CN107421895A - A kind of water quality parameter retrieving concentration method and apparatus of multiband optimum organization - Google Patents
A kind of water quality parameter retrieving concentration method and apparatus of multiband optimum organization Download PDFInfo
- Publication number
- CN107421895A CN107421895A CN201710527421.3A CN201710527421A CN107421895A CN 107421895 A CN107421895 A CN 107421895A CN 201710527421 A CN201710527421 A CN 201710527421A CN 107421895 A CN107421895 A CN 107421895A
- Authority
- CN
- China
- Prior art keywords
- wave band
- group
- water quality
- quality parameter
- inverse model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
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
-
- 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
- G01N2021/1793—Remote sensing
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Theoretical Computer Science (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention provides a kind of water quality parameter retrieving concentration method and apparatus of multiband optimum organization, this method includes:Obtain the measured concentration of sample point water quality parameter and the hyper spectral reflectance data of water body;The wave band of hyper spectral reflectance data is screened using successive projection algorithm, obtains first group of wave band;According to measured concentration and the hyper spectral reflectance data of corresponding first group of wave band, the first offset minimum binary inverse model is built;Calculate the hyper spectral reflectance data for corresponding to first group of wave band contribution margin in the first offset minimum binary inverse model;First group of wave band of hyper spectral reflectance data is screened according to contribution margin, obtains corresponding second group of wave band;Using partial least squares algorithm and according to the measured concentration of water quality parameter and the hyper spectral reflectance data of corresponding second group of wave band, the second offset minimum binary inverse model is built;The hyper spectral reflectance data input of corresponding second group of wave band to the second offset minimum binary inverse model is subjected to retrieving concentration, obtains the retrieving concentration value of water quality parameter.
Description
Technical field
The present invention relates to Remote Sensing Techniques in Determining Water Quality technical field, more particularly to a kind of water quality parameter of multiband optimum organization
Retrieving concentration method and apparatus.
Background technology
While lake storehouse type water head site provides great lot of water resources for the mankind, it may have important ecosystem function, be complete
Fullerenes circulate and the important component of Nutrient Cycle.However as the aggravation of the effect of human activity, lake storehouse type water head site constantly by
The threat of organic and inorganic matter pollution, water environmental problems getting worse.The management and protection of water environment need a wide range of, continuous
Water quality monitoring is supported, and conventional water quality monitoring uses lab analysis means more, this mode is not only time-consuming, it is laborious but also
Water quality condition on monitoring section can only be obtained, it is difficult to meet to carry out the needs of a wide range of, Multi-phases dynamic monitors to water quality.
Partial least squares algorithm can comprehensively utilize multiple band class informations and carry out inverting water quality as a kind of multiple regression procedure
The concentration of parameter, it is a kind of important water quality remote sensing inversion method.At present, relevant water quality parameter offset minimum binary inverse model is ground
Study carefully, full spectrum builds partial least square model in how directly visible light or near infrared range, or with certain wave band interval come
Extraction spectrum then mainly uses cross-verification method to build partial least square model, and in the timing really of optimal principal component.
Wherein, the wave band that directly using full spectrum wave band structure partial least square model partial least square model can be caused to model
Number is much larger than sampled point number, and existing conllinear sex chromosome mosaicism can aggravate the uncertain of partial least square model modeling between wave band
Property;And spectrum is extracted with certain wave band interval, it but can not effectively select the wave band sensitive to water quality parameter inverting;And interact
Method of inspection can be verified, and missed with validation-cross it is determined that during optimal principal component by rejecting part modeling sample for model
Difference determines optimal principal component number, and this can lose part modeling sampling point information, so as to cause partial least square model to cross plan
Conjunction or poor fitting.
As can be seen here, generally deposited in retrieving concentration of the use partial least square model to water quality parameter in the prior art
Synteny influences between spectrum, so as to cause partial least square model used it is uncertain strong the problem of.
The content of the invention
It is existing to solve the invention provides a kind of water quality parameter retrieving concentration method and apparatus of multiband optimum organization
In technology in retrieving concentration of the use partial least square model to water quality parameter, synteny influences between existing spectrum,
So as to cause partial least square model used it is uncertain strong the problem of.
In order to solve the above problems, according to an aspect of the present invention, the invention discloses a kind of multiband optimum organization
Water quality parameter retrieving concentration method, including:
Obtain the measured concentration of sample point water quality parameter and the bloom of the corresponding different-waveband of the sample point water body
Compose reflectivity data;
First time screening is carried out to the wave band of the hyper spectral reflectance data using successive projection algorithm, obtains first group
Wave band, wherein, first group of wave band includes multiple wave bands;
According to the measured concentration of the water quality parameter and the hyper spectral reflectance data of corresponding first group of wave band, structure
First offset minimum binary inverse model;
The hyper spectral reflectance data of corresponding first group of wave band are calculated in the first offset minimum binary inverse model
In contribution margin;
Programmed screening is carried out to first group of wave band of the hyper spectral reflectance data according to the contribution margin, obtained
To corresponding second group of wave band, wherein, the quantity of the wave band included by second group of wave band is wrapped less than first group of wave band
The quantity of the wave band included;
Using partial least squares algorithm and according to the measured concentration and correspondingly second group of wave band of the water quality parameter
Hyper spectral reflectance data, build the second offset minimum binary inverse model;
The sample point is corresponded into the hyper spectral reflectance data input of second group of wave band to described second partially most
A young waiter in a wineshop or an inn multiplies inverse model and carries out retrieving concentration, obtains the retrieving concentration value of the water quality parameter.
According to another aspect of the present invention, the invention also discloses a kind of water quality parameter concentration of multiband optimum organization is anti-
Device is drilled, including:
Acquisition module, for obtain sample point water quality parameter measured concentration and the sample point water body correspondence not
With the hyper spectral reflectance data of wave band;
First screening module, for carrying out first to the wave band of the hyper spectral reflectance data using successive projection algorithm
Secondary screening, first group of wave band is obtained, wherein, first group of wave band includes multiple wave bands;
First structure module, for the measured concentration according to the water quality parameter and the bloom of corresponding first group of wave band
Reflectivity data is composed, builds the first offset minimum binary inverse model;
Computing module, the hyper spectral reflectance data for calculating corresponding first group of wave band are partially minimum described first
Two multiply the contribution margin in inverse model;
Second screening module, for first group of wave band according to the contribution margin to the hyper spectral reflectance data
Programmed screening is carried out, obtains corresponding second group of wave band, wherein, the quantity of the wave band included by second group of wave band is less than institute
State the quantity of the wave band included by first group of wave band;
Second structure module, for using partial least squares algorithm and according to the measured concentration of the water quality parameter and correspondingly
The hyper spectral reflectance data of second group of wave band, build the second offset minimum binary inverse model;
Inverting module, for the sample point to be corresponded to the hyper spectral reflectance data input of second group of wave band extremely
The second offset minimum binary inverse model carries out retrieving concentration, obtains the retrieving concentration value of the water quality parameter.
Compared with prior art, the present invention includes advantages below:
The embodiment of the present invention is anti-by using the full wave EO-1 hyperion of sample point of the successive projection algorithm to getting
The first time screening that rate carries out wave band is penetrated, and the hyper spectral reflectance based on the corresponding wave band after first time screening is created at it
Offset minimum binary inverse model in contribution margin carry out the wave band after being screened to first time and carry out postsearch screening, it is final so as to obtain
Corresponding wave band for model construction;And hyper spectral reflectance based on the corresponding wave band after postsearch screening and sample point
The measured concentration of water quality parameter carries out the structure of offset minimum binary inverse model, realizes the retrieving concentration to water quality parameter, keeps away
Exempt from the synteny influence between spectrum in modeling, significantly reduce modeling wave band number, it is dense so as to reduce progress water quality parameter
Spend the uncertainty of the partial least square model modeling of inverting.
Brief description of the drawings
Fig. 1 is a kind of step flow of the water quality parameter retrieving concentration embodiment of the method for multiband optimum organization of the present invention
Figure;
Fig. 2 is a kind of structural frames of the water quality parameter retrieving concentration device embodiment of multiband optimum organization of the present invention
Figure.
Embodiment
In order to facilitate the understanding of the purposes, features and advantages of the present invention, it is below in conjunction with the accompanying drawings and specific real
Applying mode, the present invention is further detailed explanation.
Reference picture 1, show a kind of water quality parameter retrieving concentration embodiment of the method for multiband optimum organization of the present invention
Step flow chart, specifically may include steps of:
Step 101, the measured concentration and the different ripples of correspondence of the sample point water body of sample point water quality parameter are obtained
The hyper spectral reflectance data of section;
Wherein it is possible to lay sampled point in research area, sampling and spectrum data gathering on the spot are carried out, obtains sample point water
The measured concentration of matter parameter and the hyper spectral reflectance data of different-waveband;
Lay sampled point:Uniform stationing principle can be used to lay sampled point in research area;
The water quality parameter that uniform stationing carries out water quality sampling acquisition can at utmost characterize whole research area's water quality parameter
Distribution, uniform stationing scheme can be laid along parallel longitude or latitude, can also be according to the location distribution edge in lake
A certain fixed-direction is layouted with specific interval.
In an instantiation, can using some lake as research area, and according to lay sampled point principle according to this
The location distribution in lake, uniformly lay multiple sampled points (such as 40 sampled points).
Sampling of water quality and analysis:Water sample is gathered at or near the sampled point of laying, sealing refrigeration, sends lab analysis back to
So as to the measured concentration in obtaining step 101, wherein, water quality parameter can include chlorophyll-a concentration, total suspended matter concentration and turbid
Degree etc. in any one, wherein, turbidity can be obtained with in-site measurement;
In addition, in order to ensure sampling quality, during sampling, sampling boat can be made to be located at sampled point downstream direction, to avoid ship
Body agitation basic sediment causes water sample to pollute.Sampler will be washed first 2-3 times with water sample container with sampling pond during sampling, and paste
Good label, gathers 3 parts of water samples to each sampled point simultaneously, the measured concentration of the concentration averages of 3 parts of water samples as the sampled point, and
Each sample point coordinate is recorded using GPS.Record the sampled point date and time, sampling point number and remarks (sampled point periphery
Statement, photographic intelligence etc.), and taken pictures with camera, photo numbering remarks are on the label of corresponding sampling bottle.All samples
It should be placed in incubator and refrigerate.The water quality parameter mainly measured has chlorophyll-a concentration, total suspended matter concentration, wherein turbidity, turbidity
Using nephelometer in-site measurement, chlorophyll-a concentration be able to can be used in use for laboratory metric measurement, total suspended matter concentration
Weight method measures.
After the water sample of each sample point is gathered, step 101 can of the embodiment of the present invention obtains each sampled point
Water quality parameter measured concentration.
Obtain actual measurement high-spectral data:While sample point gathers water sample, the embodiment of the present invention can also utilize light
Spectrometer gathers the hyper spectral reflectance data of each sampled point actual measurement;
In an instantiation, while water sample is gathered, can be used using SVC HR1024 spectrometers the water surface with
Upper mensuration measures the up spoke brightness L of the water surface successivelysw, sky optical signal LskyAnd standard hawk spoke brightness Lp, then according to such as
Lower formula calculates water body Remote Sensing Reflectance Rrs, i.e. hyper spectral reflectance.
Wherein, r is gas-water interface to the reflectivity of skylight, by solar zenith angle, azimuth and observed bearing etc. because
Element determines, when actually measuring, when zenith angle is 40 °, and r ≈ 0.0245, ρpFor 30% standard hawk reflectivity.
Certainly, in this example it is measurement that reflectivity is carried out using standard hawk as reflecting plate, according to hawk in practical application
Material is different, ρpReflectivity can also change therewith, typically based on 20% or 30%.
Or the satellite hyper spectral reflectance data in covering research area, for example, HJ1A HSI can also be utilized
(Hyperspectral Imager) is that China passes in the EO-1 hyperion carried on the environment disaster reduction satellite 1A stars that in September, 2008 is launched
Sensor, the spectral resolution of spatial resolution, the temporal resolution of 4 days and 10nm or so with 100m, can meet big model
The Remote Sensing Techniques in Determining Water Quality enclosed.HJ1A can be from Chinese Resources satellite application center website (http://www.cresda.com/
CN/) directly download, geometric accurate correction, atmospheric correction are carried out to the HSI images of download, then according to the latitude and longitude coordinates of actual measurement
Extract the hyper spectral reflectance data of each sample point.With the continuous development of satellite technology, available satellite EO-1 hyperion number
It will be on the increase according to source.
That is, the hyper spectral reflectance data of each sample point obtained in this example can be actual measurement high-spectral data,
Can be satellite hyper spectral reflectance data.
So-called high-spectral data reflectivity data, compared to common multispectral reflectivity data, high-spectral data reflectivity
The resolution ratio of data is higher, and the spectral resolution of high-spectral data reflectivity is generally less than 10nm, i.e., average just right every 10nm
1 spectral reflectivity is answered, the spectral resolution for surveying hyper spectral reflectance generally to be preferred over 5nm.
In addition, the hyper spectral reflectance of the embodiment of the present invention is corresponding different-waveband, its wavelength band is in 400-850nm
In the range of.
And in order to avoid the ripple of modeling can be caused in conventional art using the partial least square model constructed by full spectrum wave band
Hop count is much larger than sampled point number, and conllinear sex chromosome mosaicism between wave band be present and cause the uncertain aggravation of partial least square model
The problem of, the wave band of hyper spectral reflectance data can be carried out twice by 102~step 105 of following step in the present embodiment
Screening, so as to avoid the conllinear sex chromosome mosaicism between wave band, reduce the uncertainty of partial least square model modeling.Specifically:
Step 102, first time screening is carried out to the wave band of the hyper spectral reflectance data using successive projection algorithm, obtained
To first group of wave band;
Wherein it is possible to using successive projection algorithm (SPA, successive projections algorithm) to step
The wave band of the hyper spectral reflectance got in 101 carries out first time screening, so as to remove some wave bands, obtains first group of ripple
Section, wherein, first group of wave band here includes multiple wave bands, but the quantity of the plurality of wave band is certainly less than in step 101 and obtained
The wave band sum for the hyper spectral reflectance got.
Step 103, according to the measured concentration of the water quality parameter and the hyper spectral reflectance of corresponding first group of wave band
Data, build the first offset minimum binary inverse model;
Wherein it is possible to according to the measured concentration of the water quality parameter of sample point and the EO-1 hyperion got in a step 101
Extracted in reflectivity data to should first group of wave band hyper spectral reflectance data, it is anti-to build the first offset minimum binary
Drill model.
Hyper spectral reflectance used not full spectrum wave band when so, due to the offset minimum binary inverse model of structure, and
It is the hyper spectral reflectance after wave band screens, so as to reduce the wave band number needed for offset minimum binary modeling, reduces institute
The uncertainty of the offset minimum binary inverse model of structure.
Step 104, the hyper spectral reflectance data of corresponding first group of wave band are calculated in first offset minimum binary
Contribution margin in inverse model;
Wherein it is possible to the hyper spectral reflectance of each wave band after first time screens is calculated in constructed minimum partially
Two multiply the contribution margin in inverse model, wherein, the contribution margin of the hyper spectral reflectance data of each wave band is to pass through step 102
First time screening after each wave band hyper spectral reflectance to caused by the precision of the first offset minimum binary inverse model
Contribution, the corresponding contribution margin of hyper spectral reflectance of each wave band.
Step 105, second is carried out to first group of wave band of the hyper spectral reflectance data according to the contribution margin
Secondary screening, obtain corresponding second group of wave band;
Then, in order to ensure getting the hyper spectral reflectance of optimal wave band, there is also the need to screened according in step 102
The contribution margin of the hyper spectral reflectance of obtained each wave band after first time screens come to obtaining first group of wave band progress two
Secondary screening, so as to after filtering some wave bands, obtain second group of wave band.
Wherein, the quantity of the wave band included by second group of wave band is less than the wave band included by first group of wave band
Quantity;
Step 106, using partial least squares algorithm and according to the measured concentration of the water quality parameter and corresponding described second
The hyper spectral reflectance data of group wave band, build the second offset minimum binary inverse model;
So, the optimal wave band for creating offset minimum binary inverse model is had been obtained for by step 105, needed here
The hyper spectral reflectance data for the corresponding second group of wave band of hyper spectral reflectance extracting data to be got from step 101,
And obtained using partial least squares algorithm according to the measured concentration of the water quality parameter got in step 101 and after screening twice
Each wave band hyper spectral reflectance data, to build offset minimum binary inverse model.Constructed minimum partially in this step
Two multiply the retrieving concentration that inverse model is just really used for water quality parameter.
Step 107, the sample point is corresponded to the hyper spectral reflectance data input of second group of wave band to described
Second offset minimum binary inverse model carries out retrieving concentration, obtains the retrieving concentration value of the water quality parameter.
Wherein it is possible to the hyper spectral reflectance of the different-waveband of the water body of sample point accessed from step 101
The hyper spectral reflectance corresponding to the wave band that step 102 and step 105 are screened and finally determined twice is passed through in selection in data, will
The offset minimum binary inverse model that the hyper spectral reflectance of selection is inputted into step 106 carries out the retrieving concentration of water quality parameter,
So as to obtain the retrieving concentration value for water quality parameter described in the sample point.
The embodiment of the present invention is anti-by using the full wave EO-1 hyperion of sample point of the successive projection algorithm to getting
The first time screening that rate carries out wave band is penetrated, and the hyper spectral reflectance based on the corresponding wave band after first time screening is created at it
Offset minimum binary inverse model in contribution margin carry out the wave band after being screened to first time and carry out postsearch screening, it is final so as to obtain
Corresponding wave band for model construction;And hyper spectral reflectance based on the corresponding wave band after postsearch screening and sample point
The measured concentration of water quality parameter carries out the structure of offset minimum binary inverse model, realizes the retrieving concentration to water quality parameter, keeps away
Exempt from the synteny influence between spectrum in modeling, significantly reduce modeling wave band number, it is dense so as to reduce progress water quality parameter
Spend the uncertainty of the partial least square model modeling of inverting.
Alternatively, in one embodiment, when performing step 102, can be accomplished by the following way:
For example, the hyper spectral reflectance data got in step 101 are actual measurement hyper spectral reflectance data.
Here it is possible to which multiple sampled points (such as 40) are divided into two groups, 28 sampled points of random selection are used for model structure
Build, remaining 12 sampled points are used to verify model accuracy.Therefore, can be high by the actual measurement at multiple sampled points (40) place of acquisition
The measured concentration of spectral reflectivity and corresponding water quality parameter is divided into two groups, the actual measurement hyper spectral reflectance of one group of 28 sampled point
It is used for model construction with the measured concentration of corresponding water quality parameter, the actual measurement hyper spectral reflectance of another group of 12 sampled point and right
The measured concentration for the water quality parameter answered is used to verify model accuracy.
Wherein, water quality parameter includes chlorophyll-a concentration, total suspended matter concentration and turbidity, and water quality parameter is referred to below can
To refer to any one in above-mentioned three kinds of water quality parameters;The wavelength band of accessed hyper spectral reflectance in step 101
Generally 400-850nm, because the water body spectrum later to 850nm has the strong water quality for absorbing, causing spectral reflectivity to carry
Information is less.Therefore, multiple ripples in the substantially corresponding 400-850nm wavelength bands of the hyper spectral reflectance of each sampled point
Section.
Hyper spectral reflectance (actual measurement hyper spectral reflectance or satellite hyper spectral reflectance) to modeling sampled point, structure is high
Spectral reflectance rate matrix Xm×n(m is modeling sampled point number, and n is the wave band number of the hyper spectral reflectance of sample point);
From high spectrum reflection rate matrix Xm×nIn randomly select 1 row xj(j=1,2 ..., n), is designated as xs(i), wherein, s (i)
=j, with s (i) (i=1,2, N, N<M-1) record chooses band po sition (wavelength), and i is chooses wave band to number, xs(1)Make
For initial projections vector, now i=1;
Unchecked wave band is included in set H,X is calculated respectivelyjTo collection
Close the projection of the spectral vector of H medium wave bands:
Pxj=xj-(xj Txs(i))xs(i)(xs(i) Txs(i))-1;
Wherein,For projection operator, xjRepresent not selected spectral vector, xs(i)For projection vector, i is to choose wave band
Number, now i=1.
ChooseIn maximal projection wave band, recorded using s (i), now i=2, by its projection valueChanged as next time
The projection vector in generation:
Wherein,RepresentNorm, arg max represent and take and makeMaximum j.
Utilize projection operatorInstead of the spectral vector x of set H medium wave bandsj:
Assignment i=i+1, recalculate Projection vectorProjection to the spectral vector of set H medium wave bands, until i=N,
Now, the hyper spectral reflectance of the different-waveband of each modeling sampled point in step 101 is screened to obtain height by SPA algorithms
The wave band of spectral reflectivity is { s (1), s (2) ..., s (N) }.
In the present embodiment, it is described in detail and how the hyper spectral reflectance data of different-waveband is carried out using SPA algorithms
Wave band screens, and may thereby determine that the hyper spectral reflectance data of certain wave band, for using the specific of SPA algorithms screening wave band
Details is prior art, be will not be repeated here.
Alternatively, in another embodiment, can be by following sub-step S1~S3 come real when performing step 103
It is existing:
S1, from the multiple principal components of hyper spectral reflectance extracting data of corresponding first group of wave band;
Specifically include sub-step SS1~SS4;
SS1, according to the hyper spectral reflectance data of the correspondence of the sample point water body first group of wave band, establish light
Spectrum matrix;
For example, obtain N group spectral bands based on SPA preliminary screenings, modeling data collection (including each modeling sampled point is utilized
Locate the hyper spectral reflectance data of each wave band of water body) determine the hyper spectral reflectance corresponding to the N group spectral bands and conduct
Independent variable (spectrum matrix) A, makes A={ x1,x2,…,xN};Wherein A every a line represents a modeling sampled point, each row generation
Table correspondingly models hyper spectral reflectance of the sampled point at the wave band;
SS2, according to the measured concentration of the water quality parameter of the sample point, establish water quality parameter matrix;
With corresponding water quality ginseng in modeling data collection (measured concentration for including each modeling sample point water quality parameter)
It is a column vector that number, which is used as dependent variable (water quality parameter matrix) Y, Y={ y }, dependent variable Y, represents the water quality of modeling sample point
The measured concentration of parameter.
SS3, according to the spectrum matrix and the water quality parameter matrix, calculate extraction coefficient;
Specifically, spectrum matrix A, water quality parameter matrix Y average and standard deviation is utilized respectively to be standardized A, Y
Processing (such as data normalization --- z-score) obtains matrix E0(A), i.e. E0(xi,j), matrix F0(Y), i.e. F0(yi), it is such as public
Shown in formula (1) and formula (2):
Wherein, formula (1), on the right of formula (2) equal sign in xi,j,yiI-th of modeling sampled point is represented respectively in j ripples
The measured concentration of the water quality parameter of hyper spectral reflectance and i-th of modeling sampled point at section,Represent all modeling sampled points
Hyper spectral reflectance of the spectrum at j wave bands average,Represent the measured concentration of the water quality parameter of all modeling sampled points
Average, Std (xj) represent modeling sampled point spectrum reflectivity at j wave bands standard deviation, Std (y) represent modeling sampling
The standard deviation of the measured concentration of the water quality parameter of point;The x on the equal sign left sidei,j,yiRepresent respectively to i-th of modeling sampled point in j ripples
The measured concentration standardization result of the water quality parameter of hyper spectral reflectance and i-th of modeling sampled point at section, it is respectively
Matrix E0(A), matrix F0(Y) element in.
According to E0, F0Calculate extraction coefficient W1, W1For matrix E0 T·F0·F0 T·E0Eigenvalue of maximum corresponding to feature
Vector, wherein, " " represents matrix multiplication.WiBy matrix E0And matrix F0Together decide on.
SS4, multiple principal components are extracted from the spectrum matrix according to the extraction coefficient.
From matrix E0First principal component t of middle extraction1:
t1=E0·W1;
Calculate E0To t1Regression equation, calculate residual matrix E1=E0-t1·P1 T, wherein, P1=E0 T·t1/||t1||2;
With residual matrix E1And E1 TInstead of matrix E0 T·F0·F0 T·E0In E0And E0 TMatrix multiplication operation is carried out to obtain
To W2, from E1Second principal component t of middle extraction2:
t2=E1·W2;
Wherein, W2For E1 T·F0·F0 T·E1Eigenvalue of maximum corresponding to characteristic vector;
Calculate E1To t2Regression equation, calculate residual matrix E2=E1-t2·P2 T, wherein P2=E1 T·t2/||t2||2;
Similarly, continue to extract principal component t3,…,th;
S2, the cross-verification method in partial least squares algorithm is replaced with into randomization test, obtain improving offset minimum binary
Algorithm, calculating analysis is carried out to the multiple principal component of extraction using the improvement partial least squares algorithm, determined described more
Optimal principal component in individual principal componentWherein hoptimal<H, and determine the number h of the optimal principal componentoptimal;
Wherein it is possible to it will be used to determine that the cross-verification method of optimal principal component replaces with randomized test in PLS
Method, to improve partial least squares algorithm, and using the randomization test in the improvement partial least squares algorithm, determine described more
Optimal principal component in individual principal component, and determine the number h of the optimal principal componentoptimal;
Specifically, h principal component, i.e. t are extracted in S11,t2,…,th, it is possible, firstly, to calculate each principal component and warp
Water quality parameter matrix F after standardization0Between covariance Cov:
Covp=cov (tp,F0), p=1,2 ..., h;
By F0Middle each element carries out random alignment, generates K arrangement, is designated as F0,K, calculate be based on E respectively0And F0,KExtraction
Principal component t1,K,t2,K,…,th,KAnd F0,KBetween covariance:
Covp,K=cov (tp,K,F0,K), p=1,2 ..., h;
Count CovpAnd Covp,KBetween difference:
Dp,K=Covp,k-Covp,K;
Calculate contribution rate R of p-th of principal component to partial least square models(p):
Rs(p)=(Vp/V1)×100;
Vp=std (Dp,K);
Calculate the Relative Contribution rate R between principal componentr(p)=Rs(p)/Rs(p+1) R, is drawnr(p) curve, if working as Rr(p)
When curve tends towards stability, p now is arranged to hoptimal, and determine hoptimalFor optimal principal component number,
For the optimal principal component of extraction.
S3, according to the measured concentration of the water quality parameter and the optimal principal component of the number, build a first inclined most young waiter in a wineshop or an inn
Multiply inverse model.
According to the optimal principal component of determination and the measured concentration of the water quality parameter, established using PLS polynary
Regression equation, obtain the first offset minimum binary inverse model;
Specifically, according to the optimal principal component of determinationWith the water quality parameter matrix after normalized processing
F0Establish multiple regression equation:
Wherein,It is E0Linear combination, ti=Ei-1·Wi=E0·Wi *,Will
Wi *Substitute into ti:
Make y*=F0, x* i=E0i,J=1,2 ..., N, substitute into
Then y*=β1·x1 *+β2·x2 *+…+βN·xN *
According to y*=β1·x1 *+β2·x2 *+…+βN·xN *Calculate a PLS equation (i.e. first inclined most young waiter in a wineshop or an inn
Multiply inverse model) be:
Wherein, E is represented and averaged, SyWithDependent variable and the standard deviation of each row of independent variable are represented respectively.
Alternatively, in one embodiment, can be by following sub-step S21~S25 come real when performing step 104
It is existing:
S21, each wave band in first group of wave band is removed successively, form F band combination, each band combination bag
Include F-1 wave band;
For example, first group of wave band includes F wave band, remove successively the F wave band, the F-1 wave band ..., the 1st
Wave band, so as to obtain F band combination, each band combination includes F-1 wave band.
S22, by hyper spectral reflectance data input corresponding to the measured concentration of the water quality parameter and each band combination
To the first offset minimum binary inverse model, to build the offset minimum binary inverse model of F kinds the 3rd;
As an instantiation, can be selected from modeling spectroscopic data collection (being generated by step 101) each
Band combination corresponds to the measured concentration of the spectral reflectivity of wave band and the water quality parameter of modeling sampled point;Then, by a wave band
The measured concentration of the water quality parameter of the spectral reflectivity of combination and modeling sampled point input into above-described embodiment described first
Offset minimum binary inverse model (such as the PLS equation being finally calculated in above-described embodiment), so as to
Build a kind of the 3rd offset minimum binary inverse model of band combination;Similarly, the offset minimum binary inverting mould of F kinds the 3rd can be calculated
Type.
S23, according to the measured concentration of the water quality parameter and the first offset minimum binary inverse model to the actual measurement
The inverting concentration of concentration, calculate the first residual sum of squares (RSS) of the first offset minimum binary inverse model;
As an instantiation, the measured concentration of the water quality parameter of each modeling sampled point can be inputted to step
103 the first offset minimum binary inverse models established carry out retrieving concentration, anti-so as to obtain the concentration of corresponding modeling sampled point
Drill value (i.e. inverting concentration);Then, using equation below, according to the measured concentration of the water quality parameter of modeling sampled point and corresponding
Retrieving concentration value calculates the residual sum of squares (RSS) of the first offset minimum binary inverse model:
Wherein,Be i-th modeling sample point water quality parameter measured concentration in the first offset minimum binary inverse model
In retrieving concentration value, yiFor the concentration measured value of i-th of modeling sample point water quality parameter, N is the number of modeling sampled point.
S24, according to the measured concentration of the water quality parameter and every kind of 3rd offset minimum binary inverse model to the actual measurement
The inverting concentration of concentration, calculate the 3rd residual sum of squares (RSS) of every kind of 3rd offset minimum binary inverse model;
As an instantiation, similarly, with reference to sub-step S23, F kinds the can also be calculated using the formula in S23
Three offset minimum binary inverse models (correspond to respectively successively remove come the F wave band, the F-1 wave band ..., the 1st wave band
Band combination) the 3rd residual sum of squares (RSS), be designated as V respectivelyF,VF-1,…,V1。
S25, according to first residual sum of squares (RSS) and the 3rd residual error of every kind of 3rd offset minimum binary inverse model
Quadratic sum, calculate the contribution margin of the wave band being removed corresponding to every kind of 3rd offset minimum binary inverse model.
, can be by equation below, to calculate the wave band being removed in each band combination corresponding as an example
Contribution margin in 3rd offset minimum binary inverse model:
Ci=(Vi-VTotal)/VTotal, i=1,2 ..., F.
Wherein, it is individual for the 3rd offset minimum binary inverse model corresponding to the band combination of the F wave band of removal, the F
Wave band is C to the contribution margin of the 3rd offset minimum binary inverse modelF;Corresponded to for the band combination for removing the F-1 wave band
The 3rd offset minimum binary inverse model, the F-1 wave band be to the contribution margin of the 3rd offset minimum binary inverse model
CF-1... ..., for removing the 3rd offset minimum binary inverse model corresponding to the band combination of the 1st wave band, the 1st wave band
Contribution margin to the 3rd offset minimum binary inverse model is C1。
Alternatively, when performing step 105, can realize in the following way:
Determine the target band combination corresponding to the 3rd minimum offset minimum binary inverse model of contribution margin;
First group of wave band is updated to the target band combination, until first residual sum of squares (RSS) stops subtracting
It is small;
First group of wave band after last time is updated is defined as second group of wave band.
Specifically, can be by the contribution margin C in examples detailed aboveiArranged from big to small, it is determined that minimum contribution margin institute
Corresponding band combination (such as C1Minimum, then illustrate first wave band in first group of wave band can be removed, retain other ripples
Section), the first group of wave band then replaced with the band combination in above-mentioned sub-step S21 (specifically, can be to the band group
Remaining wave band continues executing with above-mentioned sub-step S21~S25 in conjunction, i.e. continues structure using remaining wave band in the band combination
Above-mentioned 3rd offset minimum binary inverse model is built, and calculates the residual sum of squares (RSS) of each 3rd offset minimum binary inverse model, with
And a wave band in the band combination, residual sum of squares (RSS) corresponding to calculating and corresponding contribution margin are removed successively), continue executing with
In this step 105, the minimum wave band of contribution margin is rejected, i.e. be updated to band combination corresponding to contribution margin minimum new
First group of wave band, loop iteration, until residual sum of squares (RSS) VTotalWhen starting increase, circulation is terminated, retains remaining wave band.
To sum up, the embodiment of the present invention determines that partial least square model is calculated using randomization test instead of cross verification
Optimal principal component number, partial least square model is improved;And carried out using the partial least square model after improving
During water quality remote-sensing inversion, then the measured concentration information of modeling sample spectrum and water quality parameter can be comprehensively utilized, prevents that water quality is distant
Feel partial least square model over-fitting or poor fitting;
In addition, the spectral band needed for being modeled by using successive projection algorithm to partial least square model screens,
The synteny between spectral band can be eliminated;Calculating sifting wave band is to the contribution margin of the first partial least square model, final determination
For the best band of the modeling of the second offset minimum binary inverse model of water quality remote sensing, offset minimum binary mould can be significantly reduced
Type models wave band number, the modeling stability of the partial least square model for remote sensing of increasing water quality.
It should be noted that for embodiment of the method, in order to be briefly described, therefore it is all expressed as to a series of action group
Close, but those skilled in the art should know, the embodiment of the present invention is not limited by described sequence of movement, because according to
According to the embodiment of the present invention, some steps can use other orders or carry out simultaneously.Secondly, those skilled in the art also should
Know, embodiment described in this description belongs to preferred embodiment, and the involved action not necessarily present invention is implemented
Necessary to example.
It is corresponding with the method that the embodiments of the present invention are provided, reference picture 2, show that a kind of multiband of the present invention is excellent
Change the structured flowchart of the water quality parameter retrieving concentration device embodiment of combination, can specifically include following module:
Acquisition module 21, for obtaining the measured concentration of sample point water quality parameter and the correspondence of the sample point water body
The hyper spectral reflectance data of different-waveband;
First screening module 22, for carrying out the to the wave bands of the hyper spectral reflectance data using successive projection algorithm
Primary screening, first group of wave band is obtained, wherein, first group of wave band includes multiple wave bands;
First structure module 23, for the measured concentration according to the water quality parameter and the height of corresponding first group of wave band
Spectral reflectance data, build the first offset minimum binary inverse model;
Computing module 24, for calculating the hyper spectral reflectance data of corresponding first group of wave band described first partially most
A young waiter in a wineshop or an inn multiplies the contribution margin in inverse model;
Second screening module 25, for first group of ripple according to the contribution margin to the hyper spectral reflectance data
Duan Jinhang programmed screenings, corresponding second group of wave band is obtained, wherein, the quantity of the wave band included by second group of wave band is less than
The quantity of wave band included by first group of wave band;
Second structure module 26, for using partial least squares algorithm and according to the measured concentration of the water quality parameter and right
The hyper spectral reflectance data of second group of wave band are answered, build the second offset minimum binary inverse model;
Inverting module 27, for the sample point to be corresponded to the hyper spectral reflectance data input of second group of wave band
Retrieving concentration is carried out to the second offset minimum binary inverse model, obtains the retrieving concentration value of the water quality parameter.
Alternatively, the first structure module 23 includes:
Extracting sub-module, for the hyper spectral reflectance extracting data from corresponding first group of wave band it is multiple it is main into
Point;
Submodule is replaced, for the cross-verification method in partial least squares algorithm to be replaced with into randomization test, is changed
Enter partial least squares algorithm;
Analysis submodule is calculated, for entering using the improvement partial least squares algorithm to the multiple principal component of extraction
Row calculates analysis, determines optimal principal component in the multiple principal component, and determines the number of the optimal principal component;
First structure submodule, for the measured concentration according to the water quality parameter and the number it is described it is optimal it is main into
Point, build the first offset minimum binary inverse model.
Alternatively, the extracting sub-module includes:
First establishes unit, the high spectrum reflection for correspondence first group of wave band according to the sample point water body
Rate data, establish spectrum matrix;
Second establishes unit, for the measured concentration of the water quality parameter according to the sample point, establishes water quality parameter square
Battle array;
Computing unit, for according to the spectrum matrix and the water quality parameter matrix, calculating extraction coefficient;
Extraction unit, for extracting multiple principal components from the spectrum matrix according to the extraction coefficient.
Alternatively, first group of wave band includes F wave band, and the computing module 24 includes:
Submodule is removed, for removing each wave band in first group of wave band successively, forms F band combination, often
Individual band combination includes F-1 wave band;
Second structure submodule, for by EO-1 hyperion corresponding to the measured concentration of the water quality parameter and each band combination
Reflectivity data is inputted to the first offset minimum binary inverse model, to build the offset minimum binary inverse model of F kinds the 3rd;
First calculating sub module, for the measured concentration according to the water quality parameter and the first offset minimum binary inverting
Inverting concentration of the model to the measured concentration, calculate the first residual sum of squares (RSS) of the first offset minimum binary inverse model;
Second calculating sub module, for the measured concentration according to the water quality parameter and every kind of 3rd offset minimum binary inverting
Inverting concentration of the model to the measured concentration, calculate the 3rd residual sum of squares (RSS) of every kind of 3rd offset minimum binary inverse model;
3rd calculating sub module, for according to first residual sum of squares (RSS) and every kind of 3rd offset minimum binary inverting
3rd residual sum of squares (RSS) of model, calculate the contribution of the wave band being removed corresponding to every kind of 3rd offset minimum binary inverse model
Value.
Alternatively, second screening module 25 includes:
First determination sub-module, the target corresponding to the 3rd offset minimum binary inverse model minimum for determining contribution margin
Band combination;
Submodule is updated, for first group of wave band to be updated into the target band combination, until described first residual
Poor quadratic sum stops reducing;
Second determination sub-module, it is defined as second group of wave band for first group of wave band after last time is updated.
For device embodiment, because it is substantially similar to embodiment of the method, so description is fairly simple, it is related
Part illustrates referring to the part of embodiment of the method.
Each embodiment in this specification is described by the way of progressive, what each embodiment stressed be with
The difference of other embodiment, between each embodiment identical similar part mutually referring to.
It should be understood by those skilled in the art that, the embodiment of the embodiment of the present invention can be provided as method, apparatus or calculate
Machine program product.Therefore, the embodiment of the present invention can use complete hardware embodiment, complete software embodiment or combine software and
The form of the embodiment of hardware aspect.Moreover, the embodiment of the present invention can use one or more wherein include computer can
With in the computer-usable storage medium (including but is not limited to magnetic disk storage, CD-ROM, optical memory etc.) of program code
The form of the computer program product of implementation.
The embodiment of the present invention is with reference to method according to embodiments of the present invention, terminal device (system) and computer program
The flow chart and/or block diagram of product describes.It should be understood that can be by computer program instructions implementation process figure and/or block diagram
In each flow and/or square frame and the flow in flow chart and/or block diagram and/or the combination of square frame.These can be provided
Computer program instructions are set to all-purpose computer, special-purpose computer, Embedded Processor or other programmable data processing terminals
Standby processor is to produce a machine so that is held by the processor of computer or other programmable data processing terminal equipments
Capable instruction is produced for realizing in one flow of flow chart or multiple flows and/or one square frame of block diagram or multiple square frames
The device for the function of specifying.
These computer program instructions, which may be alternatively stored in, can guide computer or other programmable data processing terminal equipments
In the computer-readable memory to work in a specific way so that the instruction being stored in the computer-readable memory produces bag
The manufacture of command device is included, the command device is realized in one flow of flow chart or multiple flows and/or one side of block diagram
The function of being specified in frame or multiple square frames.
These computer program instructions can be also loaded into computer or other programmable data processing terminal equipments so that
Series of operation steps is performed on computer or other programmable terminal equipments to produce computer implemented processing, so that
The instruction performed on computer or other programmable terminal equipments is provided for realizing in one flow of flow chart or multiple flows
And/or specified in one square frame of block diagram or multiple square frames function the step of.
Although having been described for the preferred embodiment of the embodiment of the present invention, those skilled in the art once know base
This creative concept, then other change and modification can be made to these embodiments.So appended claims are intended to be construed to
Including preferred embodiment and fall into having altered and changing for range of embodiment of the invention.
Finally, it is to be noted that, herein, such as first and second or the like relational terms be used merely to by
One entity or operation make a distinction with another entity or operation, and not necessarily require or imply these entities or operation
Between any this actual relation or order be present.Moreover, term " comprising ", "comprising" or its any other variant meaning
Covering including for nonexcludability, so that process, method, article or terminal device including a series of elements are not only wrapped
Those key elements, but also the other element including being not expressly set out are included, or is also included for this process, method, article
Or the key element that terminal device is intrinsic.In the absence of more restrictions, wanted by what sentence "including a ..." limited
Element, it is not excluded that other identical element in the process including the key element, method, article or terminal device also be present.
It is more to the water quality parameter retrieving concentration method and one kind of a kind of multiband optimum organization provided by the present invention above
The water quality parameter retrieving concentration device of wave band optimum organization, is described in detail, and specific case used herein is to this hair
Bright principle and embodiment is set forth, the explanation of above example be only intended to help the method for understanding the present invention and its
Core concept;Meanwhile for those of ordinary skill in the art, according to the thought of the present invention, in embodiment and application
There will be changes in scope, in summary, this specification content should not be construed as limiting the invention.
Claims (10)
1. a kind of water quality parameter retrieving concentration method of multiband optimum organization, it is characterised in that including:
The EO-1 hyperion for obtaining the measured concentration of sample point water quality parameter and the corresponding different-waveband of the sample point water body is anti-
Penetrate rate data;
First time screening is carried out to the wave band of the hyper spectral reflectance data using successive projection algorithm, obtains first group of ripple
Section, wherein, first group of wave band includes multiple wave bands;
According to the measured concentration of the water quality parameter and the hyper spectral reflectance data of corresponding first group of wave band, structure first
Offset minimum binary inverse model;
The hyper spectral reflectance data of corresponding first group of wave band are calculated in the first offset minimum binary inverse model
Contribution margin;
Programmed screening is carried out to first group of wave band of the hyper spectral reflectance data according to the contribution margin, obtained pair
Second group of wave band is answered, wherein, the quantity of the wave band included by second group of wave band is less than included by first group of wave band
The quantity of wave band;
Using partial least squares algorithm and according to the measured concentration of the water quality parameter and the bloom of corresponding second group of wave band
Reflectivity data is composed, builds the second offset minimum binary inverse model;
The sample point is corresponded into the hyper spectral reflectance data input of second group of wave band to a described second inclined most young waiter in a wineshop or an inn
Multiply inverse model and carry out retrieving concentration, obtain the retrieving concentration value of the water quality parameter.
2. according to the method for claim 1, it is characterised in that the measured concentration according to the water quality parameter and correspondingly
The hyper spectral reflectance data of first group of wave band, the first offset minimum binary inverse model is built, including:
From the multiple principal components of hyper spectral reflectance extracting data of corresponding first group of wave band;
Cross-verification method in partial least squares algorithm is replaced with into randomization test, obtains improving partial least squares algorithm;
Calculating analysis is carried out to the multiple principal component of extraction using the improvement partial least squares algorithm, determined the multiple
Optimal principal component in principal component, and determine the number of the optimal principal component;
It is anti-according to the measured concentration of the water quality parameter and the optimal principal component of the number, the first offset minimum binary of structure
Drill model.
3. according to the method for claim 2, it is characterised in that the high spectrum reflection from corresponding first group of wave band
The multiple principal components of rate extracting data, including:
According to the hyper spectral reflectance data of the correspondence of the sample point water body first group of wave band, spectrum matrix is established;
According to the measured concentration of the water quality parameter of the sample point, water quality parameter matrix is established;
According to the spectrum matrix and the water quality parameter matrix, extraction coefficient is calculated;
Multiple principal components are extracted from the spectrum matrix according to the extraction coefficient.
4. according to the method for claim 1, it is characterised in that first group of wave band includes F wave band, the calculating pair
Contribution margin of the hyper spectral reflectance data of first group of wave band in the first offset minimum binary inverse model is answered, is wrapped
Include:
Each wave band in first group of wave band is removed successively, forms F band combination, and each band combination includes F-1
Wave band;
By hyper spectral reflectance data input corresponding to the measured concentration of the water quality parameter and each band combination to described
One offset minimum binary inverse model, to build the offset minimum binary inverse model of F kinds the 3rd;
According to the measured concentration of the water quality parameter and the first offset minimum binary inverse model to the anti-of the measured concentration
Concentration is drilled, calculates the first residual sum of squares (RSS) of the first offset minimum binary inverse model;
According to the measured concentration of the water quality parameter and every kind of 3rd offset minimum binary inverse model to the anti-of the measured concentration
Concentration is drilled, calculates the 3rd residual sum of squares (RSS) of every kind of 3rd offset minimum binary inverse model;
According to first residual sum of squares (RSS) and the 3rd residual sum of squares (RSS) of every kind of 3rd offset minimum binary inverse model, meter
Calculate the contribution margin of the wave band being removed corresponding to every kind of 3rd offset minimum binary inverse model.
5. according to the method for claim 4, it is characterised in that it is described according to the contribution margin to the hyper spectral reflectance
First group of wave band of data carries out programmed screening, obtains corresponding second group of wave band, including:
Determine the target band combination corresponding to the 3rd minimum offset minimum binary inverse model of contribution margin;
First group of wave band is updated to the target band combination, until first residual sum of squares (RSS) stops reducing;
First group of wave band after last time is updated is defined as second group of wave band.
A kind of 6. water quality parameter retrieving concentration device of multiband optimum organization, it is characterised in that including:
Acquisition module, for obtaining the measured concentration of sample point water quality parameter and the different ripples of correspondence of the sample point water body
The hyper spectral reflectance data of section;
First screening module, for carrying out first time sieve to the wave band of the hyper spectral reflectance data using successive projection algorithm
Choosing, obtains first group of wave band, wherein, first group of wave band includes multiple wave bands;
First structure module, the EO-1 hyperion for the measured concentration according to the water quality parameter and corresponding first group of wave band are anti-
Rate data are penetrated, build the first offset minimum binary inverse model;
Computing module, for calculating the hyper spectral reflectance data of corresponding first group of wave band in first offset minimum binary
Contribution margin in inverse model;
Second screening module, for being carried out according to the contribution margin to first group of wave band of the hyper spectral reflectance data
Programmed screening, corresponding second group of wave band is obtained, wherein, the quantity of the wave band included by second group of wave band is less than described the
The quantity of wave band included by one group of wave band;
Second structure module, for use partial least squares algorithm and described in the measured concentration according to the water quality parameter and correspondence
The hyper spectral reflectance data of second group of wave band, build the second offset minimum binary inverse model;
Inverting module, for the sample point to be corresponded to the hyper spectral reflectance data input of second group of wave band to described
Second offset minimum binary inverse model carries out retrieving concentration, obtains the retrieving concentration value of the water quality parameter.
7. device according to claim 6, it is characterised in that the first structure module includes:
Extracting sub-module, for the multiple principal components of hyper spectral reflectance extracting data from corresponding first group of wave band;
Submodule is replaced, for the cross-verification method in partial least squares algorithm to be replaced with into randomization test, obtains improving partially
Least-squares algorithm;
Analysis submodule is calculated, based on being carried out using the improvement partial least squares algorithm to the multiple principal component of extraction
Point counting is analysed, and determines optimal principal component in the multiple principal component, and determines the number of the optimal principal component;
First structure submodule, for the measured concentration according to the water quality parameter and the optimal principal component of the number,
Build the first offset minimum binary inverse model.
8. device according to claim 7, it is characterised in that the extracting sub-module includes:
First establishes unit, the hyper spectral reflectance number for correspondence first group of wave band according to the sample point water body
According to establishing spectrum matrix;
Second establishes unit, for the measured concentration of the water quality parameter according to the sample point, establishes water quality parameter matrix;
Computing unit, for according to the spectrum matrix and the water quality parameter matrix, calculating extraction coefficient;
Extraction unit, for extracting multiple principal components from the spectrum matrix according to the extraction coefficient.
9. device according to claim 6, it is characterised in that first group of wave band includes F wave band, the calculating mould
Block includes:
Submodule is removed, for removing each wave band in first group of wave band successively, forms F band combination, Mei Gebo
Duan Zuhe includes F-1 wave band;
Second structure submodule, for by high spectrum reflection corresponding to the measured concentration of the water quality parameter and each band combination
Rate data input is to the first offset minimum binary inverse model, to build the offset minimum binary inverse model of F kinds the 3rd;
First calculating sub module, for the measured concentration according to the water quality parameter and the first offset minimum binary inverse model
Inverting concentration to the measured concentration, calculate the first residual sum of squares (RSS) of the first offset minimum binary inverse model;
Second calculating sub module, for the measured concentration according to the water quality parameter and every kind of 3rd offset minimum binary inverse model
Inverting concentration to the measured concentration, calculate the 3rd residual sum of squares (RSS) of every kind of 3rd offset minimum binary inverse model;
3rd calculating sub module, for according to first residual sum of squares (RSS) and every kind of 3rd offset minimum binary inverse model
The 3rd residual sum of squares (RSS), calculate the contribution margin of the wave band being removed corresponding to every kind of 3rd offset minimum binary inverse model.
10. device according to claim 9, it is characterised in that second screening module includes:
First determination sub-module, the target wave band corresponding to the 3rd offset minimum binary inverse model minimum for determining contribution margin
Combination;
Submodule is updated, for first group of wave band to be updated into the target band combination, until first residual error is put down
Side and stopping reduce;
Second determination sub-module, it is defined as second group of wave band for first group of wave band after last time is updated.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710527421.3A CN107421895A (en) | 2017-06-30 | 2017-06-30 | A kind of water quality parameter retrieving concentration method and apparatus of multiband optimum organization |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710527421.3A CN107421895A (en) | 2017-06-30 | 2017-06-30 | A kind of water quality parameter retrieving concentration method and apparatus of multiband optimum organization |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107421895A true CN107421895A (en) | 2017-12-01 |
Family
ID=60427794
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710527421.3A Pending CN107421895A (en) | 2017-06-30 | 2017-06-30 | A kind of water quality parameter retrieving concentration method and apparatus of multiband optimum organization |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107421895A (en) |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108195766A (en) * | 2017-12-18 | 2018-06-22 | 河海大学 | A kind of water quality monitoring method based on remote sensing image |
CN110567905A (en) * | 2019-09-04 | 2019-12-13 | 湖南匡楚科技有限公司 | Water quality detection method and device |
CN110988286A (en) * | 2019-12-18 | 2020-04-10 | 松辽水资源保护科学研究所 | Intelligent water resource long-term detection system |
CN111239071A (en) * | 2020-02-19 | 2020-06-05 | 中国科学院烟台海岸带研究所 | Method for detecting concentration of nitrate in seawater by spectrometry |
CN111289446A (en) * | 2020-03-30 | 2020-06-16 | 天津工业大学 | Method and system for detecting component concentration of complex solution |
CN111504915A (en) * | 2020-04-27 | 2020-08-07 | 中国科学技术大学先进技术研究院 | Method, device and equipment for inverting chlorophyll concentration of water body and storage medium |
CN111811998A (en) * | 2020-09-01 | 2020-10-23 | 中国人民解放军国防科技大学 | Method for determining strongly-absorbable biological particle component under target waveband |
CN111879709A (en) * | 2020-07-15 | 2020-11-03 | 中国科学院空天信息创新研究院 | Method and device for detecting spectral reflectivity of lake water body |
CN112215525A (en) * | 2020-11-04 | 2021-01-12 | 安徽农业大学 | Lake and reservoir water quality inversion and visual evaluation method |
CN112378866A (en) * | 2021-01-18 | 2021-02-19 | 航天宏图信息技术股份有限公司 | Water quality parameter inversion model training method, water quality monitoring method and device |
CN113340825A (en) * | 2021-06-17 | 2021-09-03 | 重庆大学 | Method for measuring and calculating chlorophyll a concentration under high-turbidity background interference |
CN117451639A (en) * | 2023-12-21 | 2024-01-26 | 内蒙古工业大学 | Water chlorophyll concentration inversion method based on remote sensing data |
CN117723497A (en) * | 2024-02-08 | 2024-03-19 | 上海阿夸斯科技有限公司 | Intelligent follow-up water quality monitoring system |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101893550A (en) * | 2010-07-14 | 2010-11-24 | 青岛海洋地质研究所 | Semi-analytical method for realizing inversion of water body chlorophyll alpha concentration |
CN102200576A (en) * | 2011-03-10 | 2011-09-28 | 王桥 | Chlorophyll a concentration inversion method and system |
CN103970994A (en) * | 2014-05-05 | 2014-08-06 | 南京师范大学 | Inland water chlorophyll concentration multi-model cooperative retrieval method based on data assimilation |
CN104122210A (en) * | 2014-07-02 | 2014-10-29 | 中国林业科学研究院林业研究所 | Hyperspectral waveband extraction method based on optimum index factor-correlation coefficient method |
CN106525762A (en) * | 2016-11-07 | 2017-03-22 | 航天恒星科技有限公司 | Water quality monitoring method and water quality monitoring device based on adaptive model |
-
2017
- 2017-06-30 CN CN201710527421.3A patent/CN107421895A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101893550A (en) * | 2010-07-14 | 2010-11-24 | 青岛海洋地质研究所 | Semi-analytical method for realizing inversion of water body chlorophyll alpha concentration |
CN102200576A (en) * | 2011-03-10 | 2011-09-28 | 王桥 | Chlorophyll a concentration inversion method and system |
CN103970994A (en) * | 2014-05-05 | 2014-08-06 | 南京师范大学 | Inland water chlorophyll concentration multi-model cooperative retrieval method based on data assimilation |
CN104122210A (en) * | 2014-07-02 | 2014-10-29 | 中国林业科学研究院林业研究所 | Hyperspectral waveband extraction method based on optimum index factor-correlation coefficient method |
CN106525762A (en) * | 2016-11-07 | 2017-03-22 | 航天恒星科技有限公司 | Water quality monitoring method and water quality monitoring device based on adaptive model |
Non-Patent Citations (3)
Title |
---|
曹引 等: "基于离散离子群和偏最小二乘的湖库型水源地水体悬浮物浓度和浊度遥感反演方法", 《水力发电学报》 * |
李丽娜 等: "一种改进的随机检验法用于主成分选择以避免光谱分析校正模型的过拟合或欠拟合", 《光谱学与光谱分析》 * |
高洪治 等: "基于连续投影算法的土壤总氮近红外特征波长的选取", 《光谱学与光谱分析》 * |
Cited By (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108195766A (en) * | 2017-12-18 | 2018-06-22 | 河海大学 | A kind of water quality monitoring method based on remote sensing image |
CN110567905A (en) * | 2019-09-04 | 2019-12-13 | 湖南匡楚科技有限公司 | Water quality detection method and device |
CN110567905B (en) * | 2019-09-04 | 2023-12-29 | 怀化水直达山泉饮用水有限公司 | Water quality detection method and device |
CN110988286A (en) * | 2019-12-18 | 2020-04-10 | 松辽水资源保护科学研究所 | Intelligent water resource long-term detection system |
CN111239071B (en) * | 2020-02-19 | 2023-03-31 | 中国科学院烟台海岸带研究所 | Method for detecting concentration of nitrate in seawater by spectrometry |
CN111239071A (en) * | 2020-02-19 | 2020-06-05 | 中国科学院烟台海岸带研究所 | Method for detecting concentration of nitrate in seawater by spectrometry |
CN111289446A (en) * | 2020-03-30 | 2020-06-16 | 天津工业大学 | Method and system for detecting component concentration of complex solution |
CN111504915A (en) * | 2020-04-27 | 2020-08-07 | 中国科学技术大学先进技术研究院 | Method, device and equipment for inverting chlorophyll concentration of water body and storage medium |
CN111879709B (en) * | 2020-07-15 | 2023-05-30 | 中国科学院空天信息创新研究院 | Lake water body spectral reflectivity inspection method and device |
CN111879709A (en) * | 2020-07-15 | 2020-11-03 | 中国科学院空天信息创新研究院 | Method and device for detecting spectral reflectivity of lake water body |
CN111811998A (en) * | 2020-09-01 | 2020-10-23 | 中国人民解放军国防科技大学 | Method for determining strongly-absorbable biological particle component under target waveband |
CN112215525A (en) * | 2020-11-04 | 2021-01-12 | 安徽农业大学 | Lake and reservoir water quality inversion and visual evaluation method |
CN112215525B (en) * | 2020-11-04 | 2023-06-23 | 安徽农业大学 | Lake and reservoir water quality inversion and visual evaluation method |
CN112378866A (en) * | 2021-01-18 | 2021-02-19 | 航天宏图信息技术股份有限公司 | Water quality parameter inversion model training method, water quality monitoring method and device |
CN113340825A (en) * | 2021-06-17 | 2021-09-03 | 重庆大学 | Method for measuring and calculating chlorophyll a concentration under high-turbidity background interference |
CN117451639A (en) * | 2023-12-21 | 2024-01-26 | 内蒙古工业大学 | Water chlorophyll concentration inversion method based on remote sensing data |
CN117723497A (en) * | 2024-02-08 | 2024-03-19 | 上海阿夸斯科技有限公司 | Intelligent follow-up water quality monitoring system |
CN117723497B (en) * | 2024-02-08 | 2024-04-26 | 上海阿夸斯科技有限公司 | Intelligent follow-up water quality monitoring system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107421895A (en) | A kind of water quality parameter retrieving concentration method and apparatus of multiband optimum organization | |
CN107271405A (en) | A kind of water quality parameter retrieving concentration method and apparatus of multi-model collaboration | |
Houborg et al. | Joint leaf chlorophyll content and leaf area index retrieval from Landsat data using a regularized model inversion system (REGFLEC) | |
CN110455722A (en) | Rubber tree blade phosphorus content EO-1 hyperion inversion method and system | |
CN104573640B (en) | Data processing system and its generating device, type of face detection method | |
RU2008105393A (en) | COMPUTER-BASED APPLICATION FORMATION AND VERIFICATION OF TRAINING IMAGES INTENDED FOR MULTI-POINT THEOSTATISTIC ANALYSIS | |
CN103235095B (en) | Water-injected meat detection method and device | |
CN109325395A (en) | The recognition methods of image, convolutional neural networks model training method and device | |
CN104132897B (en) | A kind of nitrogenous measuring method of plant leaf blade based on handheld device and device | |
Adeluyi et al. | Estimating the phenological dynamics of irrigated rice leaf area index using the combination of PROSAIL and Gaussian Process Regression | |
CN109030378A (en) | Japonica rice canopy chlorophyll content inverse model approach based on PSO-ELM | |
CN111597756A (en) | Water quality parameter inversion method based on multispectral data of unmanned aerial vehicle | |
CN112712071A (en) | Crop growth detection method and related equipment thereof | |
CN109324000A (en) | Kuerle delicious pear soluble solids content prediction technique based on CARS-MIV-SVR | |
US20200242749A1 (en) | System and method for processing images of agricultural fields for remote phenotype measurement | |
CN109557030A (en) | A kind of water quality element inversion method based on unmanned aerial vehicle remote sensing | |
CN109738881A (en) | Scaling method, device and the readable storage medium storing program for executing of flight time depth mould group | |
CN117173580B (en) | Water quality parameter acquisition method and device, image processing method and medium | |
CN108596123A (en) | A kind of fruit hardness detection method and device based on hyperspectral analysis | |
CN115496891A (en) | Wheat lodging degree grading method and device | |
CN115494007A (en) | Random forest based high-precision rapid detection method and device for soil organic matters | |
CN107677646A (en) | A kind of improvement DBPSO water quality parameter monitoring method and device | |
CN106023268A (en) | Color image graying method based on two-step parameter subspace optimization | |
Liu et al. | Mapping foliar photosynthetic capacity in sub-tropical and tropical forests with UAS-based imaging spectroscopy: Scaling from leaf to canopy | |
Feldmann et al. | Cost‐effective, high‐throughput phenotyping system for 3D reconstruction of fruit form |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171201 |
|
RJ01 | Rejection of invention patent application after publication |