Surface Water Quality Monitoring Method Based on High Spatial Resolution Satellite

Abstract

A surface water quality monitoring method based on a high spatial resolution satellite includes: step 1. building standard surface water quality pools by mixing natural water bodies with clean water in different proportions to obtain surface water quality data; step 2. obtaining high spatial resolution remote sensing images for processing; step 3. identifying remote sensing bands with higher correlation through correlation analysis by presetting ratio values of remote sensor data bands and the standard surface water quality pools; step 4. building a water quality parameter retrieval model, and comparing and retrieving the remote sensing bands with higher correlation and the surface water quality data to obtain water quality data; and step 5. identifying abnormal points of the water quality based on the water quality data and water quality data threshold of the surface water.

Claims

1. A surface water quality monitoring method based on a high spatial resolution satellite, comprising the following steps: step 1: building standard surface water quality pools by mixing natural water bodies with clean water in different proportions to obtain surface water quality data; and the standard surface water quality pools comprise: a standard pool A for containing clean water, a standard pool B for containing natural water, standard pools C, D and E for containing a mixture of clean water and natural water, and the clean water therein accounts for n.sub.1, n.sub.2 and n.sub.3 of the total mixture, respectively, with n.sub.1, n.sub.2 and n.sub.3 different from one another; and a water quality monitoring device is arranged to obtain water quality data of the natural water in the standard pool B; step 2: obtaining high spatial resolution remote sensing images of lakes, reservoirs and rivers, and preprocessing and cropping the high spatial resolution remote sensing images; step 3: identifying remote sensing bands with higher correlation through correlation analysis by presetting ratio values of remote sensor data bands obtained in the step 2 and the standard surface water quality pools in the step 1; step 4: building a water quality parameter retrieval model, and comparing and retrieving the remote sensing bands with higher correlation obtained in the step 3 and the surface water quality data obtained in the step 1 to obtain water quality data of an entire lake and river; step 5: identifying abnormal points of the water quality based on the water quality data of the entire lake and river obtained in the step 4 and a preset water quality data threshold, and identifying areas with excessive pollutant concentrations.

2. The surface water quality monitoring method based on the high spatial resolution satellite according to claim 1, wherein in the step 1, each of the standard pools A, B, C, D, and E is provided with a valve; a revisit period of a satellite is taken as a period for updating the water in each of the standard pools A, B, C, D, and E to remix and update the water bodies therein, wherein the satellite acquires the high spatial resolution remote sensing images.

3. The surface water quality monitoring method based on the high spatial resolution satellite according to claim 1, wherein in the step 1, a length and a width of each of the standard pools A, B, C, D, and E should be greater than or equal to three times resolution of the high spatial resolution remote sensing images.

4. The surface water quality monitoring method based on the high spatial resolution satellite according to claim 1, wherein in the step 1, there are more than three standard pools for containing the mixed water, and at least three of these standard pools used for containing the mixed water have different proportions of the clean water.

5. The surface water quality monitoring method based on the high spatial resolution satellite according to claim 1, wherein in the step 1, n.sub.1=?, n.sub.2=? and n.sub.3=?.

6. The surface water quality monitoring method based on the high spatial resolution satellite according to claim 1, wherein in the step 2, the obtained resolution of the high spatial resolution remote sensing images obtained should not exceed 1 m.

7. The surface water quality monitoring method based on the high spatial resolution satellite according to claim 1, wherein in the step 3, concentration ratios of pollutants in each of the standard pools A, B, C, D, and E are allowed to be determined according to a ratio of the natural water bodies to the clean water therein, remote sensing bands having better adaptability and higher correlation with the concentration ratios of pollutants are allowed to be determined through the correlation analysis, and at least top three bands are allowed to be selected, or those bands with correlation coefficients exceeding a predetermined value are allowed to be selected.

8. The surface water quality monitoring method based on the high spatial resolution satellite according to claim 1, wherein in the step 3, the correlation analysis is performed by presetting ratio values of remote sensor data bands and the standard surface water quality pools, X.sub.1 represents a remote sensing band matrix, and Y.sub.1 represents a preset ratio value matrix of pollutant concentrations; and through the correlation analysis thereof, a correlation coefficient matrix R.sub.1 of each band and the preset ratio value matrix of pollutant concentrations is obtained: $X_1=\left[\begin{array}{cccccc}{X}_A^I& X_A^{\mathrm{II}}& X_A^{\mathrm{III}}& X_A^{\mathrm{IV}}& X_A^V& X_A^{\mathrm{VI}}\\ X_B^I& X_B^{\mathrm{II}}& X_B^{\mathrm{III}}& X_B^{\mathrm{IV}}& X_B^V& X_B^{\mathrm{VI}}\\ X_C^I& X_C^{\mathrm{II}}& X_C^{\mathrm{III}}& X_C^{\mathrm{IV}}& X_C^V& X_C^{\mathrm{VI}}\\ X_D^I& X_D^{\mathrm{II}}& X_D^{\mathrm{III}}& X_D^{\mathrm{IV}}& X_D^V& X_D^{\mathrm{VI}}\\ X_E^I& X_E^{\mathrm{II}}& X_E^{\mathrm{III}}& X_E^{\mathrm{IV}}& X_E^V& X_E^{\mathrm{VI}}\end{array}\right]$ in the equation, subscripts A-E represent numbering of the standard pools, and superscripts I, II, III, IV, V and VI represent band categories, wherein X.sub.A denotes a value of a band I of the standard pool A; and $Y_1=\left[\begin{array}{c}0\\ 1\\ 1/n_1\\ 1/n_2\\ 1/n_3\end{array}\right]$ $R_1=[\begin{array}{cccccc}{r}^I& r^{\mathrm{II}}& r^{\mathrm{III}}& r^{\mathrm{IV}}& r^V& r^{\mathrm{VI}}]\end{array}$ finding and selecting a plurality of bands with values ranking top in R.sub.1 or selecting those bands with correlation coefficients exceeding a predetermined value as remote sensing bands with higher correlation.

9. The surface water quality monitoring method based on the high spatial resolution satellite according to claim 1, wherein in the step 4, data of monitoring points, that is, a spectral reflectance correlation model, is retrieved and obtained by optimizing the model according to data of the standard surface water quality pools and data of spectral reflectance of corresponding monitoring stations, and the like, and an accuracy of the model is tested with a multiple correlation coefficient R.sup.2, the optimal retrieval band and band combination of each water quality index are determined by performing statistical analysis of a correlation between each band, band combination and corresponding water quality index data, and a retrieval regression model of each pollutant concentration is built.

10. The surface water quality monitoring method based on the high spatial resolution satellite according to claim 1, wherein in the step 5, an overall water quality and water quality distribution are allowed to be identified based on the water quality data of the entire lake and river obtained in the step 4 and the preset water quality data threshold, and notification and early warning are allowed to be accordingly given.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0032] FIG. 1 is a flowchart of a surface water quality monitoring method based on a high spatial resolution satellite in an embodiment of the present disclosure.

[0033] FIG. 2 is a schematic diagram of locations of standard surface water quality pools in an embodiment of the present disclosure.

[0034] FIG. 3 is a satellite image simulation diagram in an embodiment of the present disclosure.

DETAILED DESCRIPTIONS OF THE EMBODIMENTS

[0035] A surface water quality monitoring method based on a high spatial resolution satellite provided in the present disclosure will be described in details in conjunction with the accompanying drawings.

Embodiment

[0036] As shown in FIG. 1, the surface water quality monitoring method based on a high spatial resolution satellite provided in this embodiment includes the following steps:

[0037] S1. building standard surface water quality pools by mixing natural water bodies with clean water in different proportions to obtain surface water quality data; and

[0038] the building standard surface water quality pools further includes two steps: first, building the pools and then filling the pools with water bodies of different concentration gradients, and a length and a width of each pool should be no less than three times resolution of the remote sensing images. In this embodiment, the length?width?height of each of the standard pools is 3 m?3 m?1 m.

[0039] As shown in FIG. 2, five standard pools (A-E) are built on one side of a river/lake 1, and the standard pools B-E are all connected to the river/lake 1 through a natural water body connecting pipeline 2. Specifically, the standard pool A is filled with clean water, the standard pool B is filled with natural water, 1/n.sub.1 of the standard pool C is filled with natural water and the rest thereof is filled with clean water, 1/n.sub.2 of the standard pool D is filled with natural water and the rest thereof is filled with clean water, and 1/n.sub.3 of the standard pool E is filled with natural water and the rest thereof is filled with clean water. In order to ensure that the natural water in the standard pools are consistent with that in the natural water body 1, a valve is installed at an inlet of each of the standard pools, the natural water is pumped into each of the standard pools, the water bodies in each of the pools are remixed and updated by taking a satellite revisit period as a period, and a simulated satellite image is shown in FIG. 3. A set of water quality monitoring device is arranged in the standard pool B. In this embodiment, n.sub.1, n.sub.2 and n.sub.3 are ?, ? and ?, respectively.

[0040] S2. Obtaining high spatial resolution remote sensing images of lakes, reservoirs and rivers, and preprocessing and cropping the acquired remote sensing images; and

[0041] satellite remote sensing adopts high spatial resolution remote sensing, and the accuracy of water quality monitoring can be improved when ground resolution of the high spatial resolution remote sensing is less than 1 m.

[0042] Image preprocessing involves radiometric calibration, atmospheric correction and geometric correction. Specifically, radiometric calibration is a process of converting the brightness grayscale values of the images into absolute radiance values. Atmospheric correction is a process of eliminating radiation errors caused by atmospheric effects and retrieving the true surface reflectance of a ground object. Geometric correction is a process of correcting and eliminating distortions in geometric positions, shapes, sizes, orientations, and the like of various ground objects in original images that are inconsistent with the expression requirements in a reference system, and the distortions result from deformation of photographic materials, distortion of objective lens, atmospheric refraction, Earth curvature, Earth rotation, topographic relief, and the like.

[0043] S3. Identifying remote sensing bands with higher correlation through correlation analysis by presetting ratio values of remote sensor data bands obtained in the S2 and the standard surface water quality pools in the S1; and

[0044] performing the correlation analysis by presetting ratio values of remote sensor data bands and the standard surface water quality pools, where X.sub.1 represents a remote sensing band matrix, and Y.sub.! represents a preset ratio value matrix of pollutant concentrations. Through the correlation analysis thereof, a correlation coefficient matrix R.sub.1 of each band and the preset ratio value matrix of pollutant concentrations is obtained, and three bands with the largest correlation coefficients thereof are selected.

[00003]$X_1=\left[\begin{array}{cccccc}{X}_A^I& X_A^{\mathrm{II}}& X_A^{\mathrm{III}}& X_A^{\mathrm{IV}}& X_A^V& X_A^{\mathrm{VI}}\\ X_B^I& X_B^{\mathrm{II}}& X_B^{\mathrm{III}}& X_B^{\mathrm{IV}}& X_B^V& X_B^{\mathrm{VI}}\\ X_C^I& X_C^{\mathrm{II}}& X_C^{\mathrm{III}}& X_C^{\mathrm{IV}}& X_C^V& X_C^{\mathrm{VI}}\\ X_D^I& X_D^{\mathrm{II}}& X_D^{\mathrm{III}}& X_D^{\mathrm{IV}}& X_D^V& X_D^{\mathrm{VI}}\\ X_E^I& X_E^{\mathrm{II}}& X_E^{\mathrm{III}}& X_E^{\mathrm{IV}}& X_E^V& X_E^{\mathrm{VI}}\end{array}\right]$

[0045] in the equation, subscripts A-E represent numbering of the standard pools, and superscripts I, II, III, IV, V and VI represent band categories. For example, X denotes a value of a band I of the standard pool A(3).

[00004]$Y_1=\left[\begin{array}{c}0\\ 1\\ 1/n_1\\ 1/n_2\\ 1/n_3\end{array}\right]$$R_1=[\begin{array}{cccccc}{r}^I& r^{\mathrm{II}}& r^{\mathrm{III}}& r^{\mathrm{IV}}& r^V& r^{\mathrm{VI}}]\end{array}$

[0046] Specific calculations are shown below:

[00005]$X_1=\left[\begin{array}{cccccc}360& 585& 573& 885& 546& 375\\ 498& 632& 435& 424& 283& 184\\ 190& 369& 306& 433& 310& 221\\ 355& 543& 377& 555& 298& 190\\ 443& 714& 446& 431& 175& 109\end{array}\right]$

[0047] It can be seen from the S1 that n.sub.1=?, n.sub.2=? and n.sub.3=? in this embodiment, therefore:

[00006]$Y_1=\left[\begin{array}{c}0\\ 1\\ 1/2\\ 2/3\\ 1/3\end{array}\right]$$R_1=[\begin{array}{cccccc}0.312& 0.026& 0.548& 0.713& 0.547& 0.561]\end{array}$

[0048] Bands III, IV and VI are accordingly selected.

[0049] S4. building a water quality parameter retrieval model, and comparing and retrieving the remote sensing bands with higher correlation obtained in the S and the surface water quality data obtained in the S1 to obtain water quality data of an entire lake and river; and

[0050] data of monitoring points, that is, a spectral reflectance correlation model, is retrieved and obtained by optimizing the model according to data of the standard surface water quality pools and data of spectral reflectance of the corresponding monitoring stations by mean of a partial least squares method, and the like, and the accuracy of the model is tested with a multiple correlation coefficient R.sup.2, that is, the optimal retrieval band and band combination of each water quality index are determined by performing statistical analysis of the correlation between each band, band combination and corresponding water quality index data, and a retrieval regression model of each pollutant concentration is then built.

[0051] The partial least squares method has the following steps:

[0052] 1. establishing an independent variable set (taking the largest correlation coefficient in the bands III, IV, VI as an example):

[00007]$X=\left[\begin{array}{ccc}{X}_A^{\mathrm{III}}& X_A^{\mathrm{IV}}& X_A^{\mathrm{VI}}\\ X_B^{\mathrm{III}}& X_B^{\mathrm{IV}}& X_B^{\mathrm{VI}}\\ X_C^{\mathrm{III}}& X_C^{\mathrm{IV}}& X_C^{\mathrm{VI}}\\ X_D^{\mathrm{III}}& X_D^{\mathrm{IV}}& X_D^{\mathrm{VI}}\\ X_E^{\mathrm{III}}& X_E^{\mathrm{IV}}& X_E^{\mathrm{VI}}\end{array}\right]$

[0053] in the equation, the subscript A represents numbering of the standard pool, and the superscripts I, II, III, IV, V and VI represent band categories. For example, X.sub.A.sup.III denotes a value of a band IIII of the standard pool A(3).

[0054] 2. Establishing a dependent variable set

[0055] water quality tests need to be performed on the standard pool B at regular intervals to obtain:

[00008]$Y=\left[\begin{array}{ccccc}0& C_{\mathrm{COD}}& 1/2C_{\mathrm{COD}}& 2/3C_{\mathrm{COD}}& 1/3C_{\mathrm{COD}}\\ 0& C_{\mathrm{TP}}& 1/2C_{\mathrm{TP}}& 2/3C_{\mathrm{TP}}& 1/3C_{\mathrm{TP}}\\ 0& C_{\mathrm{TN}}& 1/2C_{\mathrm{TN}}& 2/3C_{\mathrm{TN}}& 1/3C_{\mathrm{TN}}\end{array}\right]$

[0056] in the equation, C.sub.COD, CTP and C.sub.TN represent measured values of COD, TP and TN in the natural water.

[0057] 3. Standardizing X and Y as E.sub.0 and F.sub.0, respectively, and fining a rank h of X.

[0058] 4. Finding a unit eigenvector ?1 corresponding to the largest eigenvalue of the matrix E.sub.0.sup.TF.sub.0F.sub.0.sup.TE.sub.0, with a corresponding component t.sub.1:

[00009]$t_1=E_0{?}_1$$E_1=E_0-t_1{p}_1^T$$p_1=\frac{E_0^T{t}_1}{{.Math.t_1.Math.}^2}$

[0059] 5. Finding a unit eigenvector ?.sub.2 corresponding to the largest eigenvalue of the matrix E.sub.1.sup.TF.sub.0F.sub.0.sup.TE.sub.1, with a corresponding component t.sub.2:

[00010]$t_2=E_1{?}_2$$E_2=E_1-t_2{p}_2^T$$p_2=\frac{E_2^T{t}_2}{{.Math.t_2.Math.}^2}$

[0060] 6. Repeating the steps to find a unit eigenvector ?.sub.h corresponding to the largest eigenvalue of the matrix E.sub.h-1.sup.TF.sub.0F.sub.0.sup.TE.sub.h-1, with a corresponding component t.sub.h.

[0061] 7. Obtaining an ordinary least squares regression equation of F.sub.0 on t.sub.1, . . . , t.sub.h:

[00011]$F_0=t_1{r}_1^T+t_2{r}_2^T+.Math.+t_h{r}_h^T+F_m$$r_i=\frac{F_0^T{t}_i}{{.Math.t_i.Math.}^2},i=1,2,.Math.,h$

[0062] Specific calculations are shown below:

[0063] it is measured that C.sub.TN, is 1.8 mg/L.

[00012]$X_1=\left[\begin{array}{ccc}573& 885& 375\\ 435& 424& 184\\ 306& 433& 221\\ 377& 555& 190\\ 446& 431& 109\end{array}\right]$$Y=\left[\begin{array}{ccccc}0& 1.8& 0.9& 1.2& 0.6\end{array}\right]$$Y=2.2+0.000486*X^{\mathrm{IH}}-0.00411*X^{\mathrm{IV}}+0.00342X^{\mathrm{VI}}$

[0064] in the equation, Y represents a total nitrogen value, X.sup.III represents a band value at a calculation point III, X.sup.IV represents a band value at a calculation point IV, and X.sup.VI represents a band value at a calculation point VI.

[0065] A pollutant concentration value of the entire lake and river can be calculated according to a comparison expression between the obtained pollutant concentration values and the bands.

[0066] S5. Investigating and reporting abnormal points based on the water quality data of the entire lake and river obtained in the step 4 and a preset water quality data threshold.

[0067] In this embodiment, the water quality data threshold is preset as follows: the pollutant concentration reaches up to five times the surrounding pollutant concentration, and the pollutant concentration meets the standards for Class V water of the surface water.

[0068] Therefore, an abnormal point refers to a point where the pollutant concentration exceeds five times the surrounding pollutant concentration and goes against the standards for Class V water of the surface water. Any abnormal point, once being identified, should be reported for further investigating the causes, that is, the following conditions are satisfied: y.sub.i,k>5?y.sub.i,j & y.sub.i,k>x.sub.i, where y.sub.i,k represents a concentration of a pollutant i at a point K (i=1, 2, 3, denoting COD, total phosphorus and total nitrogen, respectively), y.sub.i,j represents a concentration of the pollutant i at a point J (the point J meets the conditions: D<3?d, where D represents a distance between the point J and the point K, and a point d is satellite spatial resolution), and x.sub.i represents a standard concentration of Class V water of the surface water corresponding to the pollutants (x.sub.1=40 mg/L, x.sub.2=0.4 mg/L and x.sub.3=2.0 mg/L).

[0069] The above embodiment is only illustration of the technical solution of the present disclosure. The surface water quality monitoring method based on a high spatial resolution satellite provided in the present disclosure is not limited to the content described in the above embodiments, but is subject to the scope defined by the claims. Any modifications, additions or equivalent substitutions made by those skilled in the art based on the embodiments fall within the scope of protection of the claims of the present disclosure.