Method and system for inverting daily water quality through space-ground integrated space-time statistical modeling

Abstract

The invention discloses a method and a system for inverting daily water quality through space-ground integrated space-time statistical modeling, daily water quality parameters are estimated by utilizing satellite film inversion results existing in a time sequence, firstly, satellite images (sky) and water surface observation data (ground) of a target area are obtained in a space-ground integrated manner, and the satellite images (sky) and the water surface observation data (ground) are calculated on the basis of water body indexes; modeling by utilizing an ST-SAM model to automatically extract a water area; then associating satellite image data with water surface observation data in the extracted water area through spatio-temporal information, and establishing a water environment water quality inversion model and result of the target area under the satellite data condition by using partial least squares regression; and finally, through time sequence interpolation, correction is carried out by using space-time geographically weighted regression so as to obtain a daily water environment water quality inversion result of the target area. The method can obtain the water quality inversion result of the water environment at high frequency and high precision, has the advantages of real-time monitoring, high frequency, low cost and the like, and has obvious superiority in the aspect of efficient and high-precision obtaining of the daily inversion result.

Description

Technical Field

[0001] The present invention belongs to the field of spatiotemporal geographic information modeling and environmental protection technology, and relates to a daily water quality inversion method and system, and specifically to a method and system for inverting daily water quality through space-time statistical modeling of integrated earth and sky. Background Art

[0002] Environmental remote sensing technology has been widely used in natural resource exploration, disaster prediction, urban environmental monitoring, water environment monitoring, etc., and has important value in the fields of civil, military, environmental protection, etc. Good water quality and water environment are an indispensable part of maintaining sustainable life and production in human society. They not only provide protection for human health, provide raw materials for industrial production, but also lay the foundation for the good operation of the entire ecosystem.

[0003] For the purpose of research and real-time monitoring and protection, obtaining daily water quality results has become a central need. The higher the frequency of monitoring, the more timely water quality and environmental problems can be discovered and resolved. However, due to (1) orbital repetition period and number of satellites (2) field of view width limitations (3) low satellite coverage frequency (4) satellite cloud cover, satellite images cannot be obtained frequently, and daily water quality inversion results cannot be directly obtained from satellite images and ground station data.

[0004] Although there are solutions to use drones to obtain water images to achieve high-frequency (i.e., daily) water quality inversion, drones still have the following problems:

[0005] (1) According to the law, drones need to apply for airspace from the military. It is very difficult to apply for airspace for shooting images frequently, such as every day. There are many procedures and restrictions, which is not realistic.

[0006] (2) Drones are prone to disappearance, which poses a safety hazard. First, drones are prone to disappearance when flying at high altitudes due to factors such as weak GPS signals and complex outdoor environments, and data acquisition is unstable. Second, accidental crashes of drones may threaten personal safety, cause property damage, and may also threaten the lives of wild animals. Accidental fires during crashes may cause forest fires.

[0007] (3) The existing drone technology has weak endurance. When used for monitoring large areas of water, the drones require a large number of aerial survey images, a large amount of data, and a long processing time, which is time-consuming and labor-intensive.

[0008] Considering the shortcomings of drone image acquisition, it is particularly important to achieve high-frequency water quality and water environment inversion, such as daily, without using drones. Summary of the invention

[0009] In order to solve the above technical problems, the present invention provides a method and system for inverting daily water quality through space-time statistical modeling that integrates the earth and the sky. It is a water quality inversion solution that can obtain water at high frequency, monitor in real time, and consume little time and money.

[0010] The technical solution adopted by the method of the present invention is: a method for inverting daily water quality by space-time statistical modeling of integrated earth and sky, comprising the following steps:

[0011] Step 1: Obtain satellite image data and water surface observation data of the target area, extract the water area in the satellite image data, and obtain spectral data of the water surface observation data;

[0012] Step 2: Associate satellite image data with water surface observation data in the extracted water area through temporal and spatial information, and use partial least squares regression to establish the water environment and water quality inversion model and results of the target area under the condition of satellite data;

[0013] Step 3: Interpolate the time series and use spatiotemporal geographically weighted regression for correction to obtain the daily water quality inversion results of the target area.

[0014] Preferably, in step 1, the water area in the satellite image data is automatically extracted by using the ST-SAM model; the specific implementation includes the following sub-steps:

[0015] Step 1.1: Obtain remote sensing images of the target area as image datasets;

[0016] Step 1.2: Preprocess the image, adjust the image size, normalize the pixel value, and obtain the processed data set;

[0017] Step 1.3: Use the processed data set to train the ST-SAM model, determine the loss function and optimizer, and adjust the hyperparameters;

[0018] The ST-SAM model takes multi-temporal remote sensing images and prompt information as input. First, the spatial features of the image are extracted through a lightweight image encoder, and the multimodal features provided by CLIP are used to align the image and text information. Then, the spatiotemporal dependency is captured, and the optimized SAM module generates a segmentation mask through effective prompt information.

[0019] The lightweight image encoder is composed of a plurality of sequentially connected Swing Transformer models and adapter combinations, and the model is adapted to the changes of different sensor images through a domain adaptation strategy;

[0020] Step 1.4: Evaluate the model extraction results using precision, recall, and F1 score metrics;

[0021] Step 1.5: Feedback the network based on the evaluation results and adjust the model framework, retrain until the water area extraction effect that meets the threshold requirements is achieved, and then use the trained model to extract the water area in the target area.

[0022] Preferably, the specific implementation of step 2 includes the following sub-steps:

[0023] Step 2.1: Use grey relational analysis GRA to calculate the correlation coefficient between the band and the water quality parameter, and take the m bands with the largest correlation coefficient as the final independent variable of the water quality index of the target area water environment; where m is the preset value;

[0024] Step 2.2: Use partial least squares regression (PLSR) to construct a water quality inversion model for the water quality index WQI and the determined m bands;

[0025] (WQI) I =f(b r , b s …b z );

[0026] Where, (WQI) I represents water quality parameters under satellite data conditions, I represents image, b r , b s …b z Indicates the determined m bands;

[0027] Step 2.3: training the water quality inversion model to obtain a trained water quality inversion model;

[0028] During training, select the coefficient of determination R 2 The root mean square error (RMSE) and root mean square error (RMSE) are used as two evaluation criteria to select the optimal water quality inversion model, adjust the water quality inversion model parameters, and optimize the partial least squares regression (PLSR) to obtain a trained water environment water quality inversion model for the target area.

[0029] Preferably, the specific implementation of step 3 includes the following sub-steps:

[0030] Step 3.1: Obtain the inversion results of multiple phases of satellite image data, and use the Lagrangian interpolation method of time series interpolation to estimate the water quality parameter prediction results (WQI) of unknown phases based on the inversion results of the phases. e (t);

[0031]

[0032] Where, (WQI) Ii t i Water quality parameters at time t1…t nis the time series, t i , t j is the time when the water quality parameters are known;

[0033] Step 3.2: Use spatiotemporal geographical weighted regression and the field measured water quality data of the day to calibrate the prediction results;

[0034] (WQI) dg =β0(u i , v i , t i )+β1(u i v i , t i )(WQI) e +ε i ;

[0035] Among them, β0(u i ,v i ,t i ), β1(u i ,v i ,t i ) is the coefficient to be determined, is random noise, i=1…n; is a matrix composed of the final results of the water quality parameters inverted daily in the target area, where d represents daily, g represents the field, (u i , v i ) represents a coordinate point, t i Indicates a certain moment, (u i , v i , t i ) is the space-time position at a certain moment and a certain spatial coordinate; A matrix of daily water quality parameters for interpolation prediction;

[0036]

[0037] make

[0038] Calculating space-time distance

[0039]

[0040] In the formula, λ and μ are coefficients used to balance the influence between space and time. The optimal λ and μ are selected according to the fitting effect of the model under different parameters.

[0041] Calculate the space-time distance decay function α ij ,

[0042]

[0043] in h ST is the parameter of space-time bandwidth, h S is the parameter of spatial bandwidth, h T is the parameter of time bandwidth; i = 1...n; j = 1...n;

[0044] Calculate the spatiotemporal weight matrix W(u i , v i , t i ),

[0045] W(u i , v i , t i )=diag(α i1 , α i2 , …, α in );

[0046] Calculate the coefficients of the formula.

[0047]

[0048] in, is the coefficient calculated by the formula, including β0(u i ,v i ,t i ), β1(u i ,v i ,t i ) These two coefficients are the intercept and slope in the regression equation; β0(u i ,v i ,t i ) is a constant term, β1(u i ,v i ,t i ) are coefficients related to the interpolated water quality parameters;

[0049] The coefficients obtained from nearby field observation points are used to correct the water environment and water quality parameter results of the surrounding areas, thus obtaining the daily water environment and water quality inversion results of the target area.

[0050] The technical solution adopted by the system of the present invention is: a system for inverting daily water quality by space-time statistical modeling integrated with the sky and the earth, including the following modules:

[0051] A data acquisition module is used to obtain satellite image data and water surface observation data of the target area, extract the water area in the satellite image data, and obtain spectral data of the water surface observation data;

[0052] The water quality inversion model building module is used to associate satellite image data with water surface observation data in the extracted water area through temporal and spatial information, and use partial least squares regression to establish the water environment water quality inversion model and results of the target area under the condition of satellite data;

[0053] The water quality inversion module is used to obtain the daily water quality inversion results of the target area through time series interpolation and correction using spatiotemporal geographical weighted regression.

[0054] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for inverting daily water quality by using space-ground integrated spatiotemporal statistical modeling is implemented.

[0055] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for inverting daily water quality by using space-ground integrated spatiotemporal statistical modeling is implemented.

[0056] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements the method for inverting daily water quality through space-time statistical modeling of integrated earth and sky.

[0057] Compared with the prior art, the beneficial effects of the present invention include:

[0058] (1) The present invention improves the accuracy and correctness of the results. The accuracy of manually clipping water areas with vector boundaries is low. SwinTransformer modeling is used to automatically extract water areas, reducing data errors.

[0059] (2) The present invention avoids the use of drones and does not require the application of airspace, which is convenient and time-saving, consumes less funds, and is safer;

[0060] (3) The present invention provides real-time monitoring, is efficient and time-saving, and obtains daily inversion results through spatiotemporal geographic weighted regression based on existing satellite images of non-target phases in the absence of satellite image data. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] The technical solution of the present invention is further described below using embodiments and specific implementation methods. In addition, in the process of describing the technical solution, some drawings are also used. For those skilled in the art, other drawings and the intention of the present invention can also be obtained based on these drawings without paying creative work.

[0062] Figure 1 is a flow chart of a method according to an embodiment of the present invention;

[0063] Figure 2A schematic diagram of a method according to an embodiment of the present invention;

[0064] Figure 3 It is a network structure diagram of the ST-SAM model in an embodiment of the present invention;

[0065] Figure 4 A schematic diagram of spatiotemporal statistical modeling and spatiotemporal weighting in an embodiment of the present invention;

[0066] Figure 5 This is a rendering of a water body extraction model in an embodiment of the present invention;

[0067] Figure 6 This is a diagram of the water area extraction process in an embodiment of the present invention;

[0068] Figure 7 This is a water area extraction result diagram in an embodiment of the present invention;

[0069] Figure 8 This is a schematic diagram of the preliminary results of water quality inversion of TN value at the 20200902 time phase in an embodiment of the present invention;

[0070] Fig. 9 This is a schematic diagram of the preliminary results of water quality inversion of TN value at the 20200918 time phase in an embodiment of the present invention;

[0071] Fig.10 This is a schematic diagram of the water quality inversion results of daily TN values ​​from 20200903 to 20200910 in an embodiment of the present invention. DETAILED DESCRIPTION

[0072] In order to facilitate ordinary technicians in the field to understand and implement the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the implementation examples described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0073] Please see Figure 1 and Figure 2 The present embodiment provides a method for inverting daily water quality by space-time statistical modeling based on integrated earth-sky space-time statistical modeling, comprising the following steps:

[0074] Step 1: Obtain satellite image data (sky) and water surface observation data (ground) of the target area, extract the water area in the satellite image data, and obtain the spectral data of the water surface observation data;

[0075] In one embodiment, the image data of the target area is obtained from various satellites such as LANDSAT series satellites, MODIS series satellites, Sentinel series satellites, QuickBird remote sensing satellites, Planet remote sensing satellites, GeoEye remote sensing satellites, Gaofen series satellites, etc.; Figure 1Water surface observation data, including sampling test results and spectral data, are obtained through unmanned ships, buoy stations, underwater water quality spectrometers, etc.

[0076] In one embodiment, the ST-SAM model is used to automatically extract water areas from satellite image data. The specific implementation includes the following sub-steps:

[0077] Step 1.1: Obtain remote sensing images of the target area as image datasets;

[0078] Step 1.2: Preprocess the image, adjust the image size, normalize the pixel value, and obtain the processed data set;

[0079] Step 1.3: Use the processed data set to train the ST-SAM model, determine the loss function and optimizer, and adjust the hyperparameters;

[0080] Please see Figure 3 In one embodiment, the proposed ST-SAM model combines spatiotemporal modeling with efficient segmentation technology, and is improved on the basis of the Swing Transformer model. The input of the model includes multi-temporal remote sensing images and prompt information. First, the spatial features of the image are extracted by a lightweight image encoder, and the image and text information are aligned using the multimodal features provided by CLIP. The lightweight image encoder is composed of a combination of several sequentially connected Swing Transformer models and adapters, and the model is adapted to the changes of different sensor images through a domain adaptation strategy;

[0081] Next, the spatiotemporal dependencies are captured, and the optimized SAM module generates segmentation masks through effective prompts. The domain adaptation strategy enables the model to adapt to the changes in images from different sensors and improves cross-domain generalization capabilities. The overall design ensures a balance between the model's efficiency and segmentation accuracy, making it suitable for large-scale, high-resolution remote sensing image segmentation tasks.

[0082] Step 1.4, use indicators such as precision, recall, and F1 score to evaluate the model extraction results;

[0083] Step 1.5, based on the evaluation results, the network is fed back and the model framework is adjusted, and retrained until the water area extraction effect that meets the threshold requirements is achieved, and then the trained model is used to extract the water area in the target area.

[0084] Step 2: Associate satellite image data with water surface observation data in the extracted water area through temporal and spatial information, and use partial least squares regression to establish the water environment and water quality inversion model and results of the target area under the condition of satellite data;

[0085] In one embodiment, an inversion model is established using statistical methods, such as ordinary linear regression (OLR), stepwise regression (SR), ridge estimation (RE), partial least squares regression (PLSR), geographically weighted regression (GWR), etc.

[0086] The specific implementation of step 2 includes the following sub-steps:

[0087] Step 2.1: Use Grey Relation Analysis (GRA) to calculate the correlation coefficient between the band and the water quality parameter, and take the m bands with the largest correlation coefficient as the final independent variable of the water quality index of the target area water environment; where m is the preset value;

[0088] Water quality and water environment indicators (WQI) are shown in Table 1 below.

[0089] Table 1

[0090]

[0091] Step 2.2: Use partial least squares regression PLSR (Partial Least Squares Regression) to construct a water quality inversion model for the water quality index WQI and the determined m bands;

[0092] (WQI) I =f(b r , b s …b z );

[0093] Where, (WQI) I represents water quality parameters under satellite data conditions, I represents image, b r , b s …b z Indicates the determined m bands;

[0094] Step 2.3: training the water quality inversion model to obtain a trained water quality inversion model;

[0095] During training, select the coefficient of determination R 2The root mean square error (RMSE) and root mean square error (RMSE) are used as the two main evaluation criteria to select the optimal water quality inversion model, adjust the water quality inversion model parameters, and optimize the partial least squares regression (PLSR) to obtain the trained water environment water quality inversion model of the target area.

[0096] Step 3: Through time series interpolation and correction using Geographically and Temporally Weighted Regression (GTWR), the daily water quality inversion results of the target area are obtained.

[0097] Please see Figure 4 In one implementation, the specific implementation of step 3 includes the following sub-steps:

[0098] Step 3.1: Obtain the inversion results of multiple phases of satellite image data, and use the Lagrangian interpolation method of time series interpolation to estimate the water quality parameter prediction results (WQI) of unknown phases based on the inversion results of the phases. e (t);

[0099]

[0100] Where, (WQI) Ii t i Water quality parameters at time t1…t n is the time series, t i , t j is the time when the water quality parameters are known;

[0101] Step 3.2: Use spatiotemporal geographically weighted regression (GTWR) and the day's field measured water quality data to calibrate the predicted results;

[0102] (WQI) dg =β0(u i , v i , t i )+β1(u i , v i , t i )(WQI) e +ε i ;

[0103] Among them, β0(u i ,v i ,t i ), β1(u i ,v i ,t i ) is the coefficient to be determined, is random noise, i=1…n; is a matrix composed of the final results of the water quality parameters inverted daily in the target area, where d represents daily, g represents ground, and (u i , v i ) represents a coordinate point, t i Indicates a certain moment, (u i , v i , t i ) is the space-time position at a certain moment and a certain spatial coordinate; A matrix of daily water quality parameters for interpolation prediction;

[0104]

[0105] make

[0106] Calculating space-time distance

[0107]

[0108] In the formula, θ and μ are coefficients used to balance the influence between space and time. The optimal λ and μ are selected according to the fitting effect of the model under different parameters. In some cases, if the influence of time is greater, the weight of μ should be appropriately increased; on the contrary, if the spatial relationship is more important, the weight of λ should be higher. Generally speaking, these coefficients can be adjusted by cross-validation or trial and error, that is, the optimal λ and μ are selected according to the fitting effect of the model under different parameters.

[0109] Calculate the space-time distance decay function α ij ,

[0110]

[0111] in h ST is the parameter of space-time bandwidth, h S is the parameter of spatial bandwidth, h T is the parameter of time bandwidth; i = 1...n; j = 1...n;

[0112] Calculate the spatiotemporal weight matrix W(u i , v i , t i ),

[0113] W(u i , v i , t i )=diag(α i1 , α i2 , …, α in );

[0114] Calculate the coefficients of the formula.

[0115]

[0116] in, is the coefficient calculated by the formula, including β0(u i ,v i ,t i ), β1(u i ,v i ,t i ) These two coefficients are the intercept and slope in the regression equation; β0(u i ,v i ,t i ) is a constant term, β1(u i ,v i ,t i ) are coefficients related to the interpolated water quality parameters;

[0117] The coefficients obtained from nearby field observation points are used to correct the water environment and water quality parameter results of the surrounding areas, thus obtaining the daily water environment and water quality inversion results of the target area.

[0118] This embodiment also provides a system for inverting daily water quality by space-time statistical modeling based on integrated earth-ground space-time statistics, including the following modules:

[0119] A data acquisition module is used to obtain satellite image data and water surface observation data of the target area, extract the water area in the satellite image data, and obtain spectral data of the water surface observation data;

[0120] The water quality inversion model building module is used to associate satellite image data with water surface observation data in the extracted water area through temporal and spatial information, and use partial least squares regression to establish the water environment water quality inversion model and results of the target area under the condition of satellite data;

[0121] The water quality inversion module is used to obtain the daily water quality inversion results of the target area through time series interpolation and correction using spatiotemporal geographical weighted regression.

[0122] This embodiment also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for inverting daily water quality by space-time statistical modeling based on integrated earth-ground is implemented.

[0123] This embodiment also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the method for inverting daily water quality through space-time statistical modeling based on integrated earth and sky is implemented.

[0124] This embodiment also provides a computer program product, including a computer program, which, when executed by a processor, implements the method for inverting daily water quality through space-time statistical modeling based on integrated earth-sky integration.

[0125] The present invention is further described below through specific experiments.

[0126] The experimental verification process includes:

[0127] (1) Take the water area of ​​satellite image data and obtain the spectral data of water surface observation data. The effect diagram of the water body extraction model in this embodiment is shown in Figure 5 , please see the water extraction process Figure 6 ; For water area extraction results, see Figure 7 ;

[0128] Get the spectral data of water surface observation data in the water area:

[0129] (2) Taking the two time phases of 20200902 and 20200918 as examples, a water quality inversion model of the water environment in the target area with satellite data was established, and the water quality inversion results were obtained; see Figure 8 , which is the preliminary result of water quality inversion of TN value in 20200902 phase; please see Fig. 9 , which is the preliminary result of water quality inversion of TN value at 20200918;

[0130] (3) Based on two adjacent time nodes of the satellite return cycle, the daily inversion results are interpolated through time series, and the spatiotemporal geographic weighted regression is used for correction to obtain more accurate daily water environment and water quality inversion results for the target area. Taking the two phases of 20200902 and 20200918 with satellite images as examples, the inversion results of these two phases have been obtained in the previous step. Through the interpolation and correction model, the daily water environment and water quality inversion results between 20200902 and 20200918 can be finally obtained; see Fig.10 , which are the water quality inversion results of daily TN values ​​from 20200903 to 20200910.

[0131] Experiments have shown that the present invention can obtain water environment water quality inversion results with high frequency and high precision, has the advantages of real-time monitoring, high frequency and low cost, and has obvious superiority in obtaining daily inversion results with high efficiency and high precision.

[0132] It should be understood that the embodiments described above are part of the embodiments of the present invention, rather than all of the embodiments. In addition, the technical features in the various embodiments or single embodiments provided by the present invention can be combined with each other arbitrarily to form a feasible technical solution. Such combination is not restricted by the sequence of steps and / or the structural composition mode, but must be based on the ability of ordinary technicians in the field to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such combination of technical solutions does not exist and is not within the protection scope required by the present invention.

[0133] It should be understood that the above description of the preferred embodiment is relatively detailed and cannot be regarded as limiting the scope of patent protection of the present invention. Under the enlightenment of the present invention, ordinary technicians in this field can also make substitutions or modifications without departing from the scope of protection of the claims of the present invention, which all fall within the scope of protection of the present invention. The scope of protection requested for the present invention shall be based on the attached claims.

Claims

1. A method for inverting daily water quality by space-time statistical modeling based on integrated earth-sky integration, characterized in that: The following steps are involved: Step 1: Obtain satellite image data and water surface observation data of the target area, extract the water area in the satellite image data, and obtain spectral data of the water surface observation data; Step 2: Associate satellite image data with water surface observation data in the extracted water area through temporal and spatial information, and use partial least squares regression to establish the water environment and water quality inversion model and results of the target area under the condition of satellite data; Step 3: Interpolate the time series and use spatiotemporal geographically weighted regression for correction to obtain the daily water quality inversion results of the target area.

2. The method for inverting daily water quality by space-time statistical modeling based on integrated earth-sky information according to claim 1 is characterized by: In step 1, the ST-SAM model is used to automatically extract the water area in the satellite image data; the specific implementation includes the following sub-steps: Step 1.1: Obtain remote sensing images of the target area as image datasets; Step 1.2: Preprocess the image, adjust the image size, normalize the pixel value, and obtain the processed data set; Step 1.3: Use the processed data set to train the ST-SAM model, determine the loss function and optimizer, and adjust the hyperparameters; The ST-SAM model takes multi-temporal remote sensing images and prompt information as input. First, the spatial features of the image are extracted by a lightweight image encoder, and the multimodal features provided by CLIP are used to align the image and text information. Then, the spatiotemporal dependency is captured, and the optimized SAM module generates a segmentation mask through the prompt information. The lightweight image encoder is composed of a plurality of sequentially connected Swing Transformer models and adapter combinations, and the model is adapted to the changes of different sensor images through a domain adaptation strategy; Step 1.4: Evaluate the model extraction results using precision, recall, and F1 score metrics; Step 1.5: Feedback the network based on the evaluation results and adjust the model framework, retrain until the water area extraction effect that meets the threshold requirements is achieved, and then use the trained model to extract the water area in the target area.

3. The method for inverting daily water quality by space-time statistical modeling based on integrated earth-sky information according to claim 1 is characterized by: The specific implementation of step 2 includes the following sub-steps: Step 2.1: Use grey relational analysis GRA to calculate the correlation coefficient between the band and the water quality parameter, and take the m bands with the largest correlation coefficient as the final independent variable of the water quality index of the target area water environment; where m is the preset value; Step 2.2: Use partial least squares regression (PLSR) to construct a water quality inversion model for the water quality index WQI and the determined m bands; (WQI) I =f(b r ,b s …b z ); Where, (WQI) I represents water quality parameters under satellite data conditions, I represents image, b r , b s …b z Indicates the determined m bands; Step 2.3: training the water quality inversion model to obtain a trained water quality inversion model; During training, the determination coefficient R2 and the root mean square error RMSE (Root Mean Squared Error) are selected as two evaluation criteria to select the optimal water quality inversion model; the cumulative variance contribution rate is statistically calculated, and the principal components greater than a certain threshold (for example, 85%) are selected as independent variables, and the water quality inversion model parameters are adjusted to optimize the partial least squares regression PLSR, and finally the trained water environment water quality inversion model of the target area is obtained.

4. The method for inverting daily water quality by space-time statistical modeling based on integrated earth-sky information according to claim 1 is characterized by: The specific implementation of step 3 includes the following sub-steps: Step 3.1: Obtain the inversion results of multiple phases of satellite image data, and use the Lagrangian interpolation method of time series interpolation to estimate the water quality parameter prediction results (WQI) of unknown phases based on the inversion results of the phases. e (t); Where, (WQI) Ii t i Water quality parameters at time t1…t n is the time series, t i , t j is the time when the water quality parameters are known; Step 3.2: Use spatiotemporal geographical weighted regression and the field measured water quality data of the day to calibrate the prediction results; (WQI) dg =β0(u i v i ,t i )+β1(u i v i ,t i )(WQI) e +ε i 4 Among them, β0(u i ,v i ,t i ), β1(u i ,v i ,t i ) is the coefficient to be determined, is random noise, i=1…n; is a matrix composed of the final results of the water quality parameters inverted daily in the target area, where d represents daily, g represents the field, (u i , v i ) represents a coordinate point, t i Indicates a certain moment, (u i , v i , t i ) is the space-time position at a certain moment and a certain spatial coordinate; A matrix of daily water quality parameters for interpolation prediction; make Calculating space-time distance In the formula, λ and μ are coefficients used to balance the influence between space and time. The optimal λ and μ are selected according to the fitting effect of the model under different parameters. Calculate the space-time distance decay function α ij , in h ST is the parameter of space-time bandwidth, h S is the parameter of spatial bandwidth, h T is the parameter of time bandwidth; i = 1...n; j = 1...n; Calculate the spatiotemporal weight matrix W(u i , v i , t i ), W(u i ,v i ,t i )=diag(a i1 ,a i2 ,…,a in ); Calculate the coefficients of the formula. in, is the coefficient calculated by the formula, including β0(u i ,v i ,t i ), β1(u i ,v i ,t i ) These two coefficients are the intercept and slope in the regression equation; β0(u i ,v i ,t i ) is a constant term, β1(u i ,v i ,t i ) are coefficients related to the interpolated water quality parameters; The coefficients obtained from nearby field observation points are used to correct the water environment and water quality parameter results of the surrounding areas, thus obtaining the daily water environment and water quality inversion results of the target area.

5. A system for inverting daily water quality by space-time statistical modeling based on integrated earth-sky integration, used to implement the method described in any one of claims 1 to 4; characterized in that: Includes the following modules: A data acquisition module is used to obtain satellite image data and water surface observation data of the target area, extract the water area in the satellite image data, and obtain spectral data of the water surface observation data; The water quality inversion model building module is used to associate satellite image data with water surface observation data in the extracted water area through temporal and spatial information, and use partial least squares regression to establish the water environment water quality inversion model and results of the target area under the condition of satellite data; The water quality inversion module is used to obtain the daily water quality inversion results of the target area through time series interpolation and correction using spatiotemporal geographical weighted regression.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method for inverting daily water quality by space-time statistical modeling based on the integration of heaven and earth as described in any one of claims 1 to 4 is implemented.

7. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for inverting daily water quality through space-time statistical modeling based on the integration of heaven and earth as described in any one of claims 1 to 4 is implemented.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for inverting daily water quality by space-time statistical modeling based on the integration of heaven and earth as claimed in any one of claims 1 to 4 is implemented.

Images