Water quality parameter hyperspectral analysis method and system based on machine learning model

Abstract

The invention relates to the technical field of hyperspectral analysis, and particularly discloses a water quality parameter hyperspectral analysis method and system based on a machine learning model, and the method comprises the following steps: S1, selecting a plurality of water quality parameter samples, and according to the spectrum curve of each water quality parameter sample under the spectrum bands with different signal-to-noise ratios, calculating the spectrum curve of each water quality parameter sample; dividing the plurality of water quality parameter samples into sensitive water quality parameter samples and non-sensitive water quality parameter samples; s2, acquiring a water body sample at a pollution source of the water body, and acquiring a concentration and a source spectrum curve of non-sensitive water quality parameters in the water body sample; s3, hyperspectral analysis is carried out on the collection points of the water body, a hyperspectral image is obtained, and non-sensitive pixel points are selected according to the spectral curve of each pixel point in the hyperspectral image; and S4, extracting all the non-sensitive curves, comparing the non-sensitive curves with the source spectrum curve, finding out a target curve with the highest similarity, and further predicting the predicted concentration of the non-sensitive water quality parameter at the acquisition point.

Description

Technical Field

[0001] The present invention relates to the technical field of hyperspectral analysis, and in particular to a method and system for hyperspectral analysis of water quality parameters based on a machine learning model. Background Art

[0002] Hyperspectral analysis of water quality parameters is a method that uses hyperspectral imaging technology to acquire and analyze the optical properties of water bodies and then evaluate water quality conditions; it includes the following steps: data acquisition, data preprocessing, model building and water quality parameter inversion.

[0003] In the existing technology, different water quality parameters have different sensitivities to hyperspectral analysis. Specifically, some water quality parameters have a more obvious response in hyperspectral data and can be easily accurately identified and quantified through spectral features; while other water quality parameters may be difficult to be discovered in time or accurately measured in hyperspectral analysis because their spectral features are not significant enough or are easily interfered by other factors; for water quality parameters with lower sensitivity, if their content or changes in the water body do not reach the detection threshold of hyperspectral analysis, then these parameters may be missed or misjudged, thus failing to provide comprehensive and accurate information for water quality assessment. Summary of the invention

[0004] The purpose of the present invention is to provide a water quality parameter hyperspectral analysis method and system based on a machine learning model to solve the above technical problems.

[0005] The purpose of the present invention can be achieved by the following technical solutions:

[0006] A water quality parameter hyperspectral analysis method based on a machine learning model comprises the following steps:

[0007] Step S1: Select several water quality parameter samples to obtain a water quality parameter sample solution; set a standard signal-to-noise ratio Sn0, and set several signal-to-noise ratio gradient values ​​according to the standard signal-to-noise ratio, perform hyperspectral analysis on the water quality parameter sample solution with spectral bands of different signal-to-noise ratio gradient values, and obtain a spectral curve of the water quality parameter sample solution, wherein the spectral curve is the reflectivity at different wavelengths;

[0008] Obtain the reflectance change rate of the water quality parameter sample according to each signal-to-noise ratio gradient value and the spectral curve corresponding to the water quality parameter sample solution of each signal-to-noise ratio gradient value; record the water quality parameter sample whose reflectance change rate is lower than a preset reflectance change threshold as a non-sensitive water quality parameter;

[0009] Step S2: The location of the pollution source of the water body is recorded as the pollution source point; a water sample is collected at the pollution source point to obtain the concentration c0 of the non-sensitive water quality parameter; a hyperspectral analysis is performed on the water sample with a spectral band of a standard signal-to-noise ratio to obtain a hyperspectral image, and a spectral curve within a pixel point where the non-sensitive water quality parameter is located is obtained in the hyperspectral image, which is recorded as a source spectral curve;

[0010] Step S3: any other positions of the water body are recorded as acquisition points. When performing hyperspectral analysis at the acquisition points, hyperspectral images obtained under spectral bands with different signal-to-noise ratio gradient values ​​are obtained and recorded as current hyperspectral images. The spectral curves in each pixel point of the current hyperspectral image are recorded as current spectral curves.

[0011] In the current hyperspectral image, all current spectral curves of the same pixel in spectral bands with different signal-to-noise ratio gradient values ​​are obtained to obtain the difference value:

[0012] Where n is the total number of signal-to-noise ratio gradient values, f i (λ) represents the current spectral curve under the spectral band of the i-th signal-to-noise ratio gradient value, λ is the wavelength, and [λ0, λ1] represents the wavelength range of the spectral band; the pixel points whose difference values ​​are less than the preset difference threshold are selected and recorded as non-sensitive pixel points;

[0013] Step S4: Obtain all current spectral curves of the non-sensitive pixel points, recorded as non-sensitive curves, and select the non-sensitive curve with the highest similarity to the source spectral curve, recorded as the target curve; obtain the signal-to-noise ratio gradient value corresponding to the target curve, recorded as Sn min According to the signal-to-noise ratio gradient value and the standard signal-to-noise ratio, the predicted concentration of the non-sensitive water quality parameter at the collection point is obtained: Pc = c0 + μ (Sn min -Sn0), where μ is a preset correction parameter and μ>0, and the unit is mol / (L*dB).

[0014] As a further solution of the present invention: the process of setting the signal-to-noise ratio gradient value includes:

[0015] A signal-to-noise ratio interval threshold is set, the standard signal-to-noise ratio is used as a starting value, and the starting value is increased or decreased according to the signal-to-noise ratio interval threshold to obtain a plurality of signal-to-noise ratio gradient values.

[0016] As a further solution of the present invention: the process of obtaining the reflection change rate includes:

[0017] Select several reference points on each spectral curve, and obtain the mean reflectivity according to the reflectivity at each reference point on the spectral curve. Where k represents the kth spectral curve, P xrepresents the reflectivity at the xth reference point, and m is the total number of reference points;

[0018] A coordinate system is established with the signal-to-noise ratio as the horizontal coordinate and the reflectivity as the vertical coordinate; a standard line y=Mr is obtained in the coordinate system, where y represents the vertical coordinate of the coordinate system; all spectral curves are placed in the coordinate system, and the starting points of all spectral curves coincide with the origin; in the coordinate system, the difference between each spectral curve and the standard line is obtained The reflectance change rate is Sn i represents the i-th signal-to-noise ratio gradient value, D i Represents the difference in spectral curves corresponding to the i-th signal-to-noise ratio gradient value.

[0019] As a further solution of the present invention: the water quality parameter samples whose reflectance change rate is greater than or equal to a preset reflectance change threshold are recorded as sensitive water quality parameters.

[0020] As a further solution of the present invention: the process of obtaining the spectral curve within the pixel point where the non-sensitive water quality parameter is located in the hyperspectral image includes:

[0021] All sensitive water quality parameters in the water sample are obtained, and spectral curves of the sensitive water quality parameters are obtained, which are recorded as sensitive spectral curves; pixel points where all sensitive spectral curves are located are selected in the hyperspectral image, which are recorded as sensitive pixel points, and all sensitive pixel points are eliminated, and the spectral curves in all remaining pixel points in the hyperspectral image are recorded as source spectral curves.

[0022] As a further solution of the present invention: pixels whose difference values ​​are greater than or equal to a preset difference threshold are recorded as sensitive pixels.

[0023] As a further solution of the present invention: the process of selecting the insensitive curve with the highest similarity to the source spectrum curve includes:

[0024] Select several comparison points at equal intervals on the source spectrum curve and the non-sensitive curve, and then obtain the similarity between the source spectrum curve and the non-sensitive curve. Where G(t) represents the tth comparison point on the source spectrum curve, T(t) represents the tth comparison point on the non-sensitive curve, and J is the total number of comparison points.

[0025] As a further solution of the present invention: a water quality parameter hyperspectral analysis system based on a machine learning model, comprising:

[0026] Sensitive division module: select several water quality parameter samples to obtain water quality parameter sample solutions; set a standard signal-to-noise ratio Sn0, and set several signal-to-noise ratio gradient values ​​according to the standard signal-to-noise ratio, perform hyperspectral analysis on the water quality parameter sample solutions with spectral bands of different signal-to-noise ratio gradient values, and obtain spectral curves of the water quality parameter sample solutions, wherein the spectral curves are reflectances at different wavelengths;

[0027] Obtain the reflectance change rate of the water quality parameter sample according to each signal-to-noise ratio gradient value and the spectral curve corresponding to the water quality parameter sample solution of each signal-to-noise ratio gradient value; record the water quality parameter sample whose reflectance change rate is lower than a preset reflectance change threshold as a non-sensitive water quality parameter;

[0028] Pollution source analysis module: the location of the pollution source of the water body is recorded as the pollution source point; water samples are collected at the pollution source point to obtain the concentration c0 of the non-sensitive water quality parameter; the water samples are subjected to hyperspectral analysis with a spectral band of a standard signal-to-noise ratio to obtain a hyperspectral image, and the spectral curve within the pixel point where the non-sensitive water quality parameter is located is obtained in the hyperspectral image and recorded as the source spectral curve;

[0029] Collection point analysis module: any other position of the water body is recorded as a collection point. When performing hyperspectral analysis at the collection point, a hyperspectral image obtained under a spectral band with different signal-to-noise ratio gradient values ​​is obtained and recorded as a current hyperspectral image. The spectral curve within each pixel point of the current hyperspectral image is recorded as the current spectral curve.

[0030] In the current hyperspectral image, all current spectral curves of the same pixel in spectral bands with different signal-to-noise ratio gradient values ​​are obtained to obtain the difference value:

[0031] Where n is the total number of signal-to-noise ratio gradient values, f i (λ) represents the current spectral curve under the spectral band of the i-th signal-to-noise ratio gradient value, λ is the wavelength, and [λ0, λ1] represents the wavelength range of the spectral band; the pixel points whose difference values ​​are less than the preset difference threshold are selected and recorded as non-sensitive pixel points;

[0032] Concentration prediction module: obtain all current spectral curves of the non-sensitive pixel points, record them as non-sensitive curves, and select the non-sensitive curve with the highest similarity to the source spectral curve, record them as target curves; obtain the signal-to-noise ratio gradient value corresponding to the target curve, record them as Sn min According to the signal-to-noise ratio gradient value and the standard signal-to-noise ratio, the predicted concentration of the non-sensitive water quality parameter at the collection point is obtained: Pc = c0 + μ (Sn min -Sn0), where μ is a preset correction parameter and μ>0, and the unit is mol / (L*dB).

[0033] Beneficial effects of the present invention:

[0034] The present invention uses a machine learning model to perform hyperspectral analysis on water quality parameters, which can identify and distinguish non-sensitive water quality parameters and improve the accuracy of the analysis; using spectral bands under different signal-to-noise ratio gradient values, the reflectance change information of water samples can be obtained more accurately, thereby predicting the concentration of non-sensitive water quality parameters; water samples are collected at the pollution source, and the concentration of non-sensitive water quality parameters and the corresponding hyperspectral images are obtained. Then, hyperspectral analysis is performed at the collection point to obtain hyperspectral images under different signal-to-noise ratio gradient values, and non-sensitive curves are extracted. Finally, by comparing the non-sensitive curves of the two locations and combining the preset correction parameters, the concentration of non-sensitive water quality parameters at the collection point can be predicted. This method not only improves the inversion efficiency of water quality parameters, but also has strong adaptability and generalization ability, and is suitable for different pollution sources and collection points. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The present invention will be further described below in conjunction with the accompanying drawings.

[0036] Figure 1 It is a flow chart of a method for hyperspectral analysis of water quality parameters based on a machine learning model of the present invention. DETAILED DESCRIPTION

[0037] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0038] See also Figure 1 As shown, the present invention is a water quality parameter hyperspectral analysis method based on a machine learning model, comprising the following steps:

[0039] Step S1: Select several water quality parameter samples to obtain a water quality parameter sample solution; set a standard signal-to-noise ratio Sn0, and set several signal-to-noise ratio gradient values ​​according to the standard signal-to-noise ratio, perform hyperspectral analysis on the water quality parameter sample solution with spectral bands of different signal-to-noise ratio gradient values, and obtain a spectral curve of the water quality parameter sample solution, wherein the spectral curve is the reflectivity at different wavelengths;

[0040] Obtain the reflectance change rate of the water quality parameter sample according to each signal-to-noise ratio gradient value and the spectral curve corresponding to the water quality parameter sample solution of each signal-to-noise ratio gradient value; record the water quality parameter sample whose reflectance change rate is lower than a preset reflectance change threshold as a non-sensitive water quality parameter;

[0041] It should be noted that the water quality parameter sample includes physical or chemical components such as dissolved oxygen, metal element content, and ammonia nitrogen; the water quality parameter sample solution is a solution formed by dissolving the selected water quality parameter sample in an appropriate solvent; the standard signal-to-noise ratio is a determined benchmark signal-to-noise ratio value, which is used as the standard for analysis;

[0042] It can be understood that the water quality parameter sample solution is subjected to hyperspectral analysis using spectral bands with different signal-to-noise ratio gradient values, and the reflectance of each sample at different wavelengths is obtained to generate a spectral curve; the reflectance change rate of each water quality parameter sample is calculated according to each signal-to-noise ratio gradient value and the corresponding spectral curve; the reflectance change rate reflects the change in the spectral response of the sample under different signal-to-noise ratio conditions; the water quality parameter samples with a reflectance change rate lower than a preset reflectance change threshold are marked as non-sensitive water quality parameters, which means that these parameters change less under different signal-to-noise ratio conditions and are not easily detected;

[0043] It is worth noting that by calculating the reflectivity changes under different signal-to-noise ratio conditions, the sensitivity of water quality parameters to signal-to-noise ratio changes can be evaluated; a low reflectivity change rate means that the parameter does not change much under different conditions, that is, it is non-sensitive;

[0044] As a preferred embodiment of the present invention, the process of setting the signal-to-noise ratio gradient value includes:

[0045] Setting a signal-to-noise ratio interval threshold, taking the standard signal-to-noise ratio as a starting value, and increasing or decreasing the starting value according to the signal-to-noise ratio interval threshold to obtain a plurality of signal-to-noise ratio gradient values;

[0046] It can be understood that the signal-to-noise ratio interval threshold is used to control the interval between the signal-to-noise ratio gradient values, ensuring that each gradient value has an appropriate distance so that changes under different conditions can be captured in subsequent hyperspectral analysis; the standard signal-to-noise ratio is used as a starting point to provide a benchmark value, and all gradient values ​​are obtained by increasing or decreasing based on this benchmark value; multiple signal-to-noise ratio gradient values ​​are generated by increasing or decreasing, and these gradient values ​​will be used in different experimental conditions or analysis steps to ensure the diversity and comprehensiveness of the data;

[0047] As a preferred embodiment of the present invention, the process of obtaining the reflection change rate includes:

[0048] Select several reference points on each spectral curve, and obtain the mean reflectivity according to the reflectivity at each reference point on the spectral curve. Where k represents the kth spectral curve, P x represents the reflectivity at the xth reference point, and m is the total number of reference points;

[0049] A coordinate system is established with the signal-to-noise ratio as the horizontal coordinate and the reflectivity as the vertical coordinate; a standard line y=Mr is obtained in the coordinate system, where y represents the vertical coordinate of the coordinate system; all spectral curves are placed in the coordinate system, and the starting points of all spectral curves coincide with the origin; in the coordinate system, the difference between each spectral curve and the standard line is obtained The reflectance change rate is Sn i represents the i-th signal-to-noise ratio gradient value, D i Represents the difference of the spectral curve corresponding to the i-th signal-to-noise ratio gradient value;

[0050] It can be understood that by selecting multiple reference points and calculating their average reflectance, a reference value Mr is obtained for subsequent comparison; establishing a coordinate system with signal-to-noise ratio as the horizontal axis and reflectance as the vertical axis helps to intuitively display the changes in the spectral curve; by calculating the difference D between each spectral curve and the standard line, the degree of deviation between each curve and the reference value can be quantified; the final reflectance change rate R is calculated by comparing the difference changes under different signal-to-noise ratio gradients, reflecting the sensitivity of water quality parameter samples under different conditions;

[0051] As a preferred embodiment of the present invention, the water quality parameter sample whose reflectance change rate is greater than or equal to a preset reflectance change threshold is recorded as a sensitive water quality parameter;

[0052] Step S2: The location of the pollution source of the water body is recorded as the pollution source point; a water sample is collected at the pollution source point to obtain the concentration c0 of the non-sensitive water quality parameter; a hyperspectral analysis is performed on the water sample with a spectral band of a standard signal-to-noise ratio to obtain a hyperspectral image, and a spectral curve within a pixel point where the non-sensitive water quality parameter is located is obtained in the hyperspectral image, which is recorded as a source spectral curve;

[0053] It can be understood that a water sample is subjected to hyperspectral analysis using a spectral band with a standard signal-to-noise ratio to generate a hyperspectral image; the hyperspectral image contains rich spectral information and can be used for further analysis; spectral curves of pixel points where non-sensitive water quality parameters are located are extracted from the hyperspectral image, and these curves reflect the reflectivity of the non-sensitive water quality parameters at different wavelengths;

[0054] As a preferred embodiment of the present invention, the process of obtaining the spectral curve within the pixel point where the non-sensitive water quality parameter is located in the hyperspectral image includes:

[0055] Obtain all sensitive water quality parameters in the water sample, obtain spectral curves of the sensitive water quality parameters, and record them as sensitive spectral curves; select pixel points where all sensitive spectral curves are located in the hyperspectral image, record them as sensitive pixel points, and remove all sensitive pixel points, and record the spectral curves in all remaining pixel points in the hyperspectral image as source spectral curves;

[0056] It can be understood that after removing sensitive pixels, the spectral curves of the remaining pixels mainly reflect the characteristics of non-sensitive water quality parameters;

[0057] Step S3: any other positions of the water body are recorded as acquisition points. When performing hyperspectral analysis at the acquisition points, hyperspectral images obtained under spectral bands with different signal-to-noise ratio gradient values ​​are obtained and recorded as current hyperspectral images. The spectral curves in each pixel point of the current hyperspectral image are recorded as current spectral curves.

[0058] In the current hyperspectral image, all current spectral curves of the same pixel in spectral bands with different signal-to-noise ratio gradient values ​​are obtained to obtain the difference value:

[0059] Where n is the total number of signal-to-noise ratio gradient values, f i (λ) represents the current spectral curve under the spectral band of the i-th signal-to-noise ratio gradient value, λ is the wavelength, and [λ0, λ1] represents the wavelength range of the spectral band; the pixel points whose difference values ​​are less than the preset difference threshold are selected and recorded as non-sensitive pixel points;

[0060] It can be understood that sampling and analysis are carried out at different locations of the water body to ensure the representativeness and comprehensiveness of the data; hyperspectral analysis is carried out through spectral bands with different signal-to-noise ratio gradient values ​​to obtain detailed spectral information; the spectral curve of each pixel is recorded for subsequent difference value calculation; the difference value of the spectral curve of the same pixel under different signal-to-noise ratio gradient values ​​is calculated to quantify its degree of change; pixels with difference values ​​less than the preset threshold have small changes and are considered to be insensitive;

[0061] As a preferred embodiment of the present invention, the pixel points whose difference value is greater than or equal to the preset difference threshold are recorded as sensitive pixel points;

[0062] Step S4: Obtain all current spectral curves of the non-sensitive pixel points, recorded as non-sensitive curves, and select the non-sensitive curve with the highest similarity to the source spectral curve, recorded as the target curve; obtain the signal-to-noise ratio gradient value corresponding to the target curve, recorded as Sn min According to the signal-to-noise ratio gradient value and the standard signal-to-noise ratio, the predicted concentration of the non-sensitive water quality parameter at the collection point is obtained: Pc = c0 + μ (Sn min-Sn0), where μ is a preset correction parameter and μ>0, and the unit is mol / (L*dB);

[0063] It can be understood that the non-sensitive curve is a spectral curve obtained from non-sensitive pixel points, representing the reflectance change under different signal-to-noise ratio gradients; the predicted concentration of the non-sensitive water quality parameter at the collection point is calculated by combining the correction parameter μ with the difference between the known concentration and the signal-to-noise ratio gradient value;

[0064] As a preferred embodiment of the present invention, the process of selecting the insensitive curve having the highest similarity to the source spectrum curve includes:

[0065] Select several comparison points at equal intervals on the source spectrum curve and the non-sensitive curve, and then obtain the similarity between the source spectrum curve and the non-sensitive curve. Where G(t) represents the tth comparison point on the source spectrum curve, T(t) represents the tth comparison point on the non-sensitive curve, and J is the total number of comparison points;

[0066] It can be understood that the similarity between the two curves is quantified by calculating the difference between the comparison points. The smaller the similarity value, the more similar the two curves are.

[0067] A water quality parameter hyperspectral analysis system based on a machine learning model, comprising:

[0068] Sensitive division module: select several water quality parameter samples to obtain water quality parameter sample solutions; set a standard signal-to-noise ratio Sn0, and set several signal-to-noise ratio gradient values ​​according to the standard signal-to-noise ratio, perform hyperspectral analysis on the water quality parameter sample solutions with spectral bands of different signal-to-noise ratio gradient values, and obtain spectral curves of the water quality parameter sample solutions, wherein the spectral curves are reflectances at different wavelengths;

[0069] Obtain the reflectance change rate of the water quality parameter sample according to each signal-to-noise ratio gradient value and the spectral curve corresponding to the water quality parameter sample solution of each signal-to-noise ratio gradient value; record the water quality parameter sample whose reflectance change rate is lower than a preset reflectance change threshold as a non-sensitive water quality parameter;

[0070] Pollution source analysis module: the location of the pollution source of the water body is recorded as the pollution source point; water samples are collected at the pollution source point to obtain the concentration c0 of the non-sensitive water quality parameter; the water samples are subjected to hyperspectral analysis with a spectral band of a standard signal-to-noise ratio to obtain a hyperspectral image, and the spectral curve within the pixel point where the non-sensitive water quality parameter is located is obtained in the hyperspectral image and recorded as the source spectral curve;

[0071] Collection point analysis module: any other position of the water body is recorded as a collection point. When performing hyperspectral analysis at the collection point, a hyperspectral image obtained under a spectral band with different signal-to-noise ratio gradient values ​​is obtained and recorded as a current hyperspectral image. The spectral curve within each pixel point of the current hyperspectral image is recorded as the current spectral curve.

[0072] In the current hyperspectral image, all current spectral curves of the same pixel in spectral bands with different signal-to-noise ratio gradient values ​​are obtained to obtain the difference value:

[0073] Where n is the total number of signal-to-noise ratio gradient values, f i (λ) represents the current spectral curve under the spectral band of the i-th signal-to-noise ratio gradient value, λ is the wavelength, and [λ0, λ1] represents the wavelength range of the spectral band; the pixel points whose difference values ​​are less than the preset difference threshold are selected and recorded as non-sensitive pixel points;

[0074] Concentration prediction module: obtain all current spectral curves of the non-sensitive pixel points, record them as non-sensitive curves, and select the non-sensitive curve with the highest similarity to the source spectral curve, record them as target curves; obtain the signal-to-noise ratio gradient value corresponding to the target curve, record them as Sn min According to the signal-to-noise ratio gradient value and the standard signal-to-noise ratio, the predicted concentration of the non-sensitive water quality parameter at the collection point is obtained: Pc = c0 + μ (Sn min -Sn0), where μ is a preset correction parameter and μ>0, and the unit is mol / (L*dB).

[0075] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A water quality parameter hyperspectral analysis method based on a machine learning model, characterized in that: The following steps are involved: Step S1: Select a number of water quality parameter samples to obtain a water quality parameter sample solution; A standard signal-to-noise ratio Sn0 is set, and a number of signal-to-noise ratio gradient values ​​are set according to the standard signal-to-noise ratio, and a water quality parameter sample solution is subjected to a hyperspectral analysis using spectral bands with different signal-to-noise ratio gradient values ​​to obtain a spectral curve of the water quality parameter sample solution, wherein the spectral curve is the reflectivity at different wavelengths; Obtain the reflectance change rate of the water quality parameter sample according to each signal-to-noise ratio gradient value and the spectral curve corresponding to the water quality parameter sample solution of each signal-to-noise ratio gradient value; record the water quality parameter sample whose reflectance change rate is lower than a preset reflectance change threshold as a non-sensitive water quality parameter; Step S2: The location of the pollution source of the water body is recorded as the pollution source point; a water sample is collected at the pollution source point to obtain the concentration c0 of the non-sensitive water quality parameter; a hyperspectral analysis is performed on the water sample with a spectral band of a standard signal-to-noise ratio to obtain a hyperspectral image, and a spectral curve within a pixel point where the non-sensitive water quality parameter is located is obtained in the hyperspectral image, which is recorded as a source spectral curve; Step S3: any other positions of the water body are recorded as acquisition points. When performing hyperspectral analysis at the acquisition points, hyperspectral images obtained under spectral bands with different signal-to-noise ratio gradient values ​​are obtained and recorded as current hyperspectral images. The spectral curves in each pixel point of the current hyperspectral image are recorded as current spectral curves. In the current hyperspectral image, all current spectral curves of the same pixel in spectral bands with different signal-to-noise ratio gradient values ​​are obtained to obtain the difference value: Where n is the total number of signal-to-noise ratio gradient values, f i (λ) represents the current spectral curve under the spectral band of the i-th signal-to-noise ratio gradient value, λ is the wavelength, and [λ0, λ1] represents the wavelength range of the spectral band; the pixel points whose difference values ​​are less than the preset difference threshold are selected and recorded as non-sensitive pixel points; Step S4: Obtain all current spectral curves of the non-sensitive pixel points, recorded as non-sensitive curves, and select the non-sensitive curve with the highest similarity to the source spectral curve, recorded as the target curve; obtain the signal-to-noise ratio gradient value corresponding to the target curve, recorded as Sn min According to the signal-to-noise ratio gradient value and the standard signal-to-noise ratio, the predicted concentration of the non-sensitive water quality parameter at the collection point is obtained: Pc = c0 + μ (Sn min -Sn0), where μ is a preset correction parameter and μ>0, and the unit is mol / (L*dB).

2. The method for water quality parameter hyperspectral analysis based on a machine learning model according to claim 1, characterized in that: In step S1, the process of setting the signal-to-noise ratio gradient value includes: A signal-to-noise ratio interval threshold is set, the standard signal-to-noise ratio is used as a starting value, and the starting value is increased or decreased according to the signal-to-noise ratio interval threshold to obtain a plurality of signal-to-noise ratio gradient values.

3. The method for water quality parameter hyperspectral analysis based on a machine learning model according to claim 1, characterized in that: In step S1, the process of obtaining the reflection change rate includes: Select several reference points on each spectral curve, and obtain the mean reflectivity according to the reflectivity at each reference point on the spectral curve. Where k represents the kth spectral curve, P x represents the reflectivity at the xth reference point, and m is the total number of reference points; A coordinate system is established with the signal-to-noise ratio as the horizontal coordinate and the reflectivity as the vertical coordinate; a standard line y=Mr is obtained in the coordinate system, where y represents the vertical coordinate of the coordinate system; all spectral curves are placed in the coordinate system, and the starting points of all spectral curves coincide with the origin; in the coordinate system, the difference between each spectral curve and the standard line is obtained The reflectance change rate is Sn i represents the i-th signal-to-noise ratio gradient value, D i Represents the difference in spectral curves corresponding to the i-th signal-to-noise ratio gradient value.

4. The method for water quality parameter hyperspectral analysis based on a machine learning model according to claim 1, characterized in that: In step S1, the water quality parameter samples whose reflectance change rate is greater than or equal to a preset reflectance change threshold are recorded as sensitive water quality parameters.

5. The method for water quality parameter hyperspectral analysis based on a machine learning model according to claim 4 is characterized in that: In step S2, the process of obtaining the spectral curve within the pixel point where the non-sensitive water quality parameter is located in the hyperspectral image includes: All sensitive water quality parameters in the water sample are obtained, and spectral curves of the sensitive water quality parameters are obtained, which are recorded as sensitive spectral curves; pixel points where all sensitive spectral curves are located are selected in the hyperspectral image, which are recorded as sensitive pixel points, and all sensitive pixel points are eliminated, and the spectral curves in all remaining pixel points in the hyperspectral image are recorded as source spectral curves.

6. The method for water quality parameter hyperspectral analysis based on a machine learning model according to claim 1, characterized in that: In step S3, the pixel points whose difference values ​​are greater than or equal to a preset difference threshold are recorded as sensitive pixel points.

7. The method for water quality parameter hyperspectral analysis based on a machine learning model according to claim 1, characterized in that: In step S4, the process of selecting the insensitive curve having the highest similarity to the source spectrum curve includes: Select several comparison points at equal intervals on the source spectrum curve and the non-sensitive curve, and then obtain the similarity between the source spectrum curve and the non-sensitive curve. Where G(t) represents the tth comparison point on the source spectrum curve, T(t) represents the tth comparison point on the non-sensitive curve, and J is the total number of comparison points.

8. A water quality parameter hyperspectral analysis system based on a machine learning model, characterized in that: include: Sensitive classification module: select several water quality parameter samples and obtain water quality parameter sample solutions; A standard signal-to-noise ratio Sn0 is set, and a number of signal-to-noise ratio gradient values ​​are set according to the standard signal-to-noise ratio, and a water quality parameter sample solution is subjected to a hyperspectral analysis using spectral bands with different signal-to-noise ratio gradient values ​​to obtain a spectral curve of the water quality parameter sample solution, wherein the spectral curve is the reflectivity at different wavelengths; Obtain the reflectance change rate of the water quality parameter sample according to each signal-to-noise ratio gradient value and the spectral curve corresponding to the water quality parameter sample solution of each signal-to-noise ratio gradient value; record the water quality parameter sample whose reflectance change rate is lower than a preset reflectance change threshold as a non-sensitive water quality parameter; Pollution source analysis module: the location of the pollution source of the water body is recorded as the pollution source point; water samples are collected at the pollution source point to obtain the concentration c0 of the non-sensitive water quality parameter; the water samples are subjected to hyperspectral analysis with a spectral band of a standard signal-to-noise ratio to obtain a hyperspectral image, and the spectral curve within the pixel point where the non-sensitive water quality parameter is located is obtained in the hyperspectral image and recorded as the source spectral curve; Collection point analysis module: any other position of the water body is recorded as a collection point. When performing hyperspectral analysis at the collection point, a hyperspectral image obtained under a spectral band with different signal-to-noise ratio gradient values ​​is obtained and recorded as a current hyperspectral image. The spectral curve within each pixel point of the current hyperspectral image is recorded as the current spectral curve. In the current hyperspectral image, all current spectral curves of the same pixel in spectral bands with different signal-to-noise ratio gradient values ​​are obtained to obtain the difference value: Where n is the total number of signal-to-noise ratio gradient values, f i (λ) represents the current spectral curve under the spectral band of the i-th signal-to-noise ratio gradient value, λ is the wavelength, and [λ0, λ1] represents the wavelength range of the spectral band; the pixel points whose difference values ​​are less than the preset difference threshold are selected and recorded as non-sensitive pixel points; Concentration prediction module: obtain all current spectral curves of the non-sensitive pixel points, record them as non-sensitive curves, and select the non-sensitive curve with the highest similarity to the source spectral curve, record them as target curves; obtain the signal-to-noise ratio gradient value corresponding to the target curve, record them as Sn min According to the signal-to-noise ratio gradient value and the standard signal-to-noise ratio, the predicted concentration of the non-sensitive water quality parameter at the collection point is obtained: Pc = c0 + μ (Sn min -Sn0), where μ is a preset correction parameter and μ>0, and the unit is mol / (L*dB).

Images