A method and device for monitoring water quality parameters of brackish water lakes

Abstract

The present invention discloses a water quality parameter monitoring method and device suitable for brackish water lakes. The method comprises: collecting first hyperspectral data and water samples in a first brackish water lake, and generating a general water quality parameter inversion model; when the monitored lake is the first brackish water lake, resampling and generating a corresponding target water quality parameter inversion model; when the monitored lake is a second brackish water lake, collecting third hyperspectral data in the second brackish water lake, resampling and generating a corresponding target water quality parameter inversion model; acquiring a hyperspectral image of the monitored lake through the hyperspectral satellite, and calculating the water quality parameters of each pixel in the hyperspectral image; and obtaining the overall water quality parameter situation of the monitored lake. The water quality parameter monitoring method provided by the present invention is very suitable for brackish water lakes. The method integrates multiple models for the water quality monitoring results, which is more accurate than the existing method that only uses a single model, and can greatly improve the efficiency of water quality monitoring.

Description

Technical Field

[0001] The present application relates to the technical field of water quality research, and in particular to a method and device for monitoring water quality parameters suitable for brackish water lakes. Background Art

[0002] Water resources are important natural resources for human beings, and are also the basic conditions for the survival and development of human beings, animals, plants and other life forms. Among water resources, saltwater lakes are lakes with a salinity of more than 1‰, among which lakes with a salinity between 1‰ and 24.7‰ are slightly saline lakes.

[0003] With the development of society and industry and the growth of population, a large amount of water resources have been polluted. Therefore, as water pollution becomes increasingly serious, water quality monitoring, as a basic work in water pollution control, has become more important.

[0004] The existing methods are to obtain the water quality of lakes by measuring the water quality of lake samples by chemical methods, and to monitor the water quality of lakes by using remote sensing monitoring technology. The detection process of the water quality of water samples by chemical methods is complicated and takes a long time, so it is time-consuming and inefficient, and the water quality of the collected water samples often cannot accurately reflect the water quality of the entire lake, and it will also cause certain pollution to the environment. Although there have been some studies on the monitoring of lake water quality by remote sensing monitoring technology, the inventors have realized that there are still few studies on the use of remote sensing monitoring methods to monitor water quality parameters in brackish water conditions, and the models used in the existing monitoring methods are relatively simple. Summary of the invention

[0005] Based on this, in order to solve the above technical problems, a method and device for monitoring water quality parameters suitable for brackish water lakes are provided.

[0006] In a first aspect, a method for monitoring water quality parameters of a brackish water lake comprises:

[0007] Step S1, respectively collecting corresponding first hyperspectral data and water samples at a plurality of collection points preset in a first brackish water lake, obtaining a plurality of first hyperspectral data and a corresponding plurality of water samples, and analyzing the water samples through a water body experiment to obtain a measured value of a water quality parameter of each water sample; and generating a universal water quality parameter inversion model according to the plurality of first hyperspectral data and the measured values ​​of the water quality parameters of the corresponding water samples;

[0008] Step S2, when the monitored lake is a first brackish water lake, based on the hyperspectral satellite to be used, each first hyperspectral data is resampled according to the resolution of the hyperspectral satellite to obtain a plurality of second hyperspectral data matching the resolution of the hyperspectral satellite; and a corresponding target water quality parameter inversion model is generated according to the plurality of second hyperspectral data and the measured values ​​of the water quality parameters of the corresponding water samples;

[0009] Step S3, when the monitored lake is a second brackish water lake, corresponding third hyperspectral data are respectively collected at a plurality of collection points preset in the second brackish water lake to obtain a plurality of third hyperspectral data, and a water quality parameter inversion value corresponding to each third hyperspectral data is calculated by the universal water quality parameter inversion model, and the obtained water quality parameter inversion value corresponding to each third hyperspectral data is used as the water quality parameter measured value corresponding to each third hyperspectral data; based on the hyperspectral satellite to be used, each third hyperspectral data is resampled according to the resolution of the hyperspectral satellite to obtain a plurality of fourth hyperspectral data matching the resolution of the hyperspectral satellite; and a corresponding target water quality parameter inversion model is generated according to the plurality of fourth hyperspectral data and the measured values ​​of water quality parameters of the corresponding water samples;

[0010] Step S4, acquiring a hyperspectral image of the monitored lake through the hyperspectral satellite, and calculating the water quality parameter of each pixel in the hyperspectral image using a corresponding target water quality parameter inversion model;

[0011] Step S25, calculate the average value of the corresponding water quality parameter of each pixel according to the water quality parameter of each pixel calculated by each water quality parameter inversion model in the target water quality parameter inversion model; take the average value of the water quality parameter of each pixel as the inversion value of the water quality parameter of the pixel, thereby obtaining the overall water quality parameter situation of the monitored lake.

[0012] Optionally, the step S1 of generating a universal water quality parameter inversion model according to the plurality of first hyperspectral data and the measured values ​​of water quality parameters of corresponding water samples specifically includes:

[0013] Step S1A, preprocessing each first hyperspectral data to obtain a plurality of first preprocessed hyperspectral data;

[0014] Step S1B, performing correlation analysis on the plurality of first hyperspectral data and the corresponding measured values ​​of water quality parameters, and generating a corresponding water quality parameter inversion model; performing correlation analysis on the plurality of first pre-processed hyperspectral data and the corresponding measured values ​​of water quality parameters, and generating a corresponding water quality parameter inversion model;

[0015] Step S1C, selecting a water quality parameter inversion model whose accuracy meets a preset condition from each generated water quality parameter inversion model as a universal water quality parameter inversion model;

[0016] The step S2 of generating a corresponding target water quality parameter inversion model according to the plurality of second hyperspectral data and the measured values ​​of water quality parameters of the corresponding water samples specifically includes:

[0017] Step S2A, preprocessing each second hyperspectral data to obtain a plurality of second preprocessed hyperspectral data;

[0018] Step S2B, performing correlation analysis on the plurality of second hyperspectral data and the corresponding measured values ​​of water quality parameters, and generating a corresponding water quality parameter inversion model; performing correlation analysis on the plurality of second pre-processed hyperspectral data and the corresponding measured values ​​of water quality parameters, and generating a corresponding water quality parameter inversion model;

[0019] Step S2C, selecting a water quality parameter inversion model whose accuracy meets a preset condition from each generated water quality parameter inversion model as a target water quality parameter inversion model;

[0020] The step S3 of generating a corresponding target water quality parameter inversion model according to the plurality of fourth hyperspectral data and the measured values ​​of water quality parameters of the corresponding water samples specifically includes:

[0021] Step S3A, preprocessing each fourth hyperspectral data to obtain a plurality of fourth preprocessed hyperspectral data;

[0022] Step S3B, performing correlation analysis on the plurality of fourth hyperspectral data and the corresponding measured values ​​of water quality parameters, and generating a corresponding water quality parameter inversion model; performing correlation analysis on the plurality of fourth pre-processed hyperspectral data and the corresponding measured values ​​of water quality parameters, and generating a corresponding water quality parameter inversion model;

[0023] Step S3C, selecting a water quality parameter inversion model whose accuracy reaches a preset condition from each generated water quality parameter inversion model as a target water quality parameter inversion model.

[0024] Further optionally, the step S1A specifically includes:

[0025] Performing a first-order derivative processing on each first hyperspectral data to obtain corresponding first first-order hyperspectral data; performing a second-order derivative processing on each first hyperspectral data to obtain corresponding first second-order hyperspectral data; performing a third-order derivative processing on each first hyperspectral data to obtain corresponding first third-order hyperspectral data;

[0026] The step S2A specifically includes:

[0027] Performing a first-order derivative process on each second hyperspectral data to obtain corresponding second first-order hyperspectral data; performing a second-order derivative process on each second hyperspectral data to obtain corresponding second second-order hyperspectral data; performing a third-order derivative process on each second hyperspectral data to obtain corresponding second third-order hyperspectral data;

[0028] The step S3A specifically includes:

[0029] Perform a first-order derivative processing on each fourth hyperspectral data to obtain the corresponding fourth first-order hyperspectral data; perform a second-order derivative processing on each fourth hyperspectral data to obtain the corresponding fourth second-order hyperspectral data; perform a third-order derivative processing on each fourth hyperspectral data to obtain the corresponding fourth third-order hyperspectral data.

[0030] Further optionally, the step S1B specifically includes:

[0031] Performing correlation analysis on the plurality of first hyperspectral data and the corresponding measured values ​​of water quality parameters, and performing a one-dimensional linear fitting on the single band with the highest correlation among the plurality of first hyperspectral data and the corresponding measured values ​​of water quality parameters to obtain a corresponding water quality parameter inversion model;

[0032] Correlation analysis is performed on a plurality of first-order hyperspectral data, a plurality of first-second-order hyperspectral data, and a plurality of first-third-order hyperspectral data with the measured values ​​of water quality parameters, and a single band with the highest correlation in the first-order hyperspectral data, a single band with the highest correlation in the first-second-order hyperspectral data, and a single band with the highest correlation in the first-third-order hyperspectral data are respectively subjected to one-dimensional linear fitting with the measured values ​​of water quality parameters to obtain a corresponding water quality parameter inversion model; and a band ratio with the highest correlation in the first-order hyperspectral data, a band ratio with the highest correlation in the first-second-order hyperspectral data, and a band ratio with the highest correlation in the first-third-order hyperspectral data are respectively subjected to one-dimensional linear fitting with the measured values ​​of water quality parameters to obtain a corresponding water quality parameter inversion model;

[0033] The step S2B specifically includes:

[0034] Performing correlation analysis on the plurality of second hyperspectral data and the corresponding measured values ​​of water quality parameters, and performing a one-dimensional linear fitting on the single band with the highest correlation among the plurality of second hyperspectral data and the corresponding measured values ​​of water quality parameters to obtain a corresponding water quality parameter inversion model;

[0035] Correlation analysis is performed on multiple second-order hyperspectral data, multiple second-order hyperspectral data, and multiple second-order hyperspectral data with the measured values ​​of water quality parameters, and a single band with the highest correlation in the second-order hyperspectral data, a single band with the highest correlation in the second-order hyperspectral data, and a single band with the highest correlation in the second-order hyperspectral data are respectively fitted with the measured values ​​of water quality parameters to obtain corresponding water quality parameter inversion models; and a band ratio with the highest correlation in the second-order hyperspectral data, a band ratio with the highest correlation in the second-order hyperspectral data, and a band ratio with the highest correlation in the second-order hyperspectral data are respectively fitted with the measured values ​​of water quality parameters to obtain corresponding water quality parameter inversion models;

[0036] The step S3B specifically includes:

[0037] Performing correlation analysis on the plurality of fourth hyperspectral data and the corresponding measured values ​​of water quality parameters, and performing a one-dimensional linear fitting on the single band with the highest correlation among the plurality of fourth hyperspectral data and the corresponding measured values ​​of water quality parameters to obtain a corresponding water quality parameter inversion model;

[0038] Correlation analysis is performed on multiple fourth first-order hyperspectral data, multiple fourth second-order hyperspectral data and multiple fourth third-order hyperspectral data with the actual measured values ​​of water quality parameters, and a single band with the highest correlation in the fourth first-order hyperspectral data, a single band with the highest correlation in the fourth second-order hyperspectral data and a single band with the highest correlation in the fourth third-order hyperspectral data are respectively subjected to a one-dimensional linear fit with the actual measured values ​​of water quality parameters to obtain a corresponding water quality parameter inversion model; and a band ratio with the highest correlation in the fourth first-order hyperspectral data, a band ratio with the highest correlation in the fourth second-order hyperspectral data and a band ratio with the highest correlation in the fourth third-order hyperspectral data are respectively subjected to a one-dimensional linear fit with the actual measured values ​​of water quality parameters to obtain a corresponding water quality parameter inversion model.

[0039] Further optionally, the screened water quality parameter inversion models with accuracy that meets the preset conditions are:

[0040] Calculate the determination coefficient R of each water quality parameter inversion model generated 2 , select the determination coefficient R 2 Inversion model of target water quality parameters exceeding preset determination coefficient thresholds.

[0041] Further optionally, the coefficient of determination Where n is the number of hyperspectral data collected, t i is the measured value of water quality parameter corresponding to each hyperspectral data, y i is the water quality parameter inversion value of each hyperspectral data calculated using the water quality parameter inversion model. It is the average value of the measured values ​​of water quality parameters corresponding to all hyperspectral data; the preset determination coefficient threshold is 0.65.

[0042] Further optionally, the step S3 of calculating the water quality parameter inversion value corresponding to each third hyperspectral data by using the universal water quality parameter inversion model specifically includes:

[0043] Calculating the water quality parameters of each third hyperspectral data using a universal water quality parameter inversion model;

[0044] According to the water quality parameters of each third hyperspectral data calculated by each water quality parameter inversion model in the universal water quality parameter inversion model, the average value of the corresponding water quality parameters of each third hyperspectral data is calculated; and the average value of the water quality parameters of each third hyperspectral data is used as the inversion value of the water quality parameters of the third hyperspectral data.

[0045] Further optionally, both step S2C and step S3C further include:

[0046] Calculate the root mean square error RMSE and / or average relative error RE of each screened water quality parameter inversion model, and remove the water quality parameter inversion model whose root mean square error RMSE and / or average relative error RE exceeds the corresponding preset threshold value from the target water quality parameter inversion model; wherein:

[0047] Root mean square error

[0048] Mean relative error

[0049] Where n is the number of hyperspectral data collected, t i is the measured value of water quality parameter corresponding to each hyperspectral data, y i It is the water quality parameter inversion value of each hyperspectral data calculated by using the water quality parameter inversion model.

[0050] Further optionally, the water quality parameter is at least one of permanganate index, chemical oxygen demand, ammonia nitrogen content, total phosphorus content, total nitrogen content, suspended matter content and chlorophyll a content; the hyperspectral satellite is Ziyuan-1 satellite or Gaofen-5 satellite; the hyperspectral data are collected through a portable ground spectrometer; and the correlation analysis is performed using the Wilson correlation coefficient method.

[0051] In a second aspect, a water quality parameter monitoring device suitable for brackish water lakes comprises:

[0052] A universal water quality parameter inversion model generation module is used to collect corresponding first hyperspectral data and water samples at multiple collection points preset in a first brackish water lake, respectively, to obtain multiple first hyperspectral data and corresponding multiple water samples, and to analyze the water samples through water body experiments to obtain the measured value of the water quality parameters of each water sample; and to generate a universal water quality parameter inversion model according to the multiple first hyperspectral data and the measured values ​​of the water quality parameters of the corresponding water samples;

[0053] A first target water quality parameter inversion model generation module is used for, when the monitored lake is a first brackish water lake, resampling each first hyperspectral data according to the resolution of the hyperspectral satellite based on the hyperspectral satellite to be used, to obtain a plurality of second hyperspectral data matching the resolution of the hyperspectral satellite; and generating a corresponding target water quality parameter inversion model according to the plurality of second hyperspectral data and the measured values ​​of the water quality parameters of the corresponding water samples;

[0054] A second target water quality parameter inversion model generation module is used for, when the monitored lake is a second brackish water lake, respectively collecting corresponding third hyperspectral data at a plurality of collection points preset in the second brackish water lake to obtain a plurality of third hyperspectral data, and calculating a water quality parameter inversion value corresponding to each third hyperspectral data through the universal water quality parameter inversion model, and using the obtained water quality parameter inversion value corresponding to each third hyperspectral data as the measured water quality parameter value corresponding to each third hyperspectral data; based on the hyperspectral satellite to be used, resampling each third hyperspectral data according to the resolution of the hyperspectral satellite to obtain a plurality of fourth hyperspectral data matching the resolution of the hyperspectral satellite; and generating a corresponding target water quality parameter inversion model according to the plurality of fourth hyperspectral data and the measured water quality parameter values ​​of the corresponding water samples;

[0055] A hyperspectral image acquisition module, which acquires a hyperspectral image of the monitored lake through the hyperspectral satellite, and calculates the water quality parameter of each pixel in the hyperspectral image using a corresponding target water quality parameter inversion model;

[0056] The inversion value determination module calculates the average value of the corresponding water quality parameters of each pixel based on the water quality parameters of each pixel calculated by each water quality parameter inversion model in the target water quality parameter inversion model; the average value of the water quality parameters of each pixel is used as the inversion value of the water quality parameters of the pixel, thereby obtaining the overall water quality parameters of the monitored lake.

[0057] The present invention has at least the following beneficial effects:

[0058] Based on further analysis and research on the existing technical problems, the present invention recognizes that there are currently few studies on the use of remote sensing monitoring methods to monitor water quality parameters in brackish water conditions, and the models used in existing monitoring methods are relatively simple; there are relatively few phytoplankton on the surface of brackish water lakes, so that the hyperspectral images of brackish water lakes obtained by hyperspectral satellites can better reflect the conditions of the water body itself. Therefore, through the water quality parameter monitoring method provided by the present invention, the water quality parameter inversion model obtained is used to invert the water quality parameters of each pixel in the hyperspectral image, which will be more accurate, thereby being able to more accurately obtain the overall water quality parameter conditions of the monitored brackish water lakes; that is, the water quality parameter monitoring method provided by the present invention is very suitable for brackish water lakes.

[0059] At the same time, in the water quality parameter monitoring method provided by the present invention, the generated target water quality parameter inversion model includes multiple models, and the water quality parameter inversion value of each pixel is taken as the average value calculated by multiple models in the target water quality parameter inversion model. That is to say, the water quality monitoring result of the water quality monitoring method provided by the present invention is obtained by integrating the calculation results of multiple models, which is more accurate than the existing monitoring method that only uses a single model to monitor water quality parameters.

[0060] In addition, through the water quality parameter monitoring method provided by the present invention, after obtaining multiple first hyperspectral data and measured water quality parameter values ​​of multiple water samples in a lake, when the lake to be monitored is changed, there is no need to collect water samples in the new lake, only hyperspectral data need to be collected, and the corresponding measured water quality parameter values ​​can be taken as the inversion values ​​calculated by the general water quality parameter inversion model. Therefore, the present invention can greatly improve the monitoring efficiency; when different hyperspectral satellites are used to monitor the water quality parameters of the same lake, after collecting hyperspectral data in the lake, when the satellite to be used changes, it is only necessary to resample the spectral data collected by the ASD portable ground feature spectrometer according to the corresponding resolution of the satellite, and then repeat the subsequent modeling process; compared with the existing monitoring method that requires repeated sampling at the collection point for different hyperspectral satellites, the water quality parameter monitoring method provided by the present invention can greatly improve the efficiency and convenience of water quality monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 A schematic flow chart of a method for monitoring water quality parameters of brackish water lakes provided by one embodiment of the present invention;

[0062] Figure 2 A graph showing the monitoring results of the total nitrogen content of Bosten Lake provided by one embodiment of the present invention;

[0063] Figure 3 A block diagram of the module architecture of a water quality parameter monitoring device suitable for brackish water lakes provided by one embodiment of the present invention;

[0064] Figure 4 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0065] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0066] Embodiment 1

[0067] In this embodiment, if Figure 1 As shown, a method for monitoring water quality parameters applicable to brackish water lakes is provided, comprising the following steps:

[0068] Step S1, respectively collecting corresponding first hyperspectral data and water samples at multiple collection points preset in a first brackish water lake, obtaining multiple first hyperspectral data and corresponding multiple water samples, and analyzing the water samples through water body experiments to obtain measured values ​​of water quality parameters of each water sample; and generating a universal water quality parameter inversion model based on the multiple first hyperspectral data and the measured values ​​of water quality parameters of the corresponding water samples.

[0069] Specifically, the multiple collection points preset in the first brackish water lake can be 50 collection points evenly distributed in a certain area of ​​the first brackish water lake. Since the water depth near the shore is relatively shallow, the collection points should be set to avoid being close to the shore. At each collection point, a hyperspectral curve is collected at a distance of about 200 cm from the water surface. The measurement sequence is standard plate measurement (the probe is vertically downward, about 25 cm away from the standard plate), inclined water body measurement (the probe nadir angle is 40°, downward), inclined sky light measurement (probe upward, the probe zenith angle is 40°), standard plate measurement (probe vertically downward, about 25 cm away from the standard plate), inclined water body measurement (probe downward, the probe nadir angle is 40°), inclined sky light measurement (probe upward, the probe zenith angle is 40°), standard plate measurement (probe vertically downward, about 25 cm away from the standard plate), inclined water body measurement (probe downward, the probe nadir angle is 40°), inclined sky light measurement (probe upward, the probe zenith angle is 40°), standard plate measurement (probe vertically downward, about 25 cm away from the standard plate), inclined water body measurement (probe downward, the probe nadir angle is 40°), inclined sky light measurement (probe upward, the probe zenith angle is 40°). After calculation, each of the 50 collection points has a corresponding hyperspectral curve. The effective spectrum of water is 400-950nm, so the band included in each hyperspectral curve is also roughly between 400-950nm. In addition, the spectral curve is collected at the collection point through the ASD portable ground feature spectrometer. The working spectrum range of the ASD portable ground feature spectrometer is 350-2500nm, and the spectral resolution is 1nm. That is to say, the ASD portable ground feature spectrometer can identify the reflectivity of the 450nm, 451nm, 452nm... bands, and accordingly, each first hyperspectral data collected contains the reflectivity of the 450nm, 451nm, 452nm... bands. In addition, the collected water samples are generally mixed water bodies 20cm to 70cm below the water surface. After the water samples are collected, it is necessary to obtain the measured values ​​of the water quality parameters of the water samples through water body experimental analysis. The water quality parameters can be at least one of the permanganate index, chemical oxygen demand (COD), ammonia nitrogen content (NH3-N), total phosphorus content (P), total nitrogen content (N), suspended solids content (SS) and chlorophyll a content. Whichever water quality parameter needs to be monitored, the water quality parameter in the water sample can be analyzed accordingly. Of course, water quality parameters are not limited to those listed above.

[0070] Furthermore, the step S1 of generating a universal water quality parameter inversion model according to the plurality of first hyperspectral data and the measured values ​​of water quality parameters of corresponding water samples specifically includes:

[0071] Step S1A, preprocessing each first hyperspectral data to obtain a plurality of first preprocessed hyperspectral data; specifically, as follows:

[0072] Performing a first-order derivative processing on each first hyperspectral data to obtain the corresponding first first-order hyperspectral data, that is, each first first-order hyperspectral data contains the relationship between each band of the first hyperspectral data and the first-order derivative of the corresponding reflectivity; performing a second-order derivative processing on each first hyperspectral data to obtain the corresponding first second-order hyperspectral data, that is, each first second-order hyperspectral data contains the relationship between each band of the first hyperspectral data and the second-order derivative of the corresponding reflectivity; performing a third-order derivative processing on each first hyperspectral data to obtain the corresponding first third-order hyperspectral data, that is, each first third-order hyperspectral data contains the relationship between each band of the first hyperspectral data and the third-order derivative of the corresponding reflectivity;

[0073] Step S1B, performing correlation analysis on the plurality of first hyperspectral data and the corresponding measured values ​​of water quality parameters, and generating a corresponding water quality parameter inversion model; performing correlation analysis on the plurality of first pre-processed hyperspectral data and the corresponding measured values ​​of water quality parameters, and generating a corresponding water quality parameter inversion model; specifically:

[0074] By using the Pearson correlation coefficient method, the multiple first hyperspectral data are subjected to correlation analysis with the corresponding measured values ​​of water quality parameters, and the single band with the highest correlation among the multiple first hyperspectral data is found, and the single band with the highest correlation among the multiple first hyperspectral data is subjected to a one-dimensional linear fitting with the corresponding measured values ​​of water quality parameters to obtain a corresponding water quality parameter inversion model, and generate a corresponding linear model: y=a1x+b1, y is the water quality parameter, x is the reflectance corresponding to the single band with the highest correlation, and a1 and b1 are constants;

[0075] A correlation analysis is performed on multiple first-order hyperspectral data and the measured values ​​of water quality parameters to find the single band with the highest correlation among the multiple first-order hyperspectral data, and a one-dimensional linear fitting is performed on the single band with the highest correlation among the multiple first-order hyperspectral data and the corresponding measured values ​​of water quality parameters to obtain the corresponding water quality parameter inversion model, and a corresponding linear model is generated: y=a2x+b2, y is the water quality parameter, x is the absolute value of the first-order derivative of the reflectivity corresponding to the single band with the highest correlation, and a2 and b2 are constants; a correlation analysis is performed on multiple first- and second-order hyperspectral data and the measured values ​​of water quality parameters to find the single band with the highest correlation among the multiple first- and second-order hyperspectral data, and a one-dimensional linear fitting is performed on the single band with the highest correlation among the multiple first- and second-order hyperspectral data and the corresponding measured values ​​of water quality parameters. A one-dimensional linear fitting is performed to obtain the corresponding water quality parameter inversion model, and a corresponding linear model is generated: y=a3x+b3, y is the water quality parameter, x is the absolute value of the second-order derivative of the reflectivity corresponding to the single band with the highest correlation, and a3 and b3 are constants; similarly, a plurality of first- and third-order hyperspectral data are subjected to correlation analysis with the measured values ​​of water quality parameters to find the single band with the highest correlation among the plurality of third-order hyperspectral data, and a one-dimensional linear fitting is performed between the single band with the highest correlation among the plurality of first- and third-order hyperspectral data and the measured values ​​of the corresponding water quality parameters to obtain the corresponding water quality parameter inversion model, and a corresponding linear model is generated: y=a4x+b4, y is the water quality parameter, x is the absolute value of the third-order derivative of the reflectivity corresponding to the single band with the highest correlation, and a4 and b4 are constants;

[0076] At the same time, find out the band ratio with the highest correlation in multiple first-order hyperspectral data. The band ratio in the first-order hyperspectral data is the absolute value of the ratio of the first-order derivatives of the reflectance corresponding to the two bands. The band ratio can be defined as QRVI (|B1 / B2|) or QNDVI (|(B1-B2) / (B1+B2)|), where B1 and B2 are the first-order derivatives of the reflectance corresponding to the two bands; perform a one-dimensional linear fit on the band ratio with the highest correlation in multiple first-order hyperspectral data and the corresponding measured values ​​of water quality parameters to obtain the corresponding water quality parameters. The linear inversion model is used to generate the corresponding linear model: y=a5x+b5, where y is the water quality parameter, x is the band ratio with the highest correlation, and a5 and b5 are constants. Similarly, the band ratio with the highest correlation in multiple first-order and second-order hyperspectral data is found. The band ratio in the first-order and second-order hyperspectral data is the absolute value of the ratio of the second-order derivatives of the reflectivity corresponding to the two bands. The band ratio can be defined as QRVI (|B1 / B2|) or QNDVI (|(B1-B2) / (B1+B2)|), where B1 and B2 are the second-order derivatives of the reflectivity corresponding to the two bands. The band ratio with the highest correlation in multiple first-order and second-order hyperspectral data is fitted with the corresponding measured values ​​of water quality parameters to obtain the corresponding water quality parameter inversion model, and the corresponding linear model is generated: y = a6x + b6, y is the water quality parameter, x is the band ratio with the highest correlation, a6 and b6 are constants; similarly, the band ratio with the highest correlation in multiple first-order and third-order hyperspectral data is found. The band ratio in the first-order and third-order hyperspectral data is the absolute value of the ratio of the third-order derivative of the reflectivity corresponding to the two bands. The band ratio can be defined as QRVI (|B 1 / B2|) or QNDVI(|(B1-B2) / (B1+B2)|), where B1 and B2 are the third-order derivatives of reflectance corresponding to the two bands; a linear fit is performed on the band ratio with the highest correlation in the multiple first and third order hyperspectral data and the corresponding measured values ​​of water quality parameters to obtain the corresponding water quality parameter inversion model, and a corresponding linear model is generated: y=a7x+b7, y is the water quality parameter, x is the band ratio with the highest correlation, and a7 and b7 are constants; that is, after step S1B, at least seven water quality parameter inversion models can be obtained;

[0077] Step S1C, selecting a water quality parameter inversion model whose accuracy meets a preset condition from each generated water quality parameter inversion model as a universal water quality parameter inversion model;

[0078] Specifically, the water quality parameter inversion model selected in step S1C with accuracy meeting the preset conditions specifically includes:

[0079] For each generated water quality parameter inversion model, calculate the corresponding determination coefficient R 2 , Where n is the number of hyperspectral data collected in step S1, that is, the number of water samples collected in step S1; t i is the measured value of the water quality parameter corresponding to each hyperspectral data, that is, the measured value of the water quality parameter of each water sample; i is the water quality parameter inversion value of each hyperspectral data calculated using the water quality parameter inversion model, that is, the water quality parameter inversion value of each water sample calculated using the water quality parameter inversion model; is the average value of the measured values ​​of water quality parameters corresponding to all hyperspectral data, that is, the average value of the measured values ​​of water quality parameters of all water samples;

[0080] Coefficient of determination R 2 It is used to measure the accuracy of the generated water quality parameter inversion model, R 2 It has no unit and its value is between 0 and 1. The closer it is to 1, the better the regression fitting effect of the water quality parameter inversion model. 2 The water quality parameter inversion model with a value exceeding 0.65 is used as the general water quality parameter inversion model.

[0081] Step S2, when the monitored lake is a first brackish water lake, based on the hyperspectral satellite to be used, each first hyperspectral data is resampled according to the resolution of the hyperspectral satellite to obtain a plurality of second hyperspectral data matching the resolution of the hyperspectral satellite; and based on the plurality of second hyperspectral data and the measured values ​​of the water quality parameters of the corresponding water samples, a corresponding target water quality parameter inversion model is generated.

[0082] Step S2 is applicable to the case where the lake to be monitored is the first brackish water lake in step S1. Since the monitored lake is the same as the lake where the hyperspectral data and water samples are sampled to establish the universal water quality parameter inversion model in step S1, there is no need to resample the hyperspectral data and water samples in the monitored lake. It is only necessary to resample the first hyperspectral data collected in step S1 and then perform modeling. Since the present application monitors the water quality parameters in brackish water lakes by remote sensing monitoring methods, it is necessary to use hyperspectral satellites to obtain hyperspectral images of the monitored brackish water lakes. The spectral resolution of hyperspectral satellites is low, and the resolution is usually 5nm, 10nm or 20nm, etc., so it cannot achieve the wide working spectral range and 1nm resolution of the ASD portable ground feature spectrometer. That is to say, there will be a lot of bands of data that cannot be recognized by hyperspectral satellites in the spectral data collected by the ASD portable ground object spectrometer. Therefore, it is necessary to resample each first hyperspectral data collected by the ASD portable ground object spectrometer so that the bands in each first hyperspectral data can be recognized by the hyperspectral satellite, thereby ensuring the accuracy of water quality monitoring using hyperspectral satellites. In other words, resampling is to extract the bands and corresponding reflectances that can be recognized by the hyperspectral satellite used in each first hyperspectral data collected by the ASD portable ground object spectrometer, so as to obtain the corresponding second hyperspectral data for each first hyperspectral data. In addition, the commonly used hyperspectral satellite can be, but is not limited to, the Ziyuan-1 satellite or the Gaofen-5 satellite.

[0083] Similar to step S1, the step S2 of generating a corresponding target water quality parameter inversion model according to the plurality of second hyperspectral data and the measured values ​​of the water quality parameters of the corresponding water samples specifically includes:

[0084] Step S2A, preprocessing each second hyperspectral data to obtain a plurality of second preprocessed hyperspectral data; specifically, as follows:

[0085] Performing a first-order derivative process on each second hyperspectral data to obtain corresponding second first-order hyperspectral data; performing a second-order derivative process on each second hyperspectral data to obtain corresponding second second-order hyperspectral data; performing a third-order derivative process on each second hyperspectral data to obtain corresponding second third-order hyperspectral data;

[0086] Step S2B, performing correlation analysis on the plurality of second hyperspectral data and the corresponding measured values ​​of water quality parameters, and generating a corresponding water quality parameter inversion model; performing correlation analysis on the plurality of second pre-processed hyperspectral data and the corresponding measured values ​​of water quality parameters, and generating a corresponding water quality parameter inversion model; specifically:

[0087] Performing correlation analysis on the plurality of second hyperspectral data and the corresponding measured values ​​of water quality parameters, and performing a one-dimensional linear fitting on the single band with the highest correlation among the plurality of second hyperspectral data and the corresponding measured values ​​of water quality parameters to obtain a corresponding water quality parameter inversion model;

[0088] Correlation analysis is performed on multiple second first-order hyperspectral data, multiple second second-order hyperspectral data and multiple second third-order hyperspectral data with the measured values ​​of water quality parameters, and a single band with the highest correlation in the second first-order hyperspectral data, a single band with the highest correlation in the second second-order hyperspectral data and a single band with the highest correlation in the second third-order hyperspectral data are respectively subjected to one-dimensional linear fitting with the measured values ​​of water quality parameters to obtain corresponding water quality parameter inversion models; and a band ratio with the highest correlation in the second first-order hyperspectral data, a band ratio with the highest correlation in the second second-order hyperspectral data and a band ratio with the highest correlation in the second third-order hyperspectral data are respectively subjected to one-dimensional linear fitting with the measured values ​​of water quality parameters to obtain corresponding water quality parameter inversion models; similarly, after step S2B, at least seven water quality parameter inversion models can be obtained;

[0089] Step S2C, selecting a water quality parameter inversion model whose accuracy meets a preset condition from each generated water quality parameter inversion model as a target water quality parameter inversion model;

[0090] Similarly, the water quality parameter inversion model selected in step S2C with accuracy meeting the preset conditions includes:

[0091] For each generated water quality parameter inversion model, calculate the corresponding determination coefficient R 2 , Where n is the number of hyperspectral data collected in step S1, that is, the number of water samples collected in step S1; t i is the measured value of the water quality parameter corresponding to each hyperspectral data, that is, the measured value of the water quality parameter of each water sample; i is the water quality parameter inversion value of each hyperspectral data calculated using the water quality parameter inversion model, that is, the water quality parameter inversion value of each water sample calculated using the water quality parameter inversion model; is the average value of the measured values ​​of water quality parameters corresponding to all hyperspectral data, that is, the average value of the measured values ​​of water quality parameters of all water samples;

[0092] Coefficient of determination R 2 It is used to measure the accuracy of the generated water quality parameter inversion model, R 2 It has no unit and its value is between 0 and 1. The closer it is to 1, the better the regression fitting effect of the water quality parameter inversion model. 2The water quality parameter inversion model exceeding 0.65 is used as the target water quality parameter inversion model for monitoring the monitored lakes.

[0093] After obtaining the target water quality parameter inversion model, the target water quality parameter inversion model can be optimized again to select the excellent target water quality parameter inversion model, specifically:

[0094] The root mean square error RMSE and / or average relative error RE of each screened water quality parameter inversion model is calculated, and the water quality parameter inversion model whose root mean square error RMSE and / or average relative error RE exceeds the corresponding preset threshold is removed from the target water quality parameter inversion model; the preset threshold of the average relative error RE may be 20%; the water quality parameter inversion model whose root mean square error RMSE and / or average relative error RE exceeds the corresponding preset threshold is considered to be a less excellent model among the target water quality parameter inversion models; wherein:

[0095] Root mean square error

[0096] Mean relative error

[0097] Where n is the number of hyperspectral data collected, t i is the measured value of water quality parameter corresponding to each hyperspectral data, y i is the water quality parameter inversion value of each hyperspectral data calculated by using the water quality parameter inversion model; n is the number of hyperspectral data collected in step S1, that is, the number of water samples collected in step S1; t i is the measured value of the water quality parameter corresponding to each hyperspectral data, that is, the measured value of the water quality parameter of each water sample; i It is the water quality parameter inversion value of each hyperspectral data calculated by the water quality parameter inversion model, that is, the water quality parameter inversion value of each water sample calculated by the water quality parameter inversion model; after removing the water quality parameter inversion model whose root mean square error RMSE and / or average relative error RE exceeds the corresponding preset threshold value in the target water quality parameter inversion model, the target water quality parameter inversion model is further optimized, and the selected excellent target water quality parameter inversion model is used for subsequent water quality monitoring; these optimized models that meet the quality control requirements can ensure the accuracy of water quality monitoring because the calculation precision and accuracy are sufficiently guaranteed.

[0098] Step S3, when the monitored lake is a second brackish water lake, corresponding third hyperspectral data are respectively collected at a plurality of collection points preset in the second brackish water lake to obtain a plurality of third hyperspectral data, and a water quality parameter inversion value corresponding to each third hyperspectral data is calculated by the universal water quality parameter inversion model, and the obtained water quality parameter inversion value corresponding to each third hyperspectral data is used as the measured value of the water quality parameter corresponding to each third hyperspectral data;

[0099] Based on the hyperspectral satellite to be used, each third hyperspectral data is resampled according to the resolution of the hyperspectral satellite to obtain a plurality of fourth hyperspectral data matching the resolution of the hyperspectral satellite; and based on the plurality of fourth hyperspectral data and the measured values ​​of water quality parameters of corresponding water samples, a corresponding target water quality parameter inversion model is generated.

[0100] Step S3 is applicable to the case where the lake to be monitored is not the first brackish water lake in step S1. Since the monitored lake is different from the lake where the hyperspectral data and water sample are sampled by establishing the universal water quality parameter inversion model in step S1, it is necessary to resample in the monitored lake, and the preset multiple collection points can also be 50 collection points evenly distributed in a certain area of ​​the monitored lake. Different from sampling in the first brackish water lake, at this time, only 50 corresponding third hyperspectral data need to be collected at 50 collection points in the monitored lake, and then the universal water quality parameter inversion model obtained above is used to calculate the water quality parameter inversion value corresponding to each third hyperspectral data, and then the water quality parameter inversion value corresponding to each third hyperspectral data is used as the water quality parameter measured value of each third hyperspectral data corresponding water sample, and there is no need to measure 50 water samples again and then obtain the water quality parameter measured value of the water sample through water body experimental analysis.

[0101] Specifically, the step S3 of calculating the water quality parameter inversion value corresponding to each third hyperspectral data by using the universal water quality parameter inversion model specifically includes:

[0102] Calculating the water quality parameters of each third hyperspectral data using a universal water quality parameter inversion model;

[0103] According to the water quality parameters of each third hyperspectral data calculated by each water quality parameter inversion model in the universal water quality parameter inversion model, the average value of the corresponding water quality parameters of each third hyperspectral data is calculated; and the average value of the water quality parameters of each third hyperspectral data is used as the inversion value of the water quality parameters of the third hyperspectral data.

[0104] Similar to step S2, the obtained third hyperspectral data also needs to be resampled according to the resolution of the hyperspectral satellite, where the bands and corresponding reflectances that can be recognized by the hyperspectral satellite used are cut out, so as to obtain corresponding fourth hyperspectral data for each third hyperspectral data.

[0105] Further, similar to steps S1 and S2, the step S3 of generating a corresponding target water quality parameter inversion model according to the plurality of fourth hyperspectral data and the measured values ​​of water quality parameters of the corresponding water samples specifically includes:

[0106] Step S3A, preprocessing each fourth hyperspectral data to obtain a plurality of fourth preprocessed hyperspectral data; specifically:

[0107] Performing a first-order derivative processing on each fourth hyperspectral data to obtain corresponding fourth first-order hyperspectral data; performing a second-order derivative processing on each fourth hyperspectral data to obtain corresponding fourth second-order hyperspectral data; performing a third-order derivative processing on each fourth hyperspectral data to obtain corresponding fourth third-order hyperspectral data;

[0108] Step S3B, performing correlation analysis on the plurality of fourth hyperspectral data and the corresponding measured values ​​of water quality parameters, and generating a corresponding water quality parameter inversion model; performing correlation analysis on the plurality of fourth pre-processed hyperspectral data and the corresponding measured values ​​of water quality parameters, and generating a corresponding water quality parameter inversion model; specifically:

[0109] Performing correlation analysis on the plurality of fourth hyperspectral data and the corresponding measured values ​​of water quality parameters, and performing a one-dimensional linear fitting on the single band with the highest correlation among the plurality of fourth hyperspectral data and the corresponding measured values ​​of water quality parameters to obtain a corresponding water quality parameter inversion model;

[0110] Correlation analysis is performed on multiple fourth first-order hyperspectral data, multiple fourth second-order hyperspectral data and multiple fourth third-order hyperspectral data with the actual measured values ​​of water quality parameters, and a single band with the highest correlation in the fourth first-order hyperspectral data, a single band with the highest correlation in the fourth second-order hyperspectral data and a single band with the highest correlation in the fourth third-order hyperspectral data are respectively subjected to one-dimensional linear fitting with the actual measured values ​​of water quality parameters to obtain corresponding water quality parameter inversion models; and a band ratio with the highest correlation in the fourth first-order hyperspectral data, a band ratio with the highest correlation in the fourth second-order hyperspectral data and a band ratio with the highest correlation in the fourth third-order hyperspectral data are respectively subjected to one-dimensional linear fitting with the actual measured values ​​of water quality parameters to obtain corresponding water quality parameter inversion models; similarly, after step S3B, at least seven water quality parameter inversion models can be obtained.

[0111] Step S3C, selecting a water quality parameter inversion model whose accuracy reaches a preset condition from each generated water quality parameter inversion model as a target water quality parameter inversion model.

[0112] Similarly, the water quality parameter inversion model selected in step S3C with accuracy meeting the preset conditions includes:

[0113] For each generated water quality parameter inversion model, calculate the corresponding determination coefficient R 2 , Where n is the number of hyperspectral data collected in step S3; t i is the measured value of water quality parameters corresponding to each hyperspectral data; i is the water quality parameter inversion value of each hyperspectral data calculated using the water quality parameter inversion model; is the average value of the measured values ​​of water quality parameters corresponding to all hyperspectral data; the R 2 The water quality parameter inversion model exceeding 0.65 is used as the target water quality parameter inversion model for monitoring the monitored lakes.

[0114] After obtaining the target water quality parameter inversion model, the target water quality parameter inversion model can be optimized again to select the excellent target water quality parameter inversion model, specifically:

[0115] The root mean square error RMSE and / or average relative error RE of each screened water quality parameter inversion model is calculated, and the water quality parameter inversion model whose root mean square error RMSE and / or average relative error RE exceeds the corresponding preset threshold is removed from the target water quality parameter inversion model; the preset threshold of the average relative error RE may be 20%; the water quality parameter inversion model whose root mean square error RMSE and / or average relative error RE exceeds the corresponding preset threshold is considered to be a less excellent model among the target water quality parameter inversion models; wherein:

[0116] Root mean square error

[0117] Mean relative error

[0118] Where n is the number of hyperspectral data collected in step S3; t i is the measured value of water quality parameters corresponding to each hyperspectral data; iThe water quality parameter inversion value of each hyperspectral data is calculated using the water quality parameter inversion model; after removing the water quality parameter inversion models whose root mean square error RMSE and / or average relative error RE exceeds the corresponding preset threshold value in the target water quality parameter inversion model, the target water quality parameter inversion model is further optimized, and the selected excellent target water quality parameter inversion model is used for subsequent water quality monitoring; these optimized models that meet the quality control requirements can ensure the accuracy of water quality monitoring because the calculation precision and accuracy are sufficiently guaranteed.

[0119] The derivation work of the above steps can be completed through ENVI5.6 remote sensing software.

[0120] The multiple first hyperspectral data and the measured values ​​of water quality parameters of multiple water samples obtained by step S1, and the multiple third hyperspectral data and the corresponding measured values ​​of water quality parameters obtained by step S3 are saved as a parent database. When different hyperspectral satellites are used to monitor water quality parameters, it is no longer necessary to repeatedly collect hyperspectral data or water samples at the collection point through the ASD portable object spectrometer. For each hyperspectral satellite, the data required for its modeling are all from the data collected by the ASD portable object spectrometer. That is to say, after collecting hyperspectral data in a lake, when the satellite to be used changes, it is only necessary to resample the spectral data collected by the ASD portable object spectrometer according to the corresponding resolution of the satellite, and then repeat the subsequent modeling process; in addition, after obtaining multiple first hyperspectral data and the measured values ​​of water quality parameters of multiple water samples in a lake through step S1, when the lake to be monitored is changed, it is not necessary to collect water samples in the new lake, only hyperspectral data needs to be collected, and the corresponding measured values ​​of water quality parameters can be taken as the inversion values ​​calculated by the universal water quality parameter inversion model, so the present invention can greatly improve the monitoring efficiency.

[0121] Step S4, acquiring a hyperspectral image of the monitored lake through the hyperspectral satellite, and calculating the water quality parameter of each pixel in the hyperspectral image using a corresponding target water quality parameter inversion model.

[0122] Specifically, the hyperspectral image obtained in step S4 can be a local hyperspectral image of the monitored lake obtained by a hyperspectral satellite, or a complete local hyperspectral image of the monitored lake. After the hyperspectral image is obtained by the Resource-1 satellite (ZY102D) or the Gaofen-5 satellite (GF5), the obtained hyperspectral image needs to be preprocessed by geometric correction, radiation correction, and atmospheric correction; and only the water body part in the hyperspectral image is extracted for water quality monitoring.

[0123] Step S5, calculate the average value of the corresponding water quality parameter of each pixel according to the water quality parameter of each pixel calculated by each water quality parameter inversion model in the target water quality parameter inversion model; take the average value of the water quality parameter of each pixel as the inversion value of the water quality parameter of the pixel, thereby obtaining the overall water quality parameter situation of the monitored lake.

[0124] Specifically, each pixel in the hyperspectral image has an independent hyperspectral curve, so the hyperspectral curve corresponding to each pixel in the hyperspectral image can be substituted into the target water quality parameter inversion model to invert the water quality parameters of each pixel. By combining the water quality parameters of each pixel in the hyperspectral image, the overall water quality parameters of the monitored lake in the hyperspectral image can be obtained. There may be multiple target water quality parameter inversion models. For a pixel, each target water quality parameter inversion model will invert a water quality parameter value, and the average value of multiple water quality parameters obtained for each pixel will be used as the final water quality parameter inversion value of the pixel. When calculating the average value of the corresponding water quality parameter of each pixel, obvious outliers can also be eliminated, such as values ​​that are obviously too large or too small compared with past water quality parameter monitoring data, so as to improve calculation efficiency. Through ENVI5.6 remote sensing software, each independent hyperspectral curve can be subjected to first-order, second-order, and third-order derivative processing. When the hyperspectral curve is substituted into the target water quality parameter inversion model to invert the water quality parameters, the optimal band or the absolute value of the optimal band ratio of each hyperspectral curve is obtained through the corresponding plug-in of ENVI5.6 remote sensing software.

[0125] Furthermore, after obtaining the inverted value of the water quality parameter of each pixel, each pixel can be marked with a color corresponding to the interval range of the inverted value of the water quality parameter of each pixel, so that the water quality parameter monitoring results of the lake can be displayed more intuitively. For example, when calculating the total nitrogen content of each pixel, when the calculation result of the total nitrogen content of the pixel (in mg / L) is ∈[0, 0.2], the pixel can be recorded as Class I and marked as red; when the calculation result of the total nitrogen content of the pixel is ∈(0.2, 0.5], the pixel can be recorded as Class II and marked as yellow; when the calculation result of the total nitrogen content of the pixel (in mg / L) is ∈(0.5, 1], the pixel can be recorded as Class III and marked as red; and so on. Pixels with total nitrogen contents in other ranges can also be distinguished by different categories and colors. In this way, pixels with total nitrogen contents in different ranges can be distinguished, so that the distribution of total nitrogen content in the monitored lake can be intuitively reflected.

[0126] In addition, two days before or after or on the day when the hyperspectral satellite passes over the lake, that is, two days before or after or on the day when the hyperspectral image is obtained, a certain number of collection points can be selected in the monitored lake, and the corresponding hyperspectral data and water samples can be collected using the ASD portable ground feature spectrometer, and each water sample is analyzed to obtain the measured value of the water quality parameters of each water sample, and then these collection points are used as quality control points. The number of collection points this time can be one tenth of the collection points selected in step S1 or step S3, and the measured values ​​of the water quality parameters of these collection points are compared with the water quality parameter inversion values ​​of these collection points calculated by the target water quality parameter inversion model to prove the accuracy of the target water quality parameter inversion model in calculating the water quality parameter inversion value, thereby indicating that the accuracy of the water quality parameter monitoring of the monitored lake is guaranteed.

[0127] Since brackish water lakes can inhibit the growth of algae, there are relatively few phytoplankton on the surface of brackish water lakes, so the hyperspectral images of brackish water lakes obtained by hyperspectral satellites can better reflect the conditions of the water body itself. Therefore, through the water quality parameter monitoring method provided by the present invention, the water quality parameters of each pixel in the hyperspectral image can be inverted more accurately using the obtained water quality parameter inversion model, so that the overall water quality parameters of the monitored brackish water lake can be obtained more accurately. In other words, the water quality parameter monitoring method provided by the present invention is very suitable for brackish water lakes.

[0128] At the same time, in the water quality parameter monitoring method provided by the present invention, the generated target water quality parameter inversion model includes multiple models, and the water quality parameter inversion value of each pixel is taken as the average value calculated by multiple models in the target water quality parameter inversion model. That is to say, the water quality monitoring result of the water quality monitoring method provided by the present invention is obtained by integrating the calculation results of multiple models, which is more accurate than the existing monitoring method that only uses a single model to monitor water quality parameters.

[0129] In addition, through the water quality parameter monitoring method provided by the present invention, after obtaining multiple first hyperspectral data and measured values ​​of water quality parameters of multiple water samples in a lake, when changing the lake to be monitored, it is not necessary to collect water samples in the new lake, only hyperspectral data needs to be collected, and the corresponding measured values ​​of water quality parameters can be taken as the inversion values ​​calculated by the general water quality parameter inversion model, so the present invention can greatly improve the monitoring efficiency. When different hyperspectral satellites are used to monitor the water quality parameters of the same lake, there is no need to repeatedly collect hyperspectral data at the collection point through the ASD portable object spectrometer, that is, after collecting hyperspectral data in a lake, when the satellite to be used changes, it is only necessary to resample the spectral data collected by the ASD portable object spectrometer according to the corresponding resolution of the satellite, and then repeat the subsequent modeling process; compared with the existing monitoring method, which requires repeated sampling at the collection point for different hyperspectral satellites, the water quality parameter monitoring method provided by the present invention can greatly improve the efficiency and convenience of water quality monitoring.

[0130] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1 At least part of the steps may include multiple steps or multiple stages. These steps or stages are not necessarily performed at the same time, but can be performed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed in turn or alternately with other steps or at least part of the steps or stages in other steps.

[0131] Embodiment 2

[0132] In this embodiment, illustratively, the method provided in Example 1 is applied to monitor the total nitrogen content of Bosten Lake:

[0133] Step 1: Use the ASD portable ground spectrometer to collect corresponding hyperspectral data and water samples at 50 collection points in Lake Bosten, analyze the 50 water samples, and obtain the total nitrogen content of each water sample.

[0134] Step 2: Use the Derivatives of Transform in the Modify module of The Unscrambler software to preprocess the 50 resampled hyperspectral curves using the first-order derivative, second-order derivative, and third-order derivative to obtain 50 first-order spectra, second-order spectra, and third-order spectra, respectively, in preparation for correlation analysis.

[0135] Step 3, using the relevant double variable (B) in the analysis module of SPSS software to perform correlation analysis on the 50 hyperspectral images and the corresponding 50 total nitrogen contents, the best single band with the highest correlation was 798nm, the best single band reflectance was used as the independent variable, the total nitrogen content was used as the dependent variable, and the monitoring model obtained by one-dimensional linear fitting was: y1 = 0.0844*x1 + 0.0102. The corresponding R 2 =0.02.

[0136] Step 4, in MATALB, use the code to perform correlation analysis on all the band ratios QRVI (|B1 / B2|) and QNDVI (|(B1-B2) / (B1+B2)|) of the first-order spectrum, the second-order spectrum and the third-order spectrum with the corresponding 50 total nitrogen contents, and obtain the band ratio with the highest correlation through the size of the correlation coefficient. The band ratios with the highest correlation are the first-order spectrum QNDVI (|(589-752) / (589+752)|), the first-order spectrum QRVI (|640 / 797|), the second-order spectrum QNDVI (|(588-791) / (588+791)|), the second-order spectrum QRVI (|702 / 774|), the third-order spectrum QNDVI (|(470-865) / (470+865)|), and the third-order spectrum QRVI (|415 / 738|);

[0137] Then, with the band ratio as the independent variable and the total nitrogen content as the dependent variable, the monitoring models obtained by linear fitting are:

[0138] First-order spectrum QNDVI: y2 = -0.0027*x2 + 0.7095 Corresponding R 2 =0.82

[0139] First-order spectrum QRVI: y3 = -0.0028*x3 + 0.7119 Corresponding R 2 =0.87

[0140] Second order spectrum QNDVI: y4 = -0.0044*x4 + 0.7101 Corresponding R 2 =0.80

[0141] Second order spectrum QRVI: y5 = -0.0008*x5 + 0.7073 Corresponding R 2 =0.89

[0142] Third-order spectrum QNDVI: y6 = -0.0024*x6 + 0.7075 Corresponding R 2 =0.92

[0143] The third-order spectrum QRVI: y7 = -0.0025*x7 + 0.7112 The corresponding R2 =0.88

[0144] The accuracy of the above models is summarized in Table 1:

[0145] Table 1 Summary of model accuracy

[0146]

[0147]

[0148] Which satisfies R 2 The models with values ​​above 0.65, RMSE below 0.02, and RE below 20.00% are y3, y5, y6, and y7.

[0149] Step 5, using y3, y5, y6, and y7 as the preferred target detection model for monitoring the total nitrogen content of Bosten Lake, the total nitrogen content of the entire Bosten Lake can be obtained. When using the target detection model to calculate the total nitrogen content of each pixel in the hyperspectral image of Bosten Lake, when the calculation result of the total nitrogen content of the pixel (in mg / L) is ∈[0, 0.2], the pixel is recorded as Class I and the pixel is marked red; when the calculation result of the total nitrogen content of the pixel is ∈(0.2, 0.5], the pixel is recorded as Class II and the pixel is marked yellow; when the calculation result of the total nitrogen content of the pixel (in mg / L) is ∈(0.5, 1], the pixel is recorded as Class III and the pixel is marked red; the monitoring results of the total nitrogen content of Bosten Lake are as follows Figure 2 As shown, from Figure 2 It can be seen that the total nitrogen content in all areas of Lake Bosten is in the range of (0.5, 1].

[0150] Embodiment 3

[0151] In this embodiment, if Figure 3 As shown, a water quality parameter monitoring device suitable for brackish water lakes is provided, comprising:

[0152] The universal water quality parameter inversion model generation module 301 is used to collect corresponding first hyperspectral data and water samples at multiple collection points preset in the first brackish water lake, respectively, to obtain multiple first hyperspectral data and corresponding multiple water samples, and to analyze the water samples through water body experiments to obtain the measured value of the water quality parameters of each water sample; and to generate a universal water quality parameter inversion model according to the multiple first hyperspectral data and the measured values ​​of the water quality parameters of the corresponding water samples;

[0153] The first target water quality parameter inversion model generation module 302 is used for, when the monitored lake is a first brackish water lake, resampling each first hyperspectral data according to the resolution of the hyperspectral satellite based on the hyperspectral satellite to be used, to obtain a plurality of second hyperspectral data matching the resolution of the hyperspectral satellite; and generating a corresponding target water quality parameter inversion model according to the plurality of second hyperspectral data and the measured values ​​of the water quality parameters of the corresponding water samples;

[0154] The second target water quality parameter inversion model generation module 303 is used for, when the monitored lake is a second brackish water lake, respectively collecting corresponding third hyperspectral data at a plurality of collection points preset in the second brackish water lake to obtain a plurality of third hyperspectral data, and calculating the water quality parameter inversion value corresponding to each third hyperspectral data through the universal water quality parameter inversion model, and using the obtained water quality parameter inversion value corresponding to each third hyperspectral data as the measured water quality parameter value corresponding to each third hyperspectral data; based on the hyperspectral satellite to be used, resampling each third hyperspectral data according to the resolution of the hyperspectral satellite to obtain a plurality of fourth hyperspectral data matching the resolution of the hyperspectral satellite; and generating a corresponding target water quality parameter inversion model according to the plurality of fourth hyperspectral data and the measured water quality parameter values ​​of the corresponding water samples;

[0155] A hyperspectral image acquisition module 304 acquires a hyperspectral image of the monitored lake through the hyperspectral satellite, and calculates the water quality parameter of each pixel in the hyperspectral image using a corresponding target water quality parameter inversion model;

[0156] The inversion value determination module 305 calculates the average value of the corresponding water quality parameter of each pixel based on the water quality parameter of each pixel calculated by each water quality parameter inversion model in the target water quality parameter inversion model; the average value of the water quality parameter of each pixel is used as the inversion value of the water quality parameter of the pixel, thereby obtaining the overall water quality parameter situation of the monitored lake.

[0157] For the specific definition of a water quality parameter monitoring device suitable for brackish water lakes, please refer to the definition of a water quality parameter monitoring method suitable for brackish water lakes in the above embodiment 1, which will not be repeated here. Each module in the above-mentioned water quality parameter monitoring device suitable for brackish water lakes can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0158] The technical effects of this embodiment are consistent with those of the first embodiment and will not be described in detail here.

[0159] Embodiment 4

[0160] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 4 As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc. When the computer program is executed by the processor, the following steps are implemented:

[0161] Acquire a plurality of first hyperspectral data collected in a first brackish water lake and measured values ​​of water quality parameters of a plurality of water samples obtained by corresponding analysis; generate a universal water quality parameter inversion model according to the plurality of first hyperspectral data and the measured values ​​of water quality parameters of the corresponding water samples;

[0162] When the monitored lake is a first brackish water lake, based on the hyperspectral satellite to be used, each first hyperspectral data is resampled according to the resolution of the hyperspectral satellite to obtain a plurality of second hyperspectral data matching the resolution of the hyperspectral satellite; and based on the plurality of second hyperspectral data and the measured values ​​of the water quality parameters of the corresponding water samples, a corresponding target water quality parameter inversion model is generated;

[0163] When the monitored lake is a second brackish water lake, the third hyperspectral data collected in the second brackish water lake is obtained, and the water quality parameter inversion value corresponding to each third hyperspectral data is calculated by the universal water quality parameter inversion model, and the obtained water quality parameter inversion value corresponding to each third hyperspectral data is used as the water quality parameter measured value corresponding to each third hyperspectral data; based on the hyperspectral satellite to be used, each third hyperspectral data is resampled according to the resolution of the hyperspectral satellite to obtain a plurality of fourth hyperspectral data matching the resolution of the hyperspectral satellite; and a corresponding target water quality parameter inversion model is generated according to the plurality of fourth hyperspectral data and the measured values ​​of the water quality parameters of the corresponding water samples;

[0164] Acquire a hyperspectral image of the monitored lake through the hyperspectral satellite, and calculate the water quality parameter of each pixel in the hyperspectral image using a corresponding target water quality parameter inversion model;

[0165] According to the water quality parameters of each pixel calculated by each water quality parameter inversion model in the target water quality parameter inversion model, the average value of the corresponding water quality parameters of each pixel is calculated; the average value of the water quality parameters of each pixel is used as the inversion value of the water quality parameters of the pixel, thereby obtaining the overall water quality parameters of the monitored lake.

[0166] Those skilled in the art will understand that Figure 4 The structure shown in is merely a block diagram of a portion of the structure related to the present application solution and does not constitute a limitation on the computer device to which the present application solution is applied.

[0167] The technical effects of this embodiment are consistent with those of the first embodiment and will not be described in detail here.

[0168] A person skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory.

[0169] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0170] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A method for monitoring water quality parameters of brackish water lakes, characterized in that: include: Step S1, respectively collecting corresponding first hyperspectral data and water samples at a plurality of collection points preset in a first brackish water lake, obtaining a plurality of first hyperspectral data and a corresponding plurality of water samples, and analyzing the water samples through a water body experiment to obtain a measured value of a water quality parameter of each water sample; and generating a universal water quality parameter inversion model according to the plurality of first hyperspectral data and the measured values ​​of the water quality parameters of the corresponding water samples; Step S2, when the monitored lake is a first brackish water lake, based on the hyperspectral satellite to be used, each first hyperspectral data is resampled according to the resolution of the hyperspectral satellite to obtain a plurality of second hyperspectral data matching the resolution of the hyperspectral satellite; and generating a corresponding target water quality parameter inversion model according to the plurality of second hyperspectral data and the measured values ​​of water quality parameters of the corresponding water samples; Step S3, when the monitored lake is a second brackish water lake, corresponding third hyperspectral data are respectively collected at a plurality of collection points preset in the second brackish water lake to obtain a plurality of third hyperspectral data, and a water quality parameter inversion value corresponding to each third hyperspectral data is calculated by the universal water quality parameter inversion model, and the obtained water quality parameter inversion value corresponding to each third hyperspectral data is used as the measured value of the water quality parameter corresponding to each third hyperspectral data; based on the hyperspectral satellite to be used, each third hyperspectral data is resampled according to the resolution of the hyperspectral satellite to obtain a plurality of fourth hyperspectral data matching the resolution of the hyperspectral satellite; and generating a corresponding target water quality parameter inversion model according to the plurality of fourth hyperspectral data and the measured values ​​of water quality parameters of the corresponding water samples; Step S4, acquiring a hyperspectral image of the monitored lake through the hyperspectral satellite, and calculating the water quality parameter of each pixel in the hyperspectral image using a corresponding target water quality parameter inversion model; Step S5, calculate the average value of the corresponding water quality parameter of each pixel according to the water quality parameter of each pixel calculated by each water quality parameter inversion model in the target water quality parameter inversion model; take the average value of the water quality parameter of each pixel as the inversion value of the water quality parameter of the pixel, thereby obtaining the overall water quality parameter situation of the monitored lake.

2. The water quality parameter monitoring method according to claim 1, characterized in that: The step S1 of generating a universal water quality parameter inversion model according to the plurality of first hyperspectral data and the measured values ​​of water quality parameters of corresponding water samples specifically includes: Step S1A, preprocessing each first hyperspectral data to obtain a plurality of first preprocessed hyperspectral data; Step S1B, performing correlation analysis on the plurality of first hyperspectral data and the corresponding measured values ​​of water quality parameters, and generating a corresponding water quality parameter inversion model; performing correlation analysis on the plurality of first pre-processed hyperspectral data and the corresponding measured values ​​of water quality parameters, and generating a corresponding water quality parameter inversion model; Step S1C, selecting a water quality parameter inversion model whose accuracy meets a preset condition from each generated water quality parameter inversion model as a universal water quality parameter inversion model; The step S2 of generating a corresponding target water quality parameter inversion model according to the plurality of second hyperspectral data and the measured values ​​of water quality parameters of the corresponding water samples specifically includes: Step S2A, preprocessing each second hyperspectral data to obtain a plurality of second preprocessed hyperspectral data; Step S2B, performing correlation analysis on the plurality of second hyperspectral data and the corresponding measured values ​​of water quality parameters, and generating a corresponding water quality parameter inversion model; performing correlation analysis on the plurality of second pre-processed hyperspectral data and the corresponding measured values ​​of water quality parameters, and generating a corresponding water quality parameter inversion model; Step S2C, selecting a water quality parameter inversion model whose accuracy meets a preset condition from each generated water quality parameter inversion model as a target water quality parameter inversion model; The step S3 of generating a corresponding target water quality parameter inversion model according to the plurality of fourth hyperspectral data and the measured values ​​of water quality parameters of the corresponding water samples specifically includes: Step S3A, preprocessing each fourth hyperspectral data to obtain a plurality of fourth preprocessed hyperspectral data; Step S3B, performing correlation analysis on the plurality of fourth hyperspectral data and the corresponding measured values ​​of water quality parameters, and generating a corresponding water quality parameter inversion model; performing correlation analysis on the plurality of fourth pre-processed hyperspectral data and the corresponding measured values ​​of water quality parameters, and generating a corresponding water quality parameter inversion model; Step S3C, selecting a water quality parameter inversion model whose accuracy meets a preset condition from each generated water quality parameter inversion model as a target water quality parameter inversion model.

3. The water quality parameter monitoring method according to claim 2, characterized in that: The step S1A specifically includes: Performing a first-order derivative processing on each first hyperspectral data to obtain corresponding first first-order hyperspectral data; performing a second-order derivative processing on each first hyperspectral data to obtain corresponding first second-order hyperspectral data; performing a third-order derivative processing on each first hyperspectral data to obtain corresponding first third-order hyperspectral data; The step S2A specifically includes: Performing a first-order derivative process on each second hyperspectral data to obtain corresponding second first-order hyperspectral data; performing a second-order derivative process on each second hyperspectral data to obtain corresponding second second-order hyperspectral data; performing a third-order derivative process on each second hyperspectral data to obtain corresponding second third-order hyperspectral data; The step S3A specifically includes: Perform a first-order derivative processing on each fourth hyperspectral data to obtain the corresponding fourth first-order hyperspectral data; perform a second-order derivative processing on each fourth hyperspectral data to obtain the corresponding fourth second-order hyperspectral data; perform a third-order derivative processing on each fourth hyperspectral data to obtain the corresponding fourth third-order hyperspectral data.

4. The water quality parameter monitoring method according to claim 3, characterized in that: The step S1B specifically includes: Performing correlation analysis on the plurality of first hyperspectral data and the corresponding measured values ​​of water quality parameters, and performing a one-dimensional linear fitting on the single band with the highest correlation among the plurality of first hyperspectral data and the corresponding measured values ​​of water quality parameters to obtain a corresponding water quality parameter inversion model; Correlation analysis is performed on a plurality of first-order hyperspectral data, a plurality of first-second-order hyperspectral data, and a plurality of first-third-order hyperspectral data with the measured values ​​of water quality parameters, and a single band with the highest correlation in the first-order hyperspectral data, a single band with the highest correlation in the first-second-order hyperspectral data, and a single band with the highest correlation in the first-third-order hyperspectral data are respectively subjected to one-dimensional linear fitting with the measured values ​​of water quality parameters to obtain a corresponding water quality parameter inversion model; and a band ratio with the highest correlation in the first-order hyperspectral data, a band ratio with the highest correlation in the first-second-order hyperspectral data, and a band ratio with the highest correlation in the first-third-order hyperspectral data are respectively subjected to one-dimensional linear fitting with the measured values ​​of water quality parameters to obtain a corresponding water quality parameter inversion model; The step S2B specifically includes: Performing correlation analysis on the plurality of second hyperspectral data and the corresponding measured values ​​of water quality parameters, and performing a one-dimensional linear fitting on the single band with the highest correlation among the plurality of second hyperspectral data and the corresponding measured values ​​of water quality parameters to obtain a corresponding water quality parameter inversion model; Correlation analysis is performed on multiple second-order hyperspectral data, multiple second-order hyperspectral data, and multiple second-order hyperspectral data with the measured values ​​of water quality parameters, and a single band with the highest correlation in the second-order hyperspectral data, a single band with the highest correlation in the second-order hyperspectral data, and a single band with the highest correlation in the second-order hyperspectral data are respectively fitted with the measured values ​​of water quality parameters to obtain corresponding water quality parameter inversion models; and a band ratio with the highest correlation in the second-order hyperspectral data, a band ratio with the highest correlation in the second-order hyperspectral data, and a band ratio with the highest correlation in the second-order hyperspectral data are respectively fitted with the measured values ​​of water quality parameters to obtain corresponding water quality parameter inversion models; The step S3B specifically includes: Performing correlation analysis on the plurality of fourth hyperspectral data and the corresponding measured values ​​of water quality parameters, and performing a one-dimensional linear fitting on the single band with the highest correlation among the plurality of fourth hyperspectral data and the corresponding measured values ​​of water quality parameters to obtain a corresponding water quality parameter inversion model; Correlation analysis is performed on multiple fourth first-order hyperspectral data, multiple fourth second-order hyperspectral data and multiple fourth third-order hyperspectral data with the actual measured values ​​of water quality parameters, and a single band with the highest correlation in the fourth first-order hyperspectral data, a single band with the highest correlation in the fourth second-order hyperspectral data and a single band with the highest correlation in the fourth third-order hyperspectral data are respectively subjected to a one-dimensional linear fit with the actual measured values ​​of water quality parameters to obtain a corresponding water quality parameter inversion model; and a band ratio with the highest correlation in the fourth first-order hyperspectral data, a band ratio with the highest correlation in the fourth second-order hyperspectral data and a band ratio with the highest correlation in the fourth third-order hyperspectral data are respectively subjected to a one-dimensional linear fit with the actual measured values ​​of water quality parameters to obtain a corresponding water quality parameter inversion model.

5. The water quality parameter monitoring method according to claim 2, characterized in that: The water quality parameter inversion models that have been screened out and have reached the preset accuracy are: Calculate the determination coefficient R of each water quality parameter inversion model generated 2 , select the determination coefficient R 2 Inversion model of target water quality parameters exceeding preset determination coefficient thresholds.

6. The water quality parameter monitoring method according to claim 5, characterized in that: The coefficient of determination Where n is the number of hyperspectral data collected, t i is the measured value of water quality parameter corresponding to each hyperspectral data, y i is the water quality parameter inversion value of each hyperspectral data calculated using the water quality parameter inversion model. It is the average value of the measured values ​​of water quality parameters corresponding to all hyperspectral data; the preset determination coefficient threshold is 0.

65.

7. The water quality parameter monitoring method according to claim 2, characterized in that: Calculating the water quality parameter inversion value corresponding to each third hyperspectral data by the universal water quality parameter inversion model in step S3 specifically includes: Calculating the water quality parameters of each third hyperspectral data using a universal water quality parameter inversion model; According to the water quality parameters of each third hyperspectral data calculated by each water quality parameter inversion model in the universal water quality parameter inversion model, the average value of the corresponding water quality parameters of each third hyperspectral data is calculated; and the average value of the water quality parameters of each third hyperspectral data is used as the inversion value of the water quality parameters of the third hyperspectral data.

8. The water quality parameter monitoring method according to any one of claims 2 to 7, characterized in that: The steps S2C and S3C also include: Calculate the root mean square error RMSE and / or average relative error RE of each screened water quality parameter inversion model, and remove the water quality parameter inversion model whose root mean square error RMSE and / or average relative error RE exceeds the corresponding preset threshold value from the target water quality parameter inversion model; wherein: Root mean square error Mean relative error Where n is the number of hyperspectral data collected, t i is the measured value of water quality parameter corresponding to each hyperspectral data, y i It is the water quality parameter inversion value of each hyperspectral data calculated by using the water quality parameter inversion model.

9. The water quality parameter monitoring method according to any one of claims 1 to 7, characterized in that: The water quality parameter is at least one of the permanganate index, chemical oxygen demand, ammonia nitrogen content, total phosphorus content, total nitrogen content, suspended matter content and chlorophyll a content; the hyperspectral satellite is the Resources-1 satellite or the Gaofen-5 satellite; the hyperspectral data are collected through a portable ground spectrometer; and the correlation analysis is performed using the Wilson correlation coefficient method.

10. A water quality parameter monitoring device suitable for brackish water lakes, characterized in that: include: A general water quality parameter inversion model generation module is used to collect corresponding first hyperspectral data and water samples at multiple preset collection points in the first brackish water lake, obtain multiple first hyperspectral data and corresponding multiple water samples, and analyze the water samples through water body experiments to obtain measured values ​​of water quality parameters of each water sample; and generating a universal water quality parameter inversion model based on the plurality of first hyperspectral data and measured values ​​of water quality parameters of corresponding water samples; A first target water quality parameter inversion model generation module is used for, when the monitored lake is a first brackish water lake, resampling each first hyperspectral data according to the resolution of the hyperspectral satellite based on the hyperspectral satellite to be used, to obtain a plurality of second hyperspectral data matching the resolution of the hyperspectral satellite; and generating a corresponding target water quality parameter inversion model according to the plurality of second hyperspectral data and the measured values ​​of water quality parameters of the corresponding water samples; The second target water quality parameter inversion model generation module is used for, when the monitored lake is a second brackish water lake, respectively collecting corresponding third hyperspectral data at a plurality of collection points preset in the second brackish water lake to obtain a plurality of third hyperspectral data, and calculating the water quality parameter inversion value corresponding to each third hyperspectral data through the universal water quality parameter inversion model, and using the obtained water quality parameter inversion value corresponding to each third hyperspectral data as the measured value of the water quality parameter corresponding to each third hyperspectral data; based on the hyperspectral satellite to be used, resampling each third hyperspectral data according to the resolution of the hyperspectral satellite to obtain a plurality of fourth hyperspectral data matching the resolution of the hyperspectral satellite; and generating a corresponding target water quality parameter inversion model according to the plurality of fourth hyperspectral data and the measured values ​​of water quality parameters of the corresponding water samples; A hyperspectral image acquisition module, which acquires a hyperspectral image of the monitored lake through the hyperspectral satellite, and calculates the water quality parameter of each pixel in the hyperspectral image using a corresponding target water quality parameter inversion model; The inversion value determination module calculates the average value of the corresponding water quality parameters of each pixel based on the water quality parameters of each pixel calculated by each water quality parameter inversion model in the target water quality parameter inversion model; the average value of the water quality parameters of each pixel is used as the inversion value of the water quality parameters of the pixel, thereby obtaining the overall water quality parameters of the monitored lake.

Images