Method for predicting migration of pollutants in a basin based on interface distribution and river network topological constraints

Abstract

This application discloses a watershed pollutant migration prediction method based on interface allocation and river network topological constraints. The method includes: acquiring multi-source environmental data of the watershed; establishing a prediction model for the sediment-water interface allocation coefficient; constructing a migration potential index characterizing the tendency of pollutants to migrate from the sediment phase to the water phase based on the prediction model; constructing a watershed river network topology and establishing hydrological connections between river network nodes; determining the pollutant migration tendency and, combined with the watershed river network topology constraining the pollutant propagation direction, calculating the migration probability and cumulative migration effect between nodes; and finally determining the pollutant migration path information within the target watershed to be predicted. This application can achieve accurate prediction of the pollution migration process of watershed pollutants, improve the accuracy of pollutant migration path identification and prediction, and thus realize pollution early warning and control of watershed pollutants. It can be widely applied in the field of environmental data processing technology.

Description

Technical Field

[0001] This application relates to the field of environmental data processing technology, and in particular to a watershed pollutant migration prediction method based on interface assignment and river network topology constraints. Background Technology

[0002] In related technologies, the aquatic environment is an important carrier for the migration and accumulation of pollutants (such as heavy metals) in the natural environment, while sediments are important accumulation media for pollutants and can serve as potential sources of pollutant release under certain conditions. The sediment-aquatic interface partition coefficient is an important parameter characterizing the partitioning behavior and migration capacity of pollutants between the aquatic and sedimentary phases. The ecological risks posed by pollutants to the watershed aquatic environment system are closely related to their partitioning, migration, and transformation processes at the water-sediment interface.

[0003] However, existing methods for pollutant risk prediction and assessment mostly focus on pollutant content characteristics, spatial distribution analysis based on saturation levels, pollution source apportionment, and static risk assessment, while insufficiently considering the distribution behavior of pollutants at the water-sediment interface under dynamic aquatic environmental conditions. Furthermore, current technologies often fail to fully integrate the watershed network topology, upstream and downstream hydrological connections, and the migration process of pollutants along the flow direction, making it difficult to effectively characterize the migration direction, migration path, and migration probability of pollutants at the watershed scale. Therefore, these technologies still have certain limitations in identifying and predicting the migration behavior and propagation of pollutants in watersheds and require further improvement. Summary of the Invention

[0004] The main purpose of this application is to propose a watershed pollutant migration prediction method based on interface allocation and river network topology constraints, which can improve the accuracy of pollutant migration path identification and prediction, and thus realize watershed pollutant pollution early warning and control.

[0005] To achieve the above objectives, one aspect of this application proposes a watershed pollutant migration prediction method based on interface allocation and river network topology constraints, comprising: Acquire multi-source environmental data of the watershed; Construct a multi-source pollution database and establish a predictive model for sediment-water interface partition coefficients; Based on the prediction model, a migration potential index characterizing the tendency of pollutants to migrate from the sedimentary phase to the aquatic phase is constructed; Construct a river network topology for the watershed and establish hydrological connections between river network nodes; The migration tendency of pollutants is determined based on the migration potential index, and the direction of pollutant propagation is constrained by the river network topology of the watershed. The migration probability and cumulative migration effect between nodes are then calculated. Based on the migration potential index, the migration probability, and the cumulative migration effect, the pollutant migration path information within the target watershed to be predicted is determined.

[0006] Optionally, acquiring multi-source environmental data of the watershed includes: Collect pollution data within the target watershed, wherein the pollution data includes the concentration of pollutants in water samples and sediments within the watershed; Collect hydrological and meteorological data, including surface air temperature, wind speed, and precipitation rate; Collect water environment characteristic data within the target watershed; Collect sediment characteristic data for the target watershed; the sediment characteristic data includes total organic carbon and sediment grain size composition information; Collect watershed geographic information; the watershed geographic information includes land use data, municipal sewage discharge, surface runoff, and regional heavy metal geochemical background values. Based on the pollution data, hydrological and meteorological data, water environment characteristic data, sediment characteristic data, and watershed geographic information, multi-source environmental data of the watershed is constructed.

[0007] Optionally, the construction of a multi-source pollution database and the establishment of a predictive model for sediment-water interface partition coefficients include: The multi-source environmental data of the watershed are verified and cleaned, the sediment-water interface partition coefficients corresponding to each pollutant node in the watershed space are calculated, and the partition coefficients are used as target variables. The constructed multi-source pollution database and the physicochemical properties of pollutants are combined to build the input dataset for the prediction model; The input dataset is subjected to feature analysis and standardization preprocessing; a machine learning model is trained based on the preprocessed input dataset, and the hyperparameters of the machine learning model are optimized using the Bayesian optimization method to obtain the optimal hyperparameter combination; The machine learning model is retrained based on the optimal hyperparameter combination and its performance is evaluated to determine the optimal training model, thereby constructing a predictive model for sediment-water interface partition coefficients.

[0008] Optionally, the step of constructing a migration potential index characterizing the tendency of pollutants to migrate from the sedimentary phase to the aquatic phase based on the prediction model includes: Obtain the predicted results of the sediment-water environment interface distribution coefficient of the target spatial unit output by the prediction model, as well as the water flow velocity parameters and topographic slope parameters corresponding to the target spatial unit; The predicted results of the sediment-water interface partition coefficient, water flow velocity parameters, and topographic slope parameters are standardized to obtain standardized parameters. Based on the standardized parameters, the pollutant migration potential index of the target spatial unit is obtained by calculating according to the preset migration potential comprehensive calculation function. The pollutant migration potential index is used to characterize the overall tendency of pollutants to migrate from sediments to the water phase and to migrate under hydrodynamic drive. The pollutant migration potential index is jointly determined by sediment-water interface partitioning behavior parameters and hydrodynamic driving parameters.

[0009] Optionally, the construction of the watershed river network topology and the establishment of hydrological connections between river network nodes include: Spatial data of the target watershed is acquired and preprocessed; wherein, the spatial data includes a digital elevation model, watershed boundary data, and river system data; Based on the processed digital elevation model, the flow direction and the cumulative amount of runoff are determined, the watershed river network is extracted, and a node-edge directed topology is constructed. Assign hydrological attributes to the node-edge directed topology and generate a node data table, an edge data table, and a hydrological connection relationship data table. The confluence, divergence, and upstream-downstream relationships are identified based on the number of upstream and downstream connections of nodes, thereby obtaining a complete watershed river network topology. The watershed river network topology includes nodes, edges, flow directions, and hydrological connections between nodes.

[0010] Optionally, the step of determining the pollutant migration tendency based on the migration potential index, and combining this with the river network topology of the watershed to constrain the pollutant propagation direction, and calculating the migration probability and cumulative migration effect between nodes, includes: By associating the migration potential index corresponding to each river network node with the topological structure of the river network in the basin, the migration potential attributes of each river network node are obtained, so as to determine the migration tendency of pollutants at each node. Based on the hydrological connections and flow direction between river network nodes, determine the upstream and downstream node sets corresponding to each node, and construct the pollutant propagation path from upstream to downstream; The river sections between upstream and downstream nodes that meet the preset migration conditions and have hydrological connections are identified as effective migration paths. The direction of pollutant propagation is determined based on the direction of river network flow. The upstream nodes are identified as outgoing nodes and the downstream nodes are identified as incoming nodes. By combining the migration potential index, river network connectivity characteristics, and river segment migration attenuation characteristics, the migration weight of each effective migration path is calculated, and the migration weight is normalized to obtain the migration probability between nodes. Based on the migration probabilities between nodes, the migration probabilities are transferred and dynamically updated along the effective migration path. The cumulative migration effect of pollutants during the propagation process along the river network is calculated, and the cumulative migration probability and affected range of each node are obtained.

[0011] Optionally, determining the pollutant migration path information within the target watershed to be predicted based on the migration potential index, the migration probability, and the cumulative migration effect includes: The migration potential index, migration probability, and cumulative migration effect corresponding to each effective migration path are correlated according to the location to determine the comprehensive migration evaluation result of each effective migration path. Based on the comprehensive migration evaluation results, each of the effective migration pathways is compared to determine the dominant migration pathway of the pollutant. Based on the spatial location, topological connectivity, and propagation direction of the river network nodes and river segments in the dominant migration path, the starting position of pollutant migration, the river network nodes and river segments through which the pollutant migrates, the main migration direction, and the ending position of the migration are determined, thereby generating the pollutant migration path information.

[0012] Optionally, the method further includes the step of generating a watershed pollutant migration risk level map, specifically including: Based on the determined pollutant migration direction, inter-node migration probability and cumulative migration effect, the migration potential, migration impact degree and river network connectivity characteristics of each river network node or river segment are extracted to construct a pollutant migration risk assessment index. Based on each migration risk assessment indicator, the weight of the indicator that affects the degree of pollutant migration risk is determined, and the migration risk assessment indicators are comprehensively calculated to obtain the comprehensive pollutant migration risk value of each river network node or river network segment. Based on the comprehensive risk value of pollutant migration and the preset risk classification threshold, each river network node or river segment is divided into different pollutant migration risk levels. The risk level of each river network node or river segment is associated with the spatial location of the river network in the basin, and the main migration directions and high-risk migration paths are identified to generate a watershed pollutant migration risk level map.

[0013] Another aspect of this application provides a watershed pollutant migration prediction device based on interface assignment and river network topology constraints, comprising: The acquisition module is used to acquire multi-source environmental data of the watershed; The model building module is used to build a multi-source pollution database of pollutants and establish a predictive model for sediment-water interface partition coefficients. The migration potential index construction module is used to construct a migration potential index characterizing the tendency of pollutants to migrate from the sedimentary phase to the aquatic phase based on the prediction model. The hydrological connectivity construction module is used to construct the topological structure of the river network in the watershed and establish the hydrological connectivity between the river network nodes. The migration tendency determination module is used to determine the pollutant migration tendency based on the migration potential index, and to combine the pollutant propagation direction with the network constraints of the river network topology in the watershed, and to calculate the migration probability and cumulative migration effect between nodes. The migration path determination module is used to determine the pollutant migration path information within the target watershed to be predicted based on the migration potential index, the migration probability, and the cumulative migration effect.

[0014] To achieve the above objectives, another aspect of this application proposes an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method described above.

[0015] To achieve the above objectives, another aspect of this application proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the methods described above.

[0016] This application also discloses a computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions, causing the computer device to perform the aforementioned method.

[0017] This application provides at least the following beneficial effects: It offers a method for predicting watershed pollutant migration based on interface allocation and river network topology constraints. This method involves acquiring multi-source environmental data of the watershed; constructing a multi-source pollutant database and establishing a prediction model for the sediment-water interface allocation coefficient; constructing a migration potential index characterizing the tendency of pollutants to migrate from the sediment phase to the water phase based on the prediction model; constructing a watershed river network topology and establishing hydrological connections between river network nodes; determining the pollutant migration tendency based on the migration potential index; and calculating the migration probability and cumulative migration effect between nodes by combining the watershed river network topology with the pollutant propagation direction constraints; finally, determining the pollutant migration path information within the target watershed based on the migration potential index, the migration probability, and the cumulative migration effect. This application enables accurate prediction of the pollution migration process of watershed pollutants, improves the accuracy of pollutant migration path identification and prediction, and thus achieves pollution early warning and control of watershed pollutants. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of an implementation environment provided in an embodiment of this application; Figure 2 This is a flowchart of the overall steps provided in the embodiments of this application; Figure 3 These are specific implementation process diagrams provided in the embodiments of this application; Figure 4 This is a scatter plot of the model performance corresponding to the RF-based machine learning model provided in the embodiments of this application; Figure 5 A scatter plot representing the model performance corresponding to the SVR machine learning model; Figure 6 A scatter plot representing the model performance corresponding to the XGBoost machine learning model; Figure 7 This is a schematic diagram of the hardware structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0019] 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.

[0020] It is understood that 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.

[0021] 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.

[0022] 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.

[0023] 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: 1) Bayesian optimization, an optimization method for automatic hyperparameter optimization of machine learning models, which can improve parameter search efficiency and model calibration effect; 2) Z-score, standard deviation standardization, is a data preprocessing method; 3) R², coefficient of determination, is an indicator used to evaluate the fit of a prediction model; 4) MSE, Mean Squared Error, is an indicator used to evaluate the average squared error of prediction. 5) RMSE, Root Mean Square Error, is an indicator used to evaluate the overall magnitude of prediction error; 6) MAE, Mean Absolute Error, is an indicator used to evaluate the average absolute level of prediction error.

[0024] In related technologies, methods for analyzing the migration pathways of pollutants in watersheds and predicting risks are mostly based on pollutant content characteristics, combined with the spatial distribution of visible content and pollution source apportionment results for risk assessment. Because these methods do not fully consider the interfacial distribution behavior of pollutants and the dynamic water environment characteristics of the watershed, they are difficult to accurately characterize the migration process of pollutants within the watershed system, thus resulting in insufficient prediction accuracy.

[0025] In view of this, this application provides a method for analyzing and predicting watershed pollutant migration paths based on interface allocation behavior and river network topological constraints. This method acquires multi-source environmental data from the watershed, constructs a multi-source pollution database of pollutants, establishes a pollutant sediment-water environment interface allocation coefficient prediction model, and constructs a migration potential index characterizing the tendency of pollutants to migrate from the sediment phase to the water phase based on the allocation coefficient prediction results. On this basis, the migration potential index is embedded into the watershed river network topology, and the river network topological connectivity is used to constrain the migration paths and propagation directions of heavy metals, thereby enabling the prediction of the migration direction, migration path, and migration probability of watershed pollutants, providing technical support for watershed pollutant pollution early warning and risk control.

[0026] The masked autoencoder 3D multi-target tracking method for intelligent connected vehicles provided in this application relates to technical fields such as environmental monitoring, computer technology, model training, machine learning, and environmental data processing. This masked autoencoder 3D multi-target tracking method for intelligent connected vehicles can be applied to terminals, servers, or software running on either a terminal or 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 masked autoencoder 3D multi-target tracking method for intelligent connected vehicles, but is not limited to the above forms.

[0027] This application can be used in a wide variety of 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.

[0028] It should be noted that in all specific embodiments of this application, when processing data related to user identity or characteristics, such as user information, user behavior data, user historical data, and user location information, user permission or consent is obtained first. Furthermore, the collection, use, and processing of this data comply with relevant laws, regulations, and standards. In addition, when embodiments of this application require access to sensitive personal information of users, separate permission or consent from the user is obtained through pop-ups or redirection to confirmation pages. Only after obtaining the user's separate permission or consent is the necessary user-related data required for the proper functioning of these embodiments acquired.

[0029] like Figure 1 The diagram shown is a schematic representation of an implementation environment provided in an embodiment of this application. (Refer to...) Figure 1 The implementation environment includes at least one terminal 102 and a server 101. The terminal 102 and the server 101 can be connected via a network, either wirelessly or via a wired connection, to complete data transmission and exchange.

[0030] Server 101 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides 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 (Content Delivery Network), and big data and artificial intelligence platforms.

[0031] Additionally, server 101 can also be a node server in a blockchain network. Blockchain is a novel application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms.

[0032] Terminal 102 can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, etc. It can also be a vehicle-mounted terminal of the various device types described above, but is not limited to these. Terminal 102 and server 101 can be directly or indirectly connected via wired or wireless communication, and this embodiment does not impose any limitations.

[0033] For example, based on Figure 1 The implementation environment shown in this application embodiment provides a mask autoencoder three-dimensional multi-target tracking method for intelligent connected vehicles. The following description takes the application of this mask autoencoder three-dimensional multi-target tracking method for intelligent connected vehicles in server 101 as an example. It can be understood that this method can also be applied in terminal 102.

[0034] Reference Figure 2 , Figure 2 The flowchart illustrates a three-dimensional multi-target tracking method for a masked autoencoder used in intelligent connected vehicles, applied to a server, as provided in this application embodiment. The execution subject of this method can be any of the aforementioned computer devices (including servers or terminals). See also... Figure 2 The method may include the following steps: S210. Obtain multi-source environmental data of the watershed; S220. Construct a multi-source pollution database for pollutants and establish a predictive model for sediment-water interface partition coefficients; S230. Construct a migration potential index that characterizes the tendency of pollutants to migrate from sedimentary phase to aquatic phase based on the prediction model; S240. Construct the topological structure of the river network in the basin and establish the hydrological connections between the river network nodes; S250. Determine the pollutant migration tendency based on the migration potential index, and combine the pollutant propagation direction with the network constraints of the river network topology in the basin to calculate the migration probability and cumulative migration effect between nodes. S260. Based on the migration potential index, the migration probability, and the cumulative migration effect, determine the pollutant migration path information within the target watershed to be predicted.

[0035] Specifically, the implementation process for step S210 includes: Collect pollution data within the target watershed, wherein the pollution data includes the concentration of pollutants in water samples and sediments within the watershed; Collect hydrological and meteorological data, including surface air temperature, wind speed, and precipitation rate; Collect water environment characteristic data within the target watershed; Collect sediment characteristic data for the target watershed; the sediment characteristic data includes total organic carbon and sediment grain size composition information; Collect watershed geographic information; the watershed geographic information includes land use data, municipal sewage discharge, surface runoff, and regional heavy metal geochemical background values. Based on the pollution data, hydrological and meteorological data, water environment characteristic data, sediment characteristic data, and watershed geographic information, multi-source environmental data of the watershed is constructed.

[0036] Taking heavy metals as a pollutant as an example, step S210 may include: first, collecting heavy metal pollution data, including the concentration of heavy metals in water samples and sediments in the watershed; then, collecting hydrological and meteorological data, including surface air temperature, wind speed, and precipitation rate; next, collecting watershed water environment characteristic data, including pH, dissolved oxygen (DO), chemical oxygen demand (COD), and ammonia nitrogen (NH3-N); then collecting watershed sediment characteristic data, including total organic carbon (TOC) and sediment particle size distribution (clay, silt, gravel); finally, collecting watershed geographic information, including land use data, municipal sewage discharge, surface runoff, and regional heavy metal geochemical background values; wherein the land use data includes nine types of land use data, including farmland, forest, shrubland, grassland, water area, hail, bare land, impermeable surface, and wetland. The heavy metal pollution data, hydrological and meteorological data, water environment characteristic data, sediment characteristic data, and watershed geographic information are used together as multi-source environmental data for the watershed.

[0037] It should be noted that in some embodiments, in the target watershed research of the specific scenario in which this application is applied, the pollutant can be heavy metals in the watershed. In other scenarios, the pollutant can also be other types, such as nutrients, oxygen-consuming organic matter (COD / BOD type, such as protein, oil, carbohydrate), toxic inorganic matter (cyanide, fluoride, acid, alkali, salt), persistent organic pollutants (pesticides, polycyclic aromatic hydrocarbons, polychlorinated biphenyls, etc.). Therefore, this application does not limit the specific type of pollutant. When different types of pollutants are applied to the migration prediction method proposed in this application in actual scenarios, they are all within the protection scope of this application, and will not be elaborated further.

[0038] For step S220 above, the following steps may be included: S2201. Verify and clean the multi-source environmental data of the watershed, calculate the sediment-water environment interface distribution coefficient corresponding to each pollutant node in the watershed space, and use the distribution coefficient as the target variable. S2202. Combine the constructed multi-source pollution database of pollutants with the physicochemical properties of pollution to construct the input dataset for the prediction model; S2203. Perform feature analysis and standardization preprocessing on the input dataset; train a machine learning model based on the preprocessed input dataset, and use Bayesian optimization method to optimize the hyperparameters of the machine learning model to obtain the optimal hyperparameter combination; S2204. Retrain the machine learning model based on the optimal hyperparameter combination and evaluate its performance to determine the optimal training model, so as to construct a prediction model for sediment-water interface partition coefficient.

[0039] Specifically, taking heavy metals as an example, the specific implementation process of step S220 may include: Verification and cleaning of multi-source environmental data for the watershed were performed, and the sediment-water interface partition coefficient (logK) corresponding to each heavy metal node within the watershed was calculated. d The allocation coefficient is then used as the target variable.

[0040] Specifically, the logK d The calculation formula is: Log = , in, The heavy metal content in surface sediments is expressed in μg / kg. The value represents the heavy metal content in the aqueous phase, expressed in μg / L.

[0041] The heavy metal multi-source pollution database and the physicochemical properties of heavy metals are combined and used as the input dataset for the model. The physicochemical properties of heavy metals include atomic mass, electronegativity, atomic radius, ionization energy, electron affinity, ionic radius, and hydration radius. Feature analysis and standardization preprocessing are performed on the input dataset; a machine learning model is trained based on the preprocessed input dataset, and the hyperparameters of the model are optimized using the Bayesian optimization method to obtain the optimal hyperparameter combination; the machine learning model is retrained according to the optimal hyperparameter combination and its performance is evaluated to determine the optimal training model, so as to construct a prediction model for the heavy metal sediment-water interface partition coefficient.

[0042] Step S230 above may include the following implementation steps: S2301. Obtain the prediction results of the sediment-water environment interface distribution coefficient of the target spatial unit output by the prediction model, as well as the water flow velocity parameters and topographic slope parameters corresponding to the target spatial unit. S2302. The predicted results of the sediment-water interface partition coefficient, water flow velocity parameters, and topographic slope parameters are standardized to obtain standardized parameters. S2303. Based on the standardized parameters, calculate according to the preset migration potential comprehensive calculation function to obtain the pollutant migration potential index of the target spatial unit. The pollutant migration potential index is used to characterize the overall tendency of pollutants to migrate from sediments to the water phase and to migrate under hydrodynamic drive. The pollutant migration potential index is jointly determined by sediment-water interface partitioning behavior parameters and hydrodynamic driving parameters.

[0043] Specifically, taking heavy metals as an example of pollutants, the specific implementation process of the above-mentioned S230 is explained in detail as follows: First, obtain the LogK value of the target spatial cell output by the heavy metal sediment-water interface partition coefficient prediction model. d The prediction results, and the water flow velocity parameters and terrain slope parameters corresponding to the target spatial unit; for the LogK d The prediction results and parameter data are standardized to obtain the corresponding standardized parameters. Next, based on the standardized parameters, the heavy metal migration potential index (MPI) of the target spatial unit is calculated according to the preset migration potential comprehensive calculation function. The Heavy Metal Migration Potential Index (MPI) is used to characterize the overall tendency of heavy metals to migrate from sediments to the aquatic phase and to migrate under hydrodynamic driving. The MPI is jointly determined by heavy metal sediment-water interface partitioning behavior parameters and hydrodynamic driving parameters.

[0044] The heavy metal migration potential index (MPI) is calculated according to the following formula: MPI = w 1L+ w 2V+ w 3S; In the formula, L is LogK d The reverse normalization parameters are V (flow velocity normalization parameter) and S (topographic slope normalization parameter). w 1. w 2 and w 3 are LogK d The weighting coefficients corresponding to the inverse normalization parameters, velocity normalization parameters, and slope normalization parameters, and w 1+ w 2+ w 3 = 1.

[0045] In some embodiments, step S240 above may include the following steps: S2401. Acquire spatial data of the target watershed and preprocess the spatial data; wherein, the spatial data includes a digital elevation model, watershed boundary data, and river system data; S2402. Based on the processed digital elevation model, determine the flow direction and the cumulative flow, extract the watershed river network, and construct the node-edge directed topology. S2403. Assign hydrological attributes to the node-edge directed topology and generate a node data table, an edge data table, and a hydrological connection relationship data table. S2404. Identify confluence relationships, divergence relationships, and upstream-downstream relationships based on the number of upstream and downstream connections of nodes, thereby obtaining a complete watershed river network topology; wherein, the watershed river network topology includes nodes, edges, flow directions, and hydrological connections between nodes.

[0046] In some embodiments, step S250 may include the following steps: S2501. Correlate the migration potential index corresponding to each river network node with the topological structure of the river network in the basin to obtain the migration potential attribute of each river network node, so as to determine the migration tendency of pollutants at each node. S2502. Based on the hydrological connections and flow direction between river network nodes, determine the upstream and downstream node sets corresponding to each node, and construct the pollutant propagation path from upstream to downstream. S2503. The river section between upstream and downstream nodes that meet the preset migration conditions and have hydrological connection is identified as an effective migration path. The direction of pollutant propagation is determined according to the direction of river network flow. The upstream node is identified as the outgoing node and the downstream node is identified as the incoming node. S2504. Based on the comprehensive migration potential index, river network connectivity characteristics, and river segment migration attenuation characteristics, calculate the migration weight of each effective migration path, and normalize the migration weight to obtain the migration probability between nodes. S2505. Based on the migration probability between nodes, the migration probability is transferred and dynamically updated along the effective migration path to calculate the cumulative migration effect of pollutants during the propagation process along the river network, and to obtain the cumulative migration probability and affected range of each node.

[0047] In some embodiments, step S260 above may include the following steps: S2601. The migration potential index, migration probability and cumulative migration effect corresponding to each effective migration path are correlated according to the location to determine the comprehensive migration evaluation result of each effective migration path. S2602. Based on the comprehensive migration evaluation results, compare each of the effective migration paths to determine the dominant migration path of the pollutant; S2603. Based on the spatial location, topological connection relationship and propagation direction of the river network nodes and river segments in the dominant migration path, determine the starting position of the pollutant migration, the river network nodes and river segments through which the pollutant migrates, the main migration direction and the migration termination position, and generate the pollutant migration path information.

[0048] The comprehensive migration evaluation results of the effective migration paths use river network nodes as associated points, and correlate the migration potential index, migration probability, and cumulative migration effect corresponding to the same point to form a point migration characteristic sequence for each effective migration path. The dominant migration path is determined by comparing different effective migration paths and determining their migration priority according to the direction of pollutant propagation from upstream to downstream. Specifically, effective migration paths with high comprehensive migration evaluation results and continuous upstream-downstream topological connections are prioritized for retention; effective migration paths with comprehensive migration evaluation results below the preset path determination criteria are eliminated or their path priority is reduced, thereby obtaining the main propagation routes of pollutants within the target watershed.

[0049] Additionally, in some embodiments, step S270 may be included: generating a watershed pollutant migration risk level map, specifically including: S2701. Based on the determined pollutant migration direction, inter-node migration probability and cumulative migration effect, extract the migration potential, migration impact degree and river network connectivity characteristics of each river network node or river segment, and construct pollutant migration risk assessment indicators. S2702. Based on each migration risk assessment indicator, determine the weight of the indicator that affects the degree of pollutant migration risk, and perform a comprehensive calculation on the migration risk assessment indicators to obtain the comprehensive pollutant migration risk value for each river network node or river network segment. S2703. Based on the comprehensive risk value of pollutant migration and the preset risk classification threshold, each river network node or river segment is divided into different pollutant migration risk levels. S2704. Associate the risk level of each river network node or river segment with the spatial location of the river network in the basin, identify the main migration direction and high-risk migration path, and generate a river basin pollutant migration risk level map.

[0050] In summary, the method provided in this application can accurately predict the migration paths of pollutants in a target watershed and analyze the corresponding pollutant migration risks. After obtaining the pollutant migration paths using the method described in this application, it can be specifically applied in the following related fields, depending on the actual application: 1) Precise water environment management and source apportionment: By identifying the main contributors and migration paths of pollution sources, we can guide the allocation of emission reduction targets and the layout of ecological interception facilities, such as setting up buffer zones and ecological ditches for pollutants.

[0051] 2) Emergency response to sudden water pollution incidents: By predicting the diffusion path, arrival time and peak concentration of leaked substances (such as oil and heavy metals), it can support cross-regional joint prevention and control, water conservancy scheduling for interception of sewage and damage assessment.

[0052] 3) Drinking water safety and risk early warning: The method of this application is used to assess the fate of pollutants in water sources, construct a digital twin watershed to realize advanced water quality forecasting, and prevent the risk of algal blooms and exposure to new pollutants (such as antibiotics and microplastics).

[0053] 4) Environmental impact assessment of major water conservancy projects and planning: The method in this application can be used to predict changes in the transport pattern of pollutants in the original water system caused by water diversion projects and reservoir construction, and to assess secondary risks such as downstream water quality deterioration, sediment accumulation and soil salinization.

[0054] 5) Ecological Damage Assessment and Restoration Decision-Making: Quantifying the scope of pollution and ecological loss, determining the necessity and boundaries of projects such as sediment dredging and topsoil remediation, and providing a scientific basis for environmental justice.

[0055] The following, with reference to the accompanying drawings, uses heavy metals as an example to illustrate the specific implementation process of the watershed pollutant migration prediction method based on interface allocation and river network topology constraints proposed in this application: refer to Figure 3 The following is a detailed implementation process diagram of this application, illustrating the specific implementation process of this embodiment. In step S6, the migration trend of heavy metals can be determined, thereby determining the migration trajectory of heavy metals. At the same time, the migration risk distribution of pollutants can also be constructed. The steps are integrated and explained together here. S1. Collect multi-source environmental data on heavy metals in the watershed: Collect environmental factors that influence the distribution of heavy metals in sediment-water interface in the watershed, including heavy metal pollution data, hydrological and meteorological data, water environment characteristic data, sediment characteristic data, and watershed geographic information; Specifically, the raw data used in this embodiment includes, but is not limited to, the concentration of heavy metals in water samples and sediments, surface air temperature, wind speed, precipitation rate, pH, dissolved oxygen (DO), chemical oxygen demand (COD), ammonia nitrogen (NH3-N), organic carbon (TOC), sediment particle size distribution (clay, silt, gravel), land use data, municipal sewage discharge, surface runoff, and geochemical background values ​​of heavy metals.

[0056] S2. Calculate the LogK of each heavy metal node in the watershed space. d Value and construct a multi-source heavy metal pollution database; S2.1 Verification and cleaning of multi-source environmental data in the watershed; the verification and cleaning of multi-source environmental data in the watershed includes unifying data units, correcting data types, handling duplicate and missing values, and identifying and handling outliers.

[0057] Specifically, the unit for heavy metal concentration in water samples was standardized to μg / L, while for sediments it was standardized to mg / kg. For cases where data on certain pollutants in water samples or surface sediments were recorded as "below the limit of detection (LOQ)" or "not detected (ND)" in the literature, data below the LOQ were replaced with half of the LOQ value, and data not detected (ND) were replaced with 0. Outliers were removed using the interquartile range (IQR) method.

[0058] S2.2 Calculate the sediment-water interface partition coefficient logK corresponding to each heavy metal node in the watershed space. d value; Specifically, logK d The calculation formula is: Log =

[0059] in The heavy metal content in surface sediments is expressed in μg / kg. The value represents the heavy metal content in the aqueous phase, expressed in μg / L.

[0060] S3. Establish the heavy metal sediment-water interface partition coefficient (LogK). d A prediction model was developed, and a migration potential index characterizing the tendency of heavy metals to migrate from sedimentary facies to aquatic facies was constructed. S3.1. Combine the heavy metal multi-source pollution database and the physicochemical properties of heavy metals as the input dataset for the model to determine the target area for prediction as heavy metals; Specifically, the physicochemical properties of the heavy metals include atomic mass, electronegativity, atomic radius, ionization energy, electron affinity, ionic radius, and hydration radius, which can reflect the influence of the physicochemical properties of heavy metals on their adsorption, desorption, and partitioning behavior between sediments and water bodies.

[0061] S3.2. Perform model preprocessing on the training data; the model preprocessing includes feature analysis and data standardization.

[0062] S3.2.1 Based on the constructed multi-source heavy metal pollution database, Pearson correlation analysis is used to assess the linear correlation between various characteristic variables and between characteristic variables and the target variable, in order to characterize the intrinsic associations among variables and their relationship with core operating parameters. The formula for calculating the Pearson correlation coefficient is: ; In the formula: and For variables and variables In the The values ​​taken from each sample; and Variables and variables The mean of the Pearson correlation coefficient. The range of values ​​is , Indicates a positive correlation. Indicates a negative correlation. The closer the absolute value is to 1, the stronger the linear correlation.

[0063] S3.2.2 The Z-score normalization method is used to preprocess the heavy metal multi-source pollution database to eliminate differences in physical dimensions and numerical ranges among different indicators. The calculation formula is: ; In the formula: The original value; This is the mean of the indicator; This represents the standard deviation of the indicator. These are the standardized values.

[0064] S3.3 In this embodiment, Support Vector Regression (SVR), Random Forest, and Extreme Gradient Boosting (XGBoost) are selected as machine learning models. Bayesian hyperparameter optimization is performed on the three machine learning models respectively to obtain the logK coefficient used to predict the heavy metal sediment-water interface partitioning coefficient. d The optimal combination of model parameters was determined; based on the optimal hyperparameter combination corresponding to each model, SVR, RF and XGBoost heavy metal sediment-water interface partition coefficient prediction models were constructed respectively, and the performance of each model was evaluated and verified using a test set.

[0065] Specifically, this embodiment uses the coefficient of determination R², mean square error (MSE), root mean square error (RMSE), and mean absolute error (MAE) as evaluation indicators. Their calculation formulas are as follows: R² = 1 -

[0066] MSE =

[0067] RMSE= =

[0068] MAE=

[0069] Among them, y i This represents the true value of the i-th sample. i This represents the predicted value of the i-th sample. R² represents the average of the true values, and n represents the sample size. The closer R² is to 1, the better the model fit and predictive performance; the smaller MSE, RMSE, and MAE are, the smaller the deviation between the predicted and true values.

[0070] Figure 4 The three machine learning models provided in this embodiment are used to evaluate the LogK distribution coefficient of heavy metal sediment-water interface. d The prediction results, among which, Figure 4 A scatter plot representing the model performance of a RF-based machine learning model. Figure 5 A scatter plot representing the model performance corresponding to the SVR machine learning model. Figure 6 This is a scatter plot representing the performance of the XGBoost-based machine learning model. Figure 4-6 It can be seen that all three models can capture LogK d The variation patterns were observed, and the R² values ​​on the test set were all above 0.76, but the prediction performance varied.

[0071] The RF model's R², RMSE, and MAE on the training set are 0.9690, 0.1537, and 0.1095, respectively, while on the test set they are 0.7844, 0.4079, and 0.2907, indicating that its test error is larger than its training error. The SVR model's R², RMSE, and MAE on the training set are 0.9212, 0.2450, and 0.1809, respectively, while on the test set they are 0.7643, 0.4265, and 0.3186, respectively, showing overall lower prediction performance than both the RF and XGBoost models. The XGBoost model's R², RMSE, and MAE on the training set are 0.9823, 0.1160, and 0.0851, respectively, while on the test set they are 0.8263, 0.3661, and 0.2644, respectively, outperforming both the RF and SVR models in all evaluation metrics. Therefore, the XGBoost model was determined to be the optimal model for predicting the LogKd partition coefficient of heavy metal sediments-water interface.

[0072] S3.4 Establish the heavy metal sediment-water interface partition coefficient (LogK) d Predictive model: Obtain the LogK value of the target spatial cell output by the heavy metal sediment-water interface partition coefficient prediction model. d The prediction results, and the water flow velocity parameters and terrain slope parameters corresponding to the target spatial unit; for the LogK dThe prediction results and parameter data are standardized to obtain the corresponding standardized parameters. Based on the standardized parameters, the heavy metal migration potential index (MPI) of the target spatial unit is obtained by calculating according to the preset migration potential comprehensive calculation function. The Heavy Metal Migration Potential Index (MPI) is used to characterize the overall tendency of heavy metals to migrate from sediments to the aquatic phase and to migrate under hydrodynamic driving. The MPI is jointly determined by heavy metal sediment-water interface partitioning behavior parameters and hydrodynamic driving parameters.

[0073] Furthermore, the heavy metal migration potential index (MPI) is calculated according to the following formula: MPI = w 1L+ w 2V+ w 3S; In the formula, L is LogK d The reverse normalization parameters are V (flow velocity normalization parameter) and S (topographic slope normalization parameter). w 1. w 2 and w 3 are LogK d The weighting coefficients corresponding to the inverse normalization parameters, velocity normalization parameters, and slope normalization parameters, and w 1+ w 2+ w 3 = 1.

[0074] S4. Construction of river network topology and determination of hydrological connectivity based on target watershed spatial data; S4.1 Acquire spatial data of the target watershed and perform preprocessing: Specifically, spatial data of the target watershed is collected, including digital elevation model, watershed boundary and river system data, and the data is preprocessed, including coordinate unification, spatial clipping, resolution matching and outlier removal.

[0075] S4.2. Based on the digital elevation model, extract the river network of the target watershed, construct a node-edge directed topology, assign hydrological attributes to the node-edge directed topology, and generate a node data table, an edge data table, and a hydrological connection relationship data table. S4.2.1 Calculate the flow direction and cumulative flow based on the digital elevation model, extract the river network of the target watershed, and generate vector river segments; S4.2.2 Set up topological nodes at river sources, confluences, branching points, control sections, reservoir inlets and outlets, and basin outlets, and define the river segments between adjacent nodes as topological edges; determine the flow direction of each topological edge based on the flow direction, elevation, and upstream-downstream spatial relationship, and establish a directed connection relationship from upstream nodes to downstream nodes; S4.2.3 Assign numbers and hydrological attribute information to the nodes and topological edges respectively to form a node table, an edge table and a hydrological connection relationship table; S4.3. Based on the number of upstream and downstream connections of nodes, identify confluence, diversion, and upstream-downstream relationships to obtain a complete watershed river network topology. The watershed river network topology includes nodes, edges, flow directions, and hydrological connections between nodes. S5. Utilize watershed topology to constrain the direction of heavy metal propagation and identify the migration direction and probability of pollutants between upstream and downstream areas; S5.1. Correlate the heavy metal migration potential index corresponding to each river network node with the river network topology of the basin to obtain the migration potential attribute of each river network node, which is used to characterize the migration tendency of heavy metals in each node. Specifically, based on the constructed river network topology, the upstream and downstream connections, flow directions, confluence and divergence relationships between each topological node and edge are obtained, and the heavy metal migration potential index is matched to the corresponding node or river segment in the river network topology. S5.2 Based on the hydrological connections and flow direction between river network nodes, determine the upstream and downstream node sets corresponding to each node, and construct the heavy metal propagation path from upstream to downstream; S5.2.1 Using the directed edges in the topological network as spatial constraints for pollutant migration, heavy metal pollutants are restricted to migrate only from upstream nodes to downstream nodes along the water flow direction, excluding reverse migration paths, non-connected migration paths, and cross-basin migration paths that do not conform to the upstream and downstream hydrological connection relationship. S5.2.2 For any heavy metal node, search its downstream reachable nodes and downstream reachable river segments in sequence according to the flow direction relationship to form candidate migration paths for pollutants; S5.3. Based on the comprehensive migration potential index, river network connectivity characteristics, and river segment migration attenuation characteristics, calculate the migration weight of each effective migration path, and normalize the migration weight to obtain the migration probability between nodes. S5.3.1 For each effective migration river segment, calculate the migration weight between nodes by comprehensively considering the heavy metal migration potential index of the outgoing nodes, the river network connectivity characteristics between upstream and downstream nodes, and the migration attenuation characteristics of the corresponding river segment.

[0076] Specifically, the river network connectivity characteristics include river segment flow, velocity, slope, length, and diversion ratio; the migration attenuation characteristics include hydraulic residence time, sedimentation, adsorption, dilution, and pollutant migration retention coefficient.

[0077] S5.3.2 At the confluence node, the pollution contribution ratio of different upstream nodes to the confluence node is determined based on the flow rate and heavy metal concentration of each upstream tributary. At the divergence node, the distribution ratio of pollutants to different downstream nodes is determined based on the flow distribution ratio of each downstream branch. S5.4 Based on the migration probability between nodes, the migration probability is transferred and dynamically updated along the effective migration path. The cumulative migration effect of heavy metals in the process of propagation along the river network is calculated to obtain the cumulative migration probability and affected range of each node.

[0078] S5.4.1. Following the topological order of the river network nodes, the migration probability and migration retention coefficient between nodes corresponding to each effective migration segment are passed down step by step from upstream to downstream.

[0079] S5.4.2 Further, the migration probability is corrected by combining the measured concentration changes between upstream and downstream nodes, hydrological transmission distance and water body residence time, so as to identify the migration probability of heavy metal pollutants between upstream and downstream of the watershed.

[0080] S5.4.3. River network nodes with cumulative migration probabilities not less than the preset impact probability threshold are identified as affected nodes, and the corresponding effective migration river sections are identified as pollution-affected river sections. When the hydrological conditions of the river section or the migration potential of the nodes change, the corresponding migration weights and migration probabilities between nodes are updated, and the probability transfer calculation is performed again to obtain the updated heavy metal migration results.

[0081] S6. Identify heavy metal migration trends and evaluate the risk level of nodal river sections, and construct a watershed migration risk distribution map; S6.1 Based on the determined migration direction of heavy metals, migration probability between nodes and cumulative migration effect, extract the migration potential, migration impact degree and river network connectivity characteristics of each river network node or river segment, and construct a heavy metal migration risk assessment index. Specifically, based on the identified migration direction, migration path, node impact probability, and path migration probability of heavy metal pollutants, the heavy metal migration trend information corresponding to each topological node and river segment is obtained. Combined with the heavy metal concentration value, concentration exceeding multiple, pollutant accumulation impact probability, pollution contribution ratio, hydrological transmission distance, and pollutant attenuation characteristics of each node or river segment, a comprehensive risk assessment index for characterizing the degree of heavy metal migration risk is constructed. S6.2 Determine the weight of each risk assessment indicator based on its impact on the risk of heavy metal migration, and perform a comprehensive calculation on the risk assessment indicators to obtain the comprehensive risk value of heavy metal migration for each river network node or river network segment. Specifically, the comprehensive risk assessment indicators are standardized, and the concentration risk, migration probability risk, cumulative impact risk and sensitive receptor exposure risk are weighted according to preset weights to obtain the heavy metal migration risk value of each node or river section. S6.3. Based on the comprehensive risk value of heavy metal migration and the preset risk classification threshold, each river network node or river network segment is divided into different heavy metal migration risk levels. Specifically, according to the preset risk level classification rules, the heavy metal migration risk value is divided into low risk, medium risk, relatively high risk and high risk levels, or into first-level risk, second-level risk, third-level risk and fourth-level risk levels. S6.4. Spatially associate the risk levels with the river network topology of the basin, and classify and label each node and river segment to construct a distribution map of heavy metal migration risk levels in the basin.

[0082] S6.4.1 Spatial association between the risk level results of each node or river segment and the river network topology of the basin, so that each topological node and each river segment corresponds to a unique risk level attribute. S6.4.2. Based on the risk level attributes, the river network of the basin is drawn in a hierarchical manner, and a river basin heavy metal migration risk level distribution map is generated according to the direction of water flow, migration path and risk value, which characterizes the migration trend and spatial distribution of heavy metal pollutants. S6.4.3. Based on the risk level distribution map, the differences in pollution migration risk, the potential impact range of pollutants, key risk transmission paths, and high-risk confluence areas of different river sections and nodes are displayed, thereby realizing the spatial expression and hierarchical identification of the migration risk of heavy metal pollution in the watershed.

[0083] In summary, this application provides a method for predicting watershed pollutant migration paths based on interface allocation behavior and river network topology constraints. The method includes acquiring multi-source environmental data of the watershed and verifying and cleaning this data to construct a multi-source pollutant database. Based on this database, a pollutant sediment-water interface allocation coefficient prediction model is established, and a migration potential index characterizing the tendency of pollutants to migrate from the sediment phase to the water phase is further constructed. A river network topology is constructed based on the watershed river network spatial data to determine the upstream and downstream relationships and hydrological connectivity between river network nodes. Subsequently, the migration potential index characterizes the pollutant migration tendency, and the river network topology connectivity constrains the direction of heavy metal propagation, calculating the migration probability and cumulative migration effect between nodes. Finally, based on the migration potential, migration probability, and cumulative migration effect, the pollutant migration path within the watershed is determined. This application achieves accurate prediction of the watershed pollutant migration process by establishing a pollutant sediment-water interface allocation behavior prediction model and analyzing the pollutant migration direction, migration path, and spatial distribution in conjunction with the watershed river network topology.

[0084] Another aspect of this application embodiment provides a watershed pollutant migration prediction device based on interface allocation and river network topology constraints, including: The acquisition module is used to acquire multi-source environmental data of the watershed; The model building module is used to build a multi-source pollution database of pollutants and establish a predictive model for sediment-water interface partition coefficients. The migration potential index construction module is used to construct a migration potential index characterizing the tendency of pollutants to migrate from the sedimentary phase to the aquatic phase based on the prediction model. The hydrological connectivity construction module is used to construct the topological structure of the river network in the watershed and establish the hydrological connectivity between the river network nodes. The migration tendency determination module is used to determine the pollutant migration tendency based on the migration potential index, and to combine the pollutant propagation direction with the network constraints of the river network topology in the watershed, and to calculate the migration probability and cumulative migration effect between nodes. The migration path determination module is used to determine the pollutant migration path information within the target watershed to be predicted based on the migration potential index, the migration probability, and the cumulative migration effect.

[0085] 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.

[0086] 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 above-described mask autoencoder three-dimensional multi-target tracking method for intelligent connected vehicles. This electronic device can be any intelligent terminal, including tablet computers, in-vehicle computers, etc.

[0087] It is understood that the content of the above method embodiments is applicable to this device embodiment. The specific functions implemented by this device embodiment 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] Please see Figure 7 , Figure 7 The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes: The processor 501 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 502 can be implemented as a read-only memory (ROM), static storage device, dynamic storage device, or random access memory (RAM). The memory 502 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 502 and is called and executed by the processor 501 to execute the mask autoencoder three-dimensional multi-target tracking method for intelligent networking embodiments of this application. The input / output interface 503 is used to implement information input and output; The communication interface 504 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 505 transmits information between various components of the device (e.g., processor 501, memory 502, input / output interface 503, and communication interface 504); The processor 501, memory 502, input / output interface 503, and communication interface 504 are connected to each other within the device via bus 505.

[0089] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described mask autoencoder three-dimensional multi-target tracking method for intelligent connected vehicles.

[0090] 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.

[0091] 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.

[0092] 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. Those skilled in the art will know that 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.

[0093] 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.

[0094] 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.

[0095] 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.

[0096] 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.

[0097] 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.

[0098] 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.

[0099] 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.

[0100] 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 watershed pollutant migration prediction method based on interface allocation and river network topology constraints, characterized in that, include: Acquire multi-source environmental data of the watershed; Construct a multi-source pollution database and establish a predictive model for sediment-water interface partition coefficients; Based on the prediction model, a migration potential index characterizing the tendency of pollutants to migrate from sedimentary phase to aquatic phase is constructed; Construct a river network topology for the watershed and establish hydrological connections between river network nodes; The migration tendency of pollutants is determined based on the migration potential index, and the direction of pollutant propagation is constrained by the river network topology of the watershed. The migration probability and cumulative migration effect between nodes are then calculated. Based on the migration potential index, the migration probability, and the cumulative migration effect, the pollutant migration path information within the target watershed to be predicted is determined. The construction of a multi-source pollution database and the establishment of a predictive model for sediment-water interface partition coefficients include: The multi-source environmental data of the watershed are verified and cleaned, the sediment-water interface partition coefficients corresponding to each pollutant node in the watershed space are calculated, and the partition coefficients are used as target variables. The constructed multi-source pollution database and the physicochemical properties of pollutants are combined to build the input dataset for the prediction model; The input dataset is subjected to feature analysis and standardization preprocessing; a machine learning model is trained based on the preprocessed input dataset, and the hyperparameters of the machine learning model are optimized using the Bayesian optimization method to obtain the optimal hyperparameter combination; The machine learning model is retrained based on the optimal hyperparameter combination and its performance is evaluated to determine the optimal training model, so as to construct a predictive model for sediment-water interface partition coefficients. The construction of a migration potential index characterizing the tendency of pollutants to migrate from sedimentary facies to aquatic facies based on the prediction model includes: Obtain the predicted results of the sediment-water environment interface distribution coefficient of the target spatial unit output by the prediction model, as well as the water flow velocity parameters and topographic slope parameters corresponding to the target spatial unit; The predicted results of the sediment-water interface partition coefficient, water flow velocity parameters, and topographic slope parameters are standardized to obtain standardized parameters. Based on the standardized parameters, the pollutant migration potential index of the target spatial unit is obtained by calculating according to the preset migration potential comprehensive calculation function. The pollutant migration potential index is used to characterize the overall tendency of pollutants to migrate from sediments to the water phase and to migrate under hydrodynamic drive. The pollutant migration potential index is jointly determined by sediment-water interface partitioning behavior parameters and hydrodynamic driving parameters. The step of determining the pollutant migration tendency based on the migration potential index, and combining this with the river network topology of the watershed to constrain the pollutant propagation direction, and calculating the migration probability and cumulative migration effect between nodes, includes: By associating the migration potential index corresponding to each river network node with the topological structure of the river network in the basin, the migration potential attributes of each river network node are obtained, so as to determine the migration tendency of pollutants at each node. Based on the hydrological connections and flow direction between river network nodes, determine the upstream and downstream node sets corresponding to each node, and construct the pollutant propagation path from upstream to downstream; The river sections between upstream and downstream nodes that meet the preset migration conditions and have hydrological connections are identified as effective migration paths. The direction of pollutant propagation is determined based on the direction of river network flow. The upstream nodes are identified as outgoing nodes and the downstream nodes are identified as incoming nodes. By combining the migration potential index, river network connectivity characteristics, and river segment migration attenuation characteristics, the migration weight of each effective migration path is calculated, and the migration weight is normalized to obtain the migration probability between nodes. Based on the migration probabilities between nodes, the migration probabilities are transferred and dynamically updated along the effective migration path. The cumulative migration effect of pollutants during the propagation process along the river network is calculated, and the cumulative migration probability and affected range of each node are obtained.

2. The watershed pollutant migration prediction method based on interface allocation and river network topology constraints according to claim 1, characterized in that, The acquisition of multi-source environmental data in the watershed includes: Collect pollution data within the target watershed, wherein the pollution data includes the concentration of pollutants in water samples and sediments within the watershed; Collect hydrological and meteorological data, including surface air temperature, wind speed, and precipitation rate; Collect water environment characteristic data within the target watershed; Collect sediment characteristic data for the target watershed; the sediment characteristic data includes total organic carbon and sediment grain size composition information; Collect watershed geographic information; the watershed geographic information includes land use data, municipal sewage discharge, surface runoff, and regional heavy metal geochemical background values. Based on the pollution data, hydrological and meteorological data, water environment characteristic data, sediment characteristic data, and watershed geographic information, multi-source environmental data of the watershed is constructed.

3. The watershed pollutant migration prediction method based on interface allocation and river network topology constraints according to claim 1, characterized in that, The construction of the watershed river network topology and the establishment of hydrological connections between river network nodes include: Acquire spatial data of the target watershed and preprocess the spatial data; wherein the spatial data includes a digital elevation model, watershed boundary data, and river system data; Based on the processed digital elevation model, the flow direction and cumulative runoff are determined, the watershed river network is extracted, and a node-edge directed topology is constructed. Assign hydrological attributes to the node-edge directed topology and generate a node data table, an edge data table, and a hydrological connection relationship data table. The confluence, divergence, and upstream-downstream relationships are identified based on the number of upstream and downstream connections of nodes, thereby obtaining a complete watershed river network topology. The watershed river network topology includes nodes, edges, flow directions, and hydrological connections between nodes.

4. The watershed pollutant migration prediction method based on interface allocation and river network topology constraints according to claim 1, characterized in that, The step of determining pollutant migration path information within the target watershed to be predicted based on the migration potential index, the migration probability, and the cumulative migration effect includes: The migration potential index, migration probability, and cumulative migration effect corresponding to each effective migration path are correlated according to the location to determine the comprehensive migration evaluation result of each effective migration path. Based on the comprehensive migration evaluation results, each of the effective migration pathways is compared to determine the dominant migration pathway of the pollutant. Based on the spatial location, topological connectivity, and propagation direction of the river network nodes and river segments in the dominant migration path, the starting position of pollutant migration, the river network nodes and river segments through which the pollutant migrates, the main migration direction, and the ending position of the migration are determined, thereby generating the pollutant migration path information.

5. The watershed pollutant migration prediction method based on interface allocation and river network topology constraints according to claim 1, characterized in that, The method also includes the step of generating a watershed pollutant migration risk level map, specifically including: Based on the determined pollutant migration direction, inter-node migration probability and cumulative migration effect, the migration potential, migration impact degree and river network connectivity characteristics of each river network node or river segment are extracted to construct a pollutant migration risk assessment index. Based on each migration risk assessment indicator, the weight of the indicator that affects the degree of pollutant migration risk is determined, and the migration risk assessment indicators are comprehensively calculated to obtain the comprehensive pollutant migration risk value of each river network node or river network segment. Based on the comprehensive risk value of pollutant migration and the preset risk classification threshold, each river network node or river segment is divided into different pollutant migration risk levels. The risk level of each river network node or river segment is associated with the spatial location of the river network in the basin, and the main migration directions and high-risk migration paths are identified to generate a watershed pollutant migration risk level map.

6. An apparatus for implementing the watershed pollutant migration prediction method based on interface allocation and river network topology constraints as described in any one of claims 1-5, characterized in that, include: The acquisition module is used to acquire multi-source environmental data of the watershed; The model building module is used to build a multi-source pollution database of pollutants and establish a predictive model for sediment-water interface partition coefficients. The migration potential index construction module is used to construct a migration potential index characterizing the tendency of pollutants to migrate from the sedimentary phase to the aquatic phase based on the prediction model. The hydrological connectivity construction module is used to construct the topological structure of the river network in the watershed and establish the hydrological connectivity between the river network nodes. The migration tendency determination module is used to determine the pollutant migration tendency based on the migration potential index, and to combine the pollutant propagation direction with the network constraints of the river network topology in the watershed, and to calculate the migration probability and cumulative migration effect between nodes. The migration path determination module is used to determine the pollutant migration path information within the target watershed to be predicted based on the migration potential index, the migration probability, and the cumulative migration effect.

7. An electronic device, characterized in that, Including the processor and memory; The memory is used to store programs; The processor executes the program to implement the method as described in any one of claims 1 to 5.

Images