A remote sensing driven lake hydrodynamic water quality simulation method and related device

Abstract

This application discloses a remote sensing-driven method and related equipment for simulating lake hydrodynamics and water quality, relating to the field of data processing technology. The method includes: acquiring target water quality parameters of the target lake based on multi-source satellite imagery; generating a continuous time-series water quality inversion dataset based on the target water quality parameters; determining a baseline simulation sequence of water quality for the target lake based on a hydrodynamic model; subtracting the baseline simulation sequence from the inverted water quality dataset to obtain the residual, and determining an error correction amount based on the residual; and adding the baseline simulation sequence and the error correction amount to obtain the final water quality data sequence for the target lake. This application, by inverting and simulating water quality parameters, determining the residual, and then correcting the simulated baseline water quality sequence, yields a more accurate final water quality data sequence, providing a data foundation for lake management and monitoring.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a remote sensing-driven method and related equipment for simulating lake hydrodynamics and water quality. Background Technology

[0002] Lakes, as a crucial component of regional water resource systems, play an irreplaceable role in urban water supply security, water environment regulation, water ecosystem maintenance, and landscape functions. Especially against the backdrop of rapid urbanization, urban lakes have become key ecological infrastructure supporting multiple functions. However, in recent years, driven by both climate change and human activities, lake water environment problems have become increasingly complex. Frequent extreme rainfall events cause non-point source pollutants from watersheds to concentrate and enter lakes in a short period, forming pulse-like pollution inputs. Sustained high temperatures and heat waves exacerbate water stratification stability and rapid algal proliferation, inducing the risk of algal blooms. Simultaneously, human interventions such as lake enclosure development, shoreline hardening, water conservancy project regulation, and ecological water replenishment have altered the original hydrodynamic patterns of lakes, leading to decreased water exchange capacity and increased pollutant retention. The combined effect of these multiple factors results in lake water quality evolution exhibiting characteristics of strong suddenness, significant nonlinearity, and highly uneven spatial distribution, posing higher demands on the refined management and risk prevention of lake water environments. Summary of the Invention

[0003] The main objective of this application is to propose a remote sensing-driven method and related equipment for simulating lake hydrodynamics and water quality, so as to accurately predict lake hydrodynamics and water quality.

[0004] To achieve the above objectives, one aspect of this application proposes a remote sensing-driven method for simulating lake hydrodynamics and water quality, the method comprising the following steps: The target water quality parameters of the target lake are obtained from multi-source satellite imagery of the target lake. Generate a continuous time-series water quality inversion dataset based on the target water quality parameters; Determine the water quality baseline simulation sequence of the target lake based on the hydrodynamic model of the target lake; The water quality inversion dataset is subtracted from the water quality benchmark simulation sequence as the residual, and the error correction amount is determined based on the residual; The final water quality data sequence of the target lake is obtained by adding the water quality benchmark simulation sequence and the error correction amount.

[0005] In some embodiments, obtaining the target water quality parameters of the target lake based on multi-source satellite imagery of the target lake includes the following steps: The water mask of the target lake is determined from the multi-source satellite imagery; The target water quality parameters are obtained by inverting the regional image corresponding to the water mask in the multi-source satellite imagery.

[0006] In some embodiments, generating a time-series continuous water quality inversion dataset based on the target water quality parameters includes the following steps: The inversion model is constructed using the band ratio method; The inversion model is used to generate a time-series continuous water quality inversion dataset; The expression for the inversion model is: ; in, For the water quality inversion dataset, (λ1) and (λ2) represents the remote sensing reflectance of the specified band; a , b , c These are empirical coefficients calibrated using measured data.

[0007] In some embodiments, determining the water quality baseline simulation sequence of the target lake based on the hydrodynamic model of the target lake includes the following steps: A Delft3D-Flow model of the target lake is constructed as the hydrodynamic model; The three-dimensional flow field of the target lake is output using the Delft3D-Flow model. The continuity equation for the three-dimensional flow field is: ; The horizontal momentum equation for the three-dimensional flow field is: ; in, This refers to the free water surface elevation. H Total water depth; and They are respectively x , y The vertical average velocity in the direction of flow; It is the acceleration due to gravity; f For Coriolis force parameters; The horizontal eddy viscosity coefficient; The water quality benchmark simulation sequence of the target lake is obtained based on the three-dimensional flow field calculation.

[0008] In some embodiments, the step of calculating the water quality benchmark simulation sequence of the target lake based on the three-dimensional flow field includes the following steps: The water quality baseline simulation sequence is obtained by using the Delft3D-WAQ water quality ecology module, which is driven by the three-dimensional flow field to simulate the migration and transformation process of water quality parameters. The calculation formula of the Delft3D-WAQ water quality ecological module includes: ; in, The concentration of water quality parameters derived from the Delft3D-WAQ water quality ecological module, i.e., the water quality baseline simulation sequence; t For time; u , v They are respectively x , y Flow velocity component in the direction; , The horizontal diffusion coefficient; These are source / sink terms and biochemical reaction terms; ; in, For biochemical reaction rates; The settling rate of suspended solids; The bottom sediment resuspension flux driven by wind and wave bottom shear force; H The water is deep.

[0009] In some embodiments, determining the error correction amount based on the residual includes the following steps: Construct a deep neural network; wherein the deep neural network includes an input layer, multiple hidden layers and an output layer, and the hidden layers are connected by a fully connected network structure for forward propagation; No. l The formula for calculating the state response of neurons in each hidden layer is as follows: ; in, For the first l The output response feature vector of the hidden layer; when l When =1, the initial input =x, where x is the input feature vector; For the first l Layer weight matrix; It is the bias vector; For non-linear activation functions, the ReLU function is used, i.e. ; The current water quality state variables, meteorological driving factors, and hydrological driving factors output by the hydrodynamic model are combined into the input feature vector; The input feature vector and the residual are input into the deep neural network to obtain the error correction amount output by the deep neural network.

[0010] In some embodiments, before acquiring the target water quality parameters of the target lake based on multi-source satellite imagery of the target lake, the method further includes the following steps: The external boundary inputs of the target lake are set to change the state of the target lake; wherein, the external boundary inputs include extreme increase or decrease rates of watershed rainfall, extreme high temperature heat wave events, artificial water level regulation events, and ecological water replenishment scheduling events.

[0011] To achieve the above objectives, another aspect of this application proposes a remote sensing-driven lake hydrodynamic and water quality simulation device, the device comprising: The parameter acquisition unit is used to acquire the target water quality parameters of the target lake based on multi-source satellite images of the target lake. The parameter inversion unit is used to generate a time-series continuous water quality inversion dataset based on the target water quality parameters. The parameter simulation unit is used to determine the water quality baseline simulation sequence of the target lake based on the hydrodynamic model of the target lake; An error determination unit is used to subtract the water quality benchmark simulation sequence from the water quality inversion dataset as a residual, and determine the error correction amount based on the residual. An error correction unit is used to add the water quality benchmark simulation sequence and the error correction amount to obtain the final water quality data sequence of the target lake.

[0012] To achieve the above objectives, another aspect of this application provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described method.

[0013] To achieve the above objectives, another aspect of the embodiments of this application proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0014] To achieve the above objectives, another aspect of this application provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0015] The embodiments of this application include at least the following beneficial effects: This application provides a remote sensing-driven method and related equipment for simulating lake hydrodynamics and water quality. The method involves acquiring target water quality parameters of the target lake from multi-source satellite imagery; generating a continuous time-series water quality inversion dataset based on these parameters; determining a baseline simulation sequence of the target lake's water quality based on its hydrodynamic model; subtracting the baseline simulation sequence from the inverted dataset to obtain the residual; determining an error correction amount based on the residual; and finally, adding the baseline simulation sequence and the error correction amount to obtain the final water quality data sequence for the target lake. This application, by inverting and simulating water quality parameters, determining the residual, and then correcting the baseline simulation sequence, yields a more accurate final water quality data sequence, providing a data foundation for lake management and monitoring. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart illustrating a remote sensing-driven lake hydrodynamic and water quality simulation method provided in this application embodiment; Figure 2 An example flowchart of a remote sensing-driven lake hydrodynamic and water quality simulation method provided in this application embodiment; Figure 3 A schematic diagram of the structure of a remote sensing-driven lake hydrodynamic and water quality simulation device provided in an embodiment of this application; Figure 4 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit it. In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with those of this application; they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this application as detailed in the appended claims.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0020] Before providing a detailed description of the embodiments of this application, some related technologies involved in the embodiments of this application will be described first, as follows: From an engineering practice perspective, lake water environment issues typically involve multiple application scenarios, including pollution source analysis, water quality evolution prediction, ecological regulation optimization, and emergency response decision-making. For example, in the event of a sudden algal bloom, it is necessary to identify the pollution source and determine the degree and trend of eutrophication within a short period of time; in the process of ecological water transfer and water level regulation, it is necessary to assess the impact of different regulation schemes on the lake's hydrodynamic structure and water quality changes; and in the long-term governance process, it is necessary to analyze the cumulative effects of climate change and watershed development on the evolution of the lake water environment. All these applications require that the relevant technologies possess large-scale monitoring capabilities, process simulation capabilities, and rapid response capabilities to complex changing conditions.

[0021] The technologies widely used in current engineering projects mainly include on-site monitoring, remote sensing observation, and numerical simulation. On-site monitoring can provide high-precision water quality data, but it is limited by the layout of sampling points and sampling frequency, making it difficult to achieve large-scale, continuous dynamic monitoring. Remote sensing technology can realize the spatial distribution inversion of lake surface water quality, which has significant advantages in large-scale monitoring, but it is greatly affected by weather conditions, lacks temporal continuity, and is difficult to reflect the vertical structure and dynamic processes inside the water body. Hydrodynamic-water quality numerical models can describe the flow movement and pollutant transport processes of lakes from a mechanistic perspective, playing an important role in engineering analysis and scenario simulation. However, under complex environmental conditions, its simulation accuracy depends on a large amount of parameter calibration and high-quality input data, and its response capability to sudden changes and highly nonlinear processes still has certain limitations.

[0022] With the rapid development of information technology and computing power, lake water environment engineering is gradually evolving towards digitalization, refinement, and intelligence. On the one hand, the continuous acquisition of high spatiotemporal resolution remote sensing data has continuously improved the spatial monitoring capabilities of lake water quality; on the other hand, numerical simulation technology is increasingly widely used in multi-factor coupled analysis and scenario prediction; simultaneously, artificial intelligence and data-driven methods have demonstrated strong nonlinear fitting capabilities and computational efficiency advantages in complex system modeling. Driven by practical engineering needs, single technical means are no longer sufficient to meet the analysis and prediction requirements of complex lake systems, necessitating the construction of a comprehensive technical system that can integrate the advantages of multi-source observation data and multiple types of models.

[0023] Against this backdrop, how to effectively integrate and collaboratively utilize multi-source data in engineering applications has become a key issue in the field of lake water environment. On the one hand, it is necessary to use remote sensing data to compensate for the spatial coverage deficiencies of traditional monitoring and improve the breadth of water quality information acquisition; on the other hand, it is necessary to rely on physical mechanism models to systematically characterize water flow and material transport processes to support process analysis and scenario simulation; simultaneously, it is also necessary to introduce intelligent methods capable of handling complex nonlinear relationships to improve the system's response capability to sudden events and complex environmental changes. By constructing a comprehensive technical framework integrating "data acquisition—process simulation—dynamic response—decision support," continuous characterization and high-precision prediction of lake water environment status can be achieved.

[0024] Therefore, based on engineering application needs, it is necessary to develop a comprehensive simulation and analysis scheme for practical water environment management scenarios. This scheme, supported by multi-source data, should enable spatiotemporal dynamic monitoring of lake water quality, detailed characterization of key hydrodynamic and material transport processes, and stable and reliable predictive capabilities under complex climatic conditions and human disturbance scenarios. This technology can not only support the formulation of lake pollution prevention and ecological restoration engineering schemes, but also provide crucial technical support and decision-making basis for algal bloom early warning, water diversion scheduling optimization, and emergency response to sudden water environment events. It has significant engineering implications for improving the scientific rigor and precision of lake water environment governance.

[0025] The relevant technical details are as follows: 1. Lake water quality monitoring and comprehensive assessment.

[0026] As important surface water bodies, lakes directly impact regional ecological security and water supply stability through their internal material cycles and water quality. Accurately assessing lake water quality characteristics and their spatiotemporal distribution is central to water environment management. Currently, lake water quality assessment and dynamic monitoring technologies cover a wide range, from microscopic sampling to macroscopic observation. Based on different data sources and observation scales, the main methods include: (1) Conventional monitoring methods based on on-site sampling.

[0027] Traditional water quality monitoring methods rely on field survey vessels to set up fixed points in the lake study area, collect samples, and bring them back to the laboratory for chemical analysis. This method provides data with extremely high absolute accuracy, serving as the benchmark for all water quality assessments. However, field sampling requires significant human and material resources, makes it difficult to provide large-scale and long-term continuous assessment results, and cannot capture dynamic changes in lake water quality within a short period.

[0028] (2) Water quality inversion method based on multi-source remote sensing technology.

[0029] With the development of satellite remote sensing technology, remote sensing-based lake water quality assessment methods have been widely applied. By establishing inversion algorithms between water reflectance spectra and water quality parameters, the spatial distribution characteristics of key water quality indicators such as chlorophyll a (Chl-a), total suspended solids (TSS), and colored soluble organic matter (CDOM) in large-scale lakes can be obtained. Currently, significant progress has been made in the classification and high-precision inversion algorithms for complex inland optical water bodies, significantly improving the inversion accuracy of optically active substances in complex water bodies. Meanwhile, research on remote sensing inversion of lake CDOM and chlorophyll characteristics using data from satellites such as Landsat has also confirmed its feasibility in monitoring large-scale water quality changes. However, optical remote sensing technology is highly susceptible to cloud and rain weather conditions, making it difficult to obtain high-quality data that is completely continuous over time. Furthermore, current remote sensing methods largely focus on static assessments of surface water quality, failing to explore the biochemical dynamics mechanisms within the water body.

[0030] 2. Simulation of lake hydrodynamic-water quality physical mechanisms.

[0031] To explore the dynamics of water flow and pollutant migration within lakes, researchers widely employ physical mechanism-based hydrodynamic and water quality coupling models (such as Delft3D and EFDC) for numerical simulations. These models, by solving fluid dynamics equations (such as shallow water equations) and mass transport equations, link lake meteorological drivers, inflow runoff, and changes in the lake's aquatic environment. They can characterize the flow field structure within the lake and the physical processes of pollutant convection and diffusion, and incorporate ecological modules to simulate biochemical reactions. For example, using Delft3D to construct a three-dimensional hydrodynamic-ecological coupling model can effectively assess the spatiotemporal heterogeneity of lake water quality and the changing patterns of ecological states. In shallow lakes, three-dimensional numerical models have also been successfully used to simulate complex hydrodynamic cycles and sediment suspension and transport processes. The advantages of these models are that their architecture strictly adheres to the law of conservation of mass and possesses strong physical interpretability. However, highly nonlinear biochemical processes exist in natural lakes, and existing fixed partial differential equations often fail to perfectly characterize these complex mechanisms, easily leading to systematic errors.

[0032] 3. Application of deep learning in predicting water environment evolution.

[0033] With the rapid development of environmental informatics and big data, data-driven machine learning and deep network algorithms (such as deep neural networks and random forests) are widely used in predicting complex water parameters. These methods do not require pre-setting complex fluid dynamics or physicochemical partial differential equations; instead, they directly extract latent feature information from monitoring sequences, meteorological, and hydrological driving data, fitting a highly nonlinear mapping and response relationship between external driving factors and target water quality variables. Research shows that using multi-layer networks or ensemble learning models can accurately analyze the influence weights of environmental factors (such as temperature and rainfall) on water quality evolution, demonstrating high fitting accuracy and computational efficiency in water quality prediction. Furthermore, deep learning architectures have shown potential to surpass traditional statistical models in handling complex nonlinear dynamic changes in the aquatic environment. The main characteristic of these models is their high reliance on the data itself to uncover potential response patterns; in practical applications, they typically require a large amount of continuous observational data for training.

[0034] The drawbacks of related technologies: (1) Although existing lake hydrodynamic and water quality models have a certain mechanistic basis, they are still prone to systematic errors when faced with complex systems where water flow, pollutant transport, biochemical reactions and ecological processes are coupled. At the same time, these models have many parameters, complex calibration processes, and high requirements for the accuracy of boundary conditions and basic data, resulting in insufficient adaptability and stability in complex and changing environments.

[0035] (2) Although existing machine learning or deep learning methods have strong nonlinear fitting capabilities, they usually require large-scale, long-term, continuous and high-quality data as the training basis. However, actual lake water quality monitoring data often suffers from problems such as insufficient samples, discontinuous time, and limited spatial coverage, which affect the training effect, prediction stability and application capability of the model.

[0036] (3) Existing remote sensing technologies are mostly used for lake water body identification, water quality parameter inversion and static monitoring and evaluation. They are mainly limited to the observation and identification level and have not been fully integrated into the continuous simulation process of lake hydrodynamics and water quality. Therefore, they are difficult to effectively support model-driven, dynamic correction and continuous prediction.

[0037] (4) Most existing technologies focus on the application of a single method and have not yet formed a complete technical system that organically couples multi-source remote sensing information, physical mechanism models and deep learning models. Therefore, it is difficult to simultaneously take into account mechanism interpretability, simulation accuracy, dynamic response capability and engineering application efficiency, which limits the ability to simulate and predict lake water quality with high precision under complex and changing environments.

[0038] In view of the shortcomings of the aforementioned related technologies, the technical problem to be solved by this application is explained as follows: Climate change (such as heavy rainfall and extreme heat) and human activities (such as non-point source pollution and lake development) alter the hydrological cycle and hydrodynamic conditions of lakes, causing fluctuations in the physicochemical properties of water bodies. This profoundly impacts lake water quality and the ecological environment, and can accumulate and deteriorate in localized areas, leading to eutrophication and even algal blooms, resulting in the degradation of aquatic ecosystems. The essence of this series of problems is the cascading effect of climate change and human activities on physical water transport and biochemical processes, affecting lake water quality evolution. Changing environments first alter physical parameters such as lake flow fields, water levels, and temperatures, which in turn disturb complex biochemical processes such as algal proliferation, suspended solids sedimentation, and colored soluble organic matter, inducing high spatiotemporal heterogeneity in water quality parameters. Similarly, at the model simulation level, traditional single-physical models or data-driven models suffer from persistent systematic biases and nonlinear fitting bottlenecks in responding to complex external conditions. This application constructs a multi-source remote sensing inversion-driven foundation, combines a lake hydrodynamic-water quality physics master model with a deep learning residual enhancement module, and analyzes the coupled effects and dynamic response mechanisms of changing environments on lake flow field structure, water quality material transport and biochemical reactions based on a scenario-driven scheme. It applies a two-layer collaborative mechanism of physics-led and data-driven enhancement to compensate for system simulation errors, providing a scientific basis for formulating precise lake water environment governance policies and adapting to and mitigating the impact of extreme environmental disturbances.

[0039] This application provides a remote sensing-driven method and related equipment for simulating lake hydrodynamics and water quality, relating to the field of data processing technology. The remote sensing-driven method and related equipment provided in this application can be applied to a terminal, a server, or software running on either a terminal or a server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, or vehicle terminal, but is not limited to these. The server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server can also be a node server in a blockchain network. The software can be an application implementing the remote sensing-driven method and related equipment for simulating lake hydrodynamics and water quality, but is not limited to the above forms.

[0040] This application can be applied to numerous general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics devices, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0041] Reference Figure 1 This application provides a remote sensing-driven method for simulating lake hydrodynamics and water quality. This method may include, but is not limited to, steps S100 to S140, as detailed below: S100: Obtain the target water quality parameters of the target lake based on multi-source satellite imagery of the target lake; S110: Generate a continuous time-series water quality inversion dataset based on the target water quality parameters; S120: Determine the water quality baseline simulation sequence of the target lake based on the hydrodynamic model of the target lake; S130: Subtract the water quality baseline simulation sequence from the water quality inversion dataset to obtain the residual, and determine the error correction amount based on the residual; S140: Add the water quality benchmark simulation sequence and the error correction amount to obtain the final water quality data sequence of the target lake.

[0042] Optionally, obtaining the target water quality parameters of the target lake based on multi-source satellite imagery of the target lake includes the following steps: The water mask of the target lake is determined from the multi-source satellite imagery; The target water quality parameters are obtained by inverting the regional image corresponding to the water mask in the multi-source satellite imagery.

[0043] Optionally, generating a time-series continuous water quality inversion dataset based on the target water quality parameters includes the following steps: The inversion model is constructed using the band ratio method; The inversion model is used to generate a time-series continuous water quality inversion dataset; The expression for the inversion model is: ; in, For the water quality inversion dataset, (λ1) and (λ2) represents the remote sensing reflectance of the specified band; a , b , c These are empirical coefficients calibrated using measured data.

[0044] Optionally, determining the water quality baseline simulation sequence of the target lake based on the hydrodynamic model of the target lake includes the following steps: A Delft3D-Flow model of the target lake is constructed as the hydrodynamic model; The three-dimensional flow field of the target lake is output using the Delft3D-Flow model. The continuity equation for the three-dimensional flow field is: ; The horizontal momentum equation for the three-dimensional flow field is: ; in, This refers to the free water surface elevation. H Total water depth; and They are respectively x , y The vertical average velocity in the direction of flow; It is the acceleration due to gravity; f For Coriolis force parameters; The horizontal eddy viscosity coefficient; The water quality benchmark simulation sequence of the target lake is obtained based on the three-dimensional flow field calculation.

[0045] Optionally, the step of calculating the water quality benchmark simulation sequence of the target lake based on the three-dimensional flow field includes the following steps: The water quality baseline simulation sequence is obtained by using the Delft3D-WAQ water quality ecology module, which is driven by the three-dimensional flow field to simulate the migration and transformation process of water quality parameters. The calculation formula of the Delft3D-WAQ water quality ecological module includes: ; in, The concentration of water quality parameters derived from the Delft3D-WAQ water quality ecological module, i.e., the water quality baseline simulation sequence; t For time; u , v They are respectively x , yFlow velocity component in the direction; , The horizontal diffusion coefficient; These are source / sink terms and biochemical reaction terms; ; in, For biochemical reaction rates; The settling rate of suspended solids; The bottom sediment resuspension flux driven by wind and wave bottom shear force; H The water is deep.

[0046] Optionally, determining the error correction amount based on the residual includes the following steps: Construct a deep neural network; wherein the deep neural network includes an input layer, multiple hidden layers and an output layer, and the hidden layers are connected by a fully connected network structure for forward propagation; No. l The formula for calculating the state response of neurons in each hidden layer is as follows: ; in, For the first l The output response feature vector of the hidden layer; when l When =1, the initial input =x, where x is the input feature vector; For the first l Layer weight matrix; It is the bias vector; For non-linear activation functions, the ReLU function is used, i.e. ; The current water quality state variables, meteorological driving factors, and hydrological driving factors output by the hydrodynamic model are combined into the input feature vector; The input feature vector and the residual are input into the deep neural network to obtain the error correction amount output by the deep neural network.

[0047] Optionally, before obtaining the target water quality parameters of the target lake based on multi-source satellite imagery of the target lake, the method further includes the following steps: The external boundary inputs of the target lake are set to change the state of the target lake; wherein, the external boundary inputs include extreme increase or decrease rates of watershed rainfall, extreme high temperature heat wave events, artificial water level regulation events, and ecological water replenishment scheduling events.

[0048] The following sections will provide a detailed description and explanation of some optional embodiments of this application, using specific application examples.

[0049] This embodiment discloses a remote sensing-driven method for synergistic enhancement simulation of lake hydrodynamics and water quality. This method uses multi-source remote sensing imagery as the primary external observation data source, a lake hydrodynamic-water quality physical model as the main model, and a deep learning residual model as the enhancement module. By constructing a synergistic operation mechanism of "continuous remote sensing observation—physical mechanism deduction—nonlinear error learning—dynamic feedback correction," it achieves high-precision simulation and prediction of the spatiotemporal distribution of lake water quality under complex changing environments. This method can be applied to scenarios such as water quality evolution analysis, algal bloom risk early warning, ecological regulation assessment, and emergency prediction of sudden water environment events for urban lakes, shallow lakes, and lakes significantly affected by climate change and human activities. The workflow of this embodiment is described below. Figure 2 .

[0050] Specifically, this embodiment includes the following solutions: 1. Multi-source remote sensing water quality inversion and continuous driving data construction.

[0051] 1) Acquisition and preprocessing of multi-source remote sensing data.

[0052] Sentinel-2 is a high-resolution multispectral imaging satellite equipped with a multispectral imager (MSI) featuring 13 spectral bands and a spatial resolution of up to 10 meters. Its dual-satellite network has a short revisit period, giving it a significant spectral advantage in refined monitoring of inland shallow lakes. The Landsat series satellites (such as Landsat-8 / 9) are equipped with an Online Land Imager (OLI). This sensor boasts an extremely high signal-to-noise ratio and 12-bit radiometric resolution, and features a deep blue band highly sensitive to water and atmospheric aerosols. It has significant advantages in accurately removing atmospheric interference and capturing weak spectral signals of low-concentration optically active substances (such as CDOM and total suspended matter) in complex inland water bodies, providing a stable and high-quality observational benchmark for long-term water environment evolution inversion. This embodiment collects high-resolution remote sensing images of a target lake from Sentinel-2 and Landsat as the primary data source. Atmospheric correction models such as Sen2Cor are used to eliminate the effects of scattering and absorption by atmospheric molecules and aerosols, obtaining accurate water surface remote sensing reflectance (…). The improved Normalized Difference Water Index (MNDWI) was used to extract lake water masks and remove shoreline and building backgrounds.

[0053] 2) Target water quality parameter inversion and continuous driving data construction.

[0054] To address the characteristics of non-point source pollution input and eutrophication in lakes, this study focuses on retrieving key water quality parameters such as CDOM, chlorophyll a, and TSS. Based on preprocessed remote sensing reflectance of the water surface, an inversion model is constructed using the band ratio method combined with empirical algorithms. Feature bands (such as the red and near-infrared bands) are selected to construct feature factors, and their inversion formulas are as follows: ; In the formula: These are the observed values ​​of lake water quality parameters obtained through inversion. (λ1) and (λ2) represents the remote sensing reflectance of the specified band; a, b, and c are empirical coefficients calibrated using measured data. Since optical remote sensing is susceptible to interference from cloud and rain weather, a spatiotemporal data fusion algorithm (such as the FSDAF algorithm) is used to generate a continuous time-series water quality inversion dataset. Measured data from water quality stations in the study area are used, employing root mean square error (RMSE) and coefficient of determination (R²). 2 The accuracy of inversion and fusion is evaluated: when R 2 When the value is >0.6 and the RMSE is small, the remote sensing inversion is considered accurate.

[0055] 2. Construction of lake hydrodynamic-water quality physical model.

[0056] 1) Construction of Delft3D-Flow hydrodynamic model.

[0057] Delft3D is a three-dimensional water environment numerical simulation software developed by Deltares, a Dutch company. Delft3D-Flow is its core module. This model uses orthogonal curved meshes to solve the shallow water equations in the horizontal direction, and effectively solves the horizontal (2D) and three-dimensional shallow water equations using the Alternating Direction Implicit Method (ADI). Delft3D-Flow operates by discretizing the study area into a mesh system, dividing the water body into smaller units, and calculating the water flow motion in each unit using spatial and temporal meshes. The model is easy to operate and has good applicability in shallow lake simulation. This embodiment uses high-resolution remote sensing to obtain the lake shoreline boundary, uses the Delft3D-RGFGRID module to create an orthogonal curved mesh, and performs lake basin topography interpolation based on measured elevation data. Meteorological station measured data (wind speed, rainfall, temperature) and runoff data are used as boundary conditions to drive the model. The internal hydrodynamic calculations of the model are based on the Navier-Stokes equations and the Boussinesq assumption for three-dimensional incompressible fluids. The continuity equation and horizontal momentum equation are as follows: ; ; In the formula: This refers to the free water surface elevation. H Total water depth; and They are respectively x , y The vertical average velocity in the direction of flow; It is the acceleration due to gravity; f For Coriolis force parameters; The horizontal eddy viscosity coefficient is used. Finally, the model parameters and roughness are calibrated using measured water level and flow velocity data, and the Nash coefficient (NSE) is used to evaluate the accuracy of the hydrodynamic flow field simulation.

[0058] 2) Coupled calculation of water quality and ecological dynamics module.

[0059] Based on the convergence of hydrodynamic flow field calculations, a loosely coupled approach is adopted, using the flow field results output by Delft3D-Flow as input to drive the Delft3D-WAQ water quality and ecology module. The core mechanism of the physical model is based on the convection-diffusion-reaction equation, used to simulate the migration and transformation processes of water quality parameters. Its basic equations are as follows: ; In the formula: Concentration of water quality parameters derived from physical models; t For time; u , v They are respectively x , y Flow velocity component in the direction; , The horizontal diffusion coefficient; These are source and sink terms and biochemical reaction terms, used to characterize the kinetic processes of algal growth, sedimentation, and sediment release.

[0060] The calculation of its biochemical reaction source term is expanded as follows: ; In the formula: This refers to the rate of biochemical reactions (such as the photodegradation of CDOM or the growth and death rate of algae). The settling rate of suspended solids; The bottom sediment resuspension flux driven by wind and wave bottom shear force; H Let the water depth be denoted by . From this equation, a water quality baseline simulation sequence following the law of conservation of mass can be obtained.

[0061] 3. AI residual learning module construction and dynamic correction mechanism.

[0062] 1) Extraction of system residual sequences.

[0063] Since fixed partial differential equations in physical models cannot perfectly characterize highly nonlinear biochemical processes under extreme changing environments (such as the instantaneous impact of non-point source pollution and the sudden proliferation of algal blooms caused by high temperatures), this embodiment introduces artificial intelligence algorithms for residual compensation. Under the condition of matching spatial grids with time nodes, the true values ​​of water quality remote sensing inversion observations are extracted. Compared with physical baseline simulation values Systematic errors between: ; In the formula: The system residuals of the physical model at a specific spatiotemporal node; These are the observation values ​​of key water quality parameters (such as CDOM, chlorophyll a, etc.) obtained based on remote sensing inversion; These are the baseline simulated values ​​of water quality parameters output from the pure physical master model.

[0064] 2) Construction of error response relationship based on deep neural network (DNN).

[0065] To avoid traditional time series models relying too heavily on historical data inertia and neglecting the instantaneous response to sudden environmental events, this embodiment constructs a deep neural network (DNN) to accurately characterize the highly nonlinear response relationship between external driving factors and physical model errors. The current water quality state variables output by the physical model, along with multidimensional meteorological and hydrological driving factors (such as abrupt changes in temperature, instantaneous rainfall, and wind speed), are combined into an input feature vector x. The DNN consists of an input layer, multiple hidden layers, and an output layer, with fully connected network structures used for forward propagation between the hidden layers. l The formula for calculating the state response of neurons in each hidden layer is as follows: ; In the formula: For the first l The output response feature vector of the hidden layer; when l When =1, the initial input =x; For the first l The weight matrix of the layer represents the nonlinear connections and response strengths between the features of each driving factor; It is the bias vector; For non-linear activation functions, the ReLU function is used, i.e. This is to enhance the model's sensitivity to extreme mutation signals and alleviate the gradient vanishing problem.

[0066] 4. The physical model and deep learning work together to enhance the operation mechanism.

[0067] A real-time dynamic feedback mechanism is established between the physical solver and the residual response predictor. At each time step of the simulation, the nonlinear error correction amount output by the DNN module based on the current driving factor response is adjusted. This data is then overlaid in real time onto the preliminary physical calculation results to achieve dynamic enhancement and correction of accuracy. ; In the formula: This represents the final value of the high-fidelity water quality state variable after data-driven response enhancement. Subsequently, this variable is directly updated in the initial field of the physical model for the next time step, thus forming a two-way closed-loop collaborative operation framework of "physical mechanism deduction - driving factor response - dynamic error correction - feedback iteration".

[0068] 5. Scenario-driven and extreme condition simulation.

[0069] After completing the multi-source remote sensing inversion, physical model calibration, and DNN-assisted enhancement, a multi-scenario combined driving scheme is constructed. To comprehensively assess the water quality evolution patterns of lakes under complex environments, different dimensional change conditions are introduced as external boundary inputs, including but not limited to: extreme increases and decreases in watershed rainfall (such as the pulse-like inflow of non-point source pollution caused by sudden rainstorms), extreme high-temperature heat wave events (such as inducing abnormal algal proliferation and algal blooms), and artificial water level regulation and ecological water replenishment scheduling. Under the operation of the two-layer collaborative system, the above scenario settings are transformed into meteorological and hydrological driving factors and input into the coupled model. The system can accurately simulate and output the spatial distribution characteristics and dynamic response processes of the three-dimensional flow field structure evolution, material transport flux, and key water quality indicators (CDOM, Chl-a, TSS, etc.) within the lake under the combined influence of multiple scenarios. In this way, the evolutionary cascade effects, the self-purification and regulation potential of the aquatic ecosystem, and its resilience under multiple environmental and climatic stresses of urban lakes can be quantitatively assessed, providing high-precision scientific decision-making support for the formulation of forward-looking comprehensive lake water environment management and risk prevention strategies.

[0070] In summary, this embodiment has at least the following beneficial effects: In terms of originality, this embodiment proposes a two-layer collaborative lake water quality simulation system that combines "physical dominance with data-driven enhancement." Traditional purely physical hydrodynamic water quality models often suffer from unavoidable systematic biases when depicting complex nonlinear biochemical processes, while pure deep learning models, as simulations, are prone to producing inferences that do not conform to physical laws when lacking long-term continuous monitoring data or facing extreme unknown scenarios. This scheme combines multi-source remote sensing water quality inversion, a physical mechanism master model, and an AI residual learning module. It retains the interpretability of physical mechanism models in terms of material conservation and system structure, while effectively compensating for the nonlinear systematic biases of water quality dynamic evolution using deep learning, achieving complementary advantages between mechanisms and data.

[0071] In terms of value: The core value of this technical solution lies in improving the accuracy and robustness of lake water quality simulation and overall assessment, while maintaining low data acquisition costs. By continuously retrieving key lake water quality parameters (such as colored soluble organic matter and chlorophyll a) using multi-source remote sensing technology, and combining this with a multi-index comprehensive evaluation system, a large-scale, low-cost, continuous driving and validation data foundation is provided for the physical model. This overcomes the pain point of traditional simulation methods' over-reliance on expensive and spatially sparse field measurement data. This high-precision collaborative enhancement model provides scientific data support for exploring the dominant factors of lake water quality and formulating precise water ecological governance strategies.

[0072] In terms of effectiveness: Against the backdrop of current climate change and intensified human activities, lake aquatic ecosystems frequently face extreme disturbances such as heat waves and non-point source pollution impacts caused by heavy rainfall. Traditional static models struggle to respond quickly and accurately. The effectiveness of this embodiment lies in its constructed AI residual learning dynamic correction mechanism, which can agilely and dynamically compensate for the spatial heterogeneity and extreme disturbance conditions of lake water bodies. By introducing different climate scenarios or extreme conditions as drivers, the model can promptly output high-fidelity predictions of lake water quality evolution, buying valuable time for early warning of current and future water environment emergencies, pollution prevention and source tracing, and emergency response.

[0073] In terms of systemic aspects: This embodiment employs a multi-level, comprehensive analysis method, overcoming the limitations of single-factor inversion or single underlying models. This scheme constructs a complete and rigorous operational prediction and evaluation system, building a complete technical chain from continuous inversion of remote sensing data to the construction of physical mechanism models, then dynamic enhancement through artificial intelligence residual learning, and finally implementation in multi-scenario driven prediction. This scheme not only covers all core technical nodes from air-space-ground observation data acquisition and underlying mechanism model construction to intelligent system bias correction, but also deeply couples the complex hydrodynamic structure and water quality chemical evolution processes within lakes at the system level. This forms a complete and tightly integrated lake water quality collaborative enhancement simulation framework, and also provides a valuable systematic methodology for water environment assessment and scientific prediction of other complex large water bodies.

[0074] The key technical solution of this embodiment is as follows: (1) Acquisition of multi-source remote sensing water quality information and continuous driving construction.

[0075] By acquiring multi-source remote sensing image data, the lake's water body range and key water quality parameters are extracted, inverted, and spatiotemporally fused to construct a dynamic lake water quality dataset with temporal continuity and spatial integrity, providing basic data support for subsequent model simulation, state verification, and error correction.

[0076] (2) Construction of a lake hydrodynamic-water quality coupled physical model.

[0077] Based on basic data such as lake topography, water depth, shoreline, inflow and outflow boundaries, meteorological conditions, and pollution load, a coupled simulation system of lake hydrodynamic model and water quality model is established to realize the mechanistic description of flow field, water level field, and pollutant transport, diffusion, and transformation processes.

[0078] (3) Extraction of residual samples from remote sensing observation and physical simulation.

[0079] The water quality parameters obtained by remote sensing inversion are matched with the output results of the physical model on a unified time scale and spatial grid to construct the difference residual sequence between the two, so as to characterize the systematic bias and nonlinear error features that the physical model fails to fully describe under complex and changing environments.

[0080] (4) Construction of deep learning residual compensation model.

[0081] Using the physical model output state, meteorological and hydrological driving factors, and related spatiotemporal characteristics as inputs, and the residual correction amount as output, a deep learning model is trained to learn the variation law of simulation error under complex environment, thereby improving the accuracy and stability of lake water quality simulation results.

[0082] (5) Closed-loop collaborative enhancement mechanism between physical model and intelligent model.

[0083] During the model operation, the physical model first generates the baseline simulation results for the current moment, and then the deep learning model predicts the error correction amount. The correction results are then fed back to the current output and the calculation process for the next time step, forming a dynamic closed-loop operation mode that combines mechanism simulation and intelligent correction.

[0084] (6) Scenario simulation and prediction applications in complex and changing environments.

[0085] By setting different scenario inputs such as extreme rainfall, high temperature heat waves, water level regulation, ecological water replenishment, and pollution load changes, the collaborative model is driven to run, enabling dynamic prediction, risk identification, and management decision support for the spatiotemporal evolution of lake water quality.

[0086] Reference Figure 3 This application also provides a remote sensing-driven lake hydrodynamic and water quality simulation device, which can realize the above-mentioned remote sensing-driven lake hydrodynamic and water quality simulation method. The device includes: The parameter acquisition unit is used to acquire the target water quality parameters of the target lake based on multi-source satellite images of the target lake. The parameter inversion unit is used to generate a time-series continuous water quality inversion dataset based on the target water quality parameters. The parameter simulation unit is used to determine the water quality baseline simulation sequence of the target lake based on the hydrodynamic model of the target lake; An error determination unit is used to subtract the water quality benchmark simulation sequence from the water quality inversion dataset as a residual, and determine the error correction amount based on the residual. An error correction unit is used to add the water quality benchmark simulation sequence and the error correction amount to obtain the final water quality data sequence of the target lake.

[0087] It is understood that the content of the above method embodiments is applicable to the present device embodiments. The specific functions implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0088] This application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the method of this application. This electronic device can be any smart terminal, including tablet computers, in-vehicle computers, etc.

[0089] It is understood that the content of the above method embodiments is applicable to the device embodiments. The specific functions implemented by the device embodiments are the same as those of the methods of this application, and the beneficial effects achieved are the same as those achieved by the methods of this application.

[0090] Figure 4 The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes: The processor 101 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application. The memory 102 can be implemented as a read-only memory (ROM), static storage device, dynamic storage device, or random access memory (RAM). The memory 102 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 102 and is called and executed by the processor 101. Input / output interface 103 is used to implement information input and output; The communication interface 104 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.). Bus 105 transmits information between various components of the device (e.g., processor 101, memory 102, input / output interface 103, and communication interface 104); The processor 101, memory 102, input / output interface 103 and communication interface 104 are connected to each other within the device via bus 105.

[0091] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method of this application.

[0092] It is understood that the content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0093] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0094] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0095] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0096] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0097] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0098] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0099] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0100] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0101] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0102] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0103] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0104] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0105] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. A remote sensing-driven method for simulating lake hydrodynamics and water quality, characterized in that, The method includes the following steps: The target water quality parameters of the target lake are obtained from multi-source satellite imagery of the target lake. Generate a continuous time-series water quality inversion dataset based on the target water quality parameters; Determine the water quality baseline simulation sequence of the target lake based on the hydrodynamic model of the target lake; The water quality inversion dataset is subtracted from the water quality benchmark simulation sequence as the residual, and the error correction amount is determined based on the residual; The final water quality data sequence of the target lake is obtained by adding the water quality benchmark simulation sequence and the error correction amount.

2. The remote sensing-driven lake hydrodynamic and water quality simulation method according to claim 1, characterized in that, The process of obtaining the target water quality parameters of the target lake based on multi-source satellite imagery includes the following steps: The water mask of the target lake is determined from the multi-source satellite imagery; The target water quality parameters are obtained by inverting the regional image corresponding to the water mask in the multi-source satellite imagery.

3. The remote sensing-driven lake hydrodynamic and water quality simulation method according to claim 1, characterized in that, The process of generating a continuous time-series water quality inversion dataset based on the target water quality parameters includes the following steps: The inversion model is constructed using the band ratio method; The inversion model is used to generate a time-series continuous water quality inversion dataset; The expression for the inversion model is: ； in, For the water quality inversion dataset, (λ1) and (λ2) represents the remote sensing reflectance of the specified band; a , b , c These are empirical coefficients calibrated using measured data.

4. The remote sensing-driven lake hydrodynamic and water quality simulation method according to claim 1, characterized in that, The step of determining the water quality baseline simulation sequence of the target lake based on the hydrodynamic model of the target lake includes the following steps: A Delft3D-Flow model of the target lake is constructed as the hydrodynamic model; The three-dimensional flow field of the target lake is output using the Delft3D-Flow model. The continuity equation for the three-dimensional flow field is: ； The horizontal momentum equation for the three-dimensional flow field is: ； in, This refers to the free water surface elevation. H Total water depth; and They are respectively x , y The vertical average velocity in the direction of flow; It is the acceleration due to gravity; f For Coriolis force parameters; The horizontal eddy viscosity coefficient; The water quality benchmark simulation sequence of the target lake is obtained based on the three-dimensional flow field calculation.

5. The remote sensing-driven lake hydrodynamic and water quality simulation method according to claim 4, characterized in that, The process of calculating the water quality benchmark simulation sequence of the target lake based on the three-dimensional flow field includes the following steps: The water quality baseline simulation sequence is obtained by using the Delft3D-WAQ water quality ecology module, which is driven by the three-dimensional flow field to simulate the migration and transformation process of water quality parameters. The calculation formula of the Delft3D-WAQ water quality ecological module includes: ； in, The concentration of water quality parameters derived from the Delft3D-WAQ water quality ecological module, i.e., the water quality baseline simulation sequence; t For time; u , v They are respectively x , y Flow velocity component in the direction; , The horizontal diffusion coefficient; These are source / sink terms and biochemical reaction terms; ； in, For biochemical reaction rates; The settling rate of suspended solids; The bottom sediment resuspension flux driven by wind and wave bottom shear force; H The water is deep.

6. The remote sensing-driven lake hydrodynamic and water quality simulation method according to claim 1, characterized in that, Determining the error correction amount based on the residual includes the following steps: Construct a deep neural network; wherein the deep neural network includes an input layer, multiple hidden layers and an output layer, and the hidden layers are connected by a fully connected network structure for forward propagation; No. l The formula for calculating the state response of neurons in each hidden layer is as follows: ； in, For the first l The output response feature vector of the hidden layer; when l When =1, the initial input =x, where x is the input feature vector; For the first l Layer weight matrix; It is the bias vector; For non-linear activation functions, the ReLU function is used, i.e. ; The current water quality state variables, meteorological driving factors, and hydrological driving factors output by the hydrodynamic model are combined into the input feature vector; The input feature vector and the residual are input into the deep neural network to obtain the error correction amount output by the deep neural network.

7. A remote sensing-driven method for simulating lake hydrodynamics and water quality according to any one of claims 1 to 6, characterized in that, Before acquiring the target water quality parameters of the target lake based on multi-source satellite imagery, the method further includes the following steps: The external boundary inputs of the target lake are set to change the state of the target lake; wherein, the external boundary inputs include extreme increase or decrease rates of watershed rainfall, extreme high temperature heat wave events, artificial water level regulation events, and ecological water replenishment scheduling events.

8. A remotely sensing-driven lake hydrodynamic and water quality simulation device, characterized in that, The device includes: The parameter acquisition unit is used to acquire the target water quality parameters of the target lake based on multi-source satellite images of the target lake. The parameter inversion unit is used to generate a time-series continuous water quality inversion dataset based on the target water quality parameters. The parameter simulation unit is used to determine the water quality baseline simulation sequence of the target lake based on the hydrodynamic model of the target lake; An error determination unit is used to subtract the water quality benchmark simulation sequence from the water quality inversion dataset as a residual, and determine the error correction amount based on the residual. An error correction unit is used to add the water quality benchmark simulation sequence and the error correction amount to obtain the final water quality data sequence of the target lake.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 7.

Images