334186A (en) | Infrared spectral analysis method | |
| CN107966499B (en) | Method for predicting crude oil carbon number distribution by near infrared spectrum | |
| CN114414521A (en) | Milk main component measuring method based on infrared multispectral sensor | |
| CN110987858A (en) | Method for rapidly detecting oil product by using neural network data model | |
| CN110567894A (en) | method and device for rapidly detecting pH value of aqueous solution | |
| CN111896497B (en) | Spectral data correction method based on predicted value | |
| CN110231328B (en) | Raman spectrum quantitative analysis method based on half-peak height distance method | |
| Huang et al. | Multigranularity cascade forest algorithm based on hyperspectral imaging to detect moisture content in Daqu | |
| CN102507462B (en) | Hyperspectral-analysis-based copper quality detection method and system | |
| CN111220565B (en) | CPLS-based infrared spectrum measuring instrument calibration migration method | |
| CN104181125A (en) | Method for rapidly determining Kol-bach value of beer malt | |
| CN115376005A (en) | Insulator pollution component identification method and system based on spectrum resolving network | |
| CN113588597A (en) | Method for improving analysis precision of furnace slag | |
| CN1544917A (en) | Spectrum quantitative automatic analysis method | |
| CN111426680B (en) | Method for rapidly measuring coal caking index and colloidal layer index | |
| CN117250183B (en) | Gas component analysis method, apparatus, device, and storage medium | |
| CN114282446A (en) | Fitting prediction method based on different preference spectrum models | |
| CN114839158A (en) | Fitting prediction method for spectral data of multi-sensor portable near-infrared spectrometer | |
| CN115060685A (en) | Multi-index collaborative prediction method based on spectral data | |
| CN111458308A (en) | Near infrared spectrum gentian identification method and system |
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 |