A river shoreline form prediction method and device

Abstract

The application discloses a river shoreline form prediction method and device. The river shoreline form prediction method comprises the following steps: acquiring a deposition prediction model and an erosion prediction model; training the models; acquiring current shoreline coordinate information, current water level data, current flow data, current rainfall data, current seepage pressure data, current seepage flow data and current geological data; generating fusion features according to the current shoreline coordinate information, the current water level data, the current flow data, the current rainfall data, the current seepage pressure data, the current seepage flow data and the current geological data; inputting the fusion features into the deposition prediction model and / or the erosion prediction model, so as to acquire deposition prediction information and / or erosion prediction information. The application combines neural networks and historical element data of shoreline erosion and deposition to learn and generate corresponding erosion and deposition prediction models, so that the accuracy and reliability of prediction are improved.

Description

Technical Field

[0001] This application relates to the field of river shoreline morphology technology, specifically to a method and device for predicting river shoreline morphology. Background Technology

[0002] Existing water conservancy remote sensing data processing mainly focuses on the analysis of remote sensing images, including shoreline extraction, morphological evolution prediction, and data comparative analysis. However, current processing methods rely solely on single remote sensing data for sampling, analysis, and prediction, leading to unsatisfactory results. With the rapid development of smart water conservancy and the increasing sophistication of water conservancy monitoring systems, more comprehensive and real-time information on water conditions, rainfall, and engineering conditions can be collected. Therefore, combining multi-source data such as water conservancy monitoring data, remote sensing data, geographic imagery data, and geological data for comprehensive analysis has become an important means to improve the prediction of river shoreline erosion and sedimentation.

[0003] Existing technical solutions can only address the analysis of single remote sensing imaging data and cannot achieve comprehensive data processing and analysis of structured, semi-structured, and unstructured data generated from multiple sources. Based on big data systems and machine learning algorithms, this approach aims to comprehensively process and analyze structured, semi-structured, and unstructured data by integrating multiple data types and utilizing advanced analytical techniques to extract valuable information, supporting outcome prediction and decision-making. Summary of the Invention

[0004] The purpose of this invention is to provide a method for predicting river shoreline morphology to at least solve one of the above-mentioned technical problems.

[0005] One aspect of the present invention provides a method for predicting river shoreline morphology, the method comprising:

[0006] Acquire historical coordinate information on shoreline changes, historical water level data, historical flow data, historical rainfall data, historical seepage pressure data, historical seepage data, and historical geological data;

[0007] Obtain siltation prediction models and erosion prediction models;

[0008] The sedimentation prediction model and the erosion prediction model are trained by using historical coordinate information of shoreline changes, historical water level data, historical flow data, historical rainfall data, historical seepage pressure data, historical seepage data, and historical geological data, respectively, so as to obtain trained sedimentation prediction models and trained erosion prediction models.

[0009] Obtain current shoreline coordinates, current water level, current flow rate, current rainfall, current seepage pressure, current seepage flow, and current geological data;

[0010] A fusion feature is generated based on the current shoreline coordinates, current water level, current flow rate, current rainfall, current seepage pressure, current seepage flow, and current geological data.

[0011] The fused features are input into the sedimentation prediction model and / or the erosion prediction model to obtain sedimentation prediction information and / or erosion prediction information.

[0012] Optionally, the sedimentation prediction model and erosion prediction model are trained using historical shoreline change coordinate information, historical water level data, historical flow data, historical rainfall data, historical seepage pressure data, historical seepage data, and historical geological data, respectively, to obtain trained sedimentation prediction models and erosion prediction models, including:

[0013] Obtain historical coordinate information of shoreline changes for training siltation prediction models and historical coordinate information of shoreline changes for training erosion prediction models;

[0014] Acquire historical data on water level, flow rate, rainfall, seepage pressure, seepage, and geological conditions for training purposes.

[0015] The historical coordinate information of shoreline changes used to train the siltation prediction model is preprocessed to obtain a first feature vector.

[0016] The historical data of water level, flow rate, rainfall, seepage pressure, seepage flow, and geological history used in the training are preprocessed to obtain the second feature vector.

[0017] The historical coordinate information of shoreline changes used to train the erosion prediction model is subjected to a third preprocessing step to obtain a third feature vector.

[0018] The siltation prediction model is trained using the first feature vector and the second feature vector.

[0019] The erosion prediction model is trained using the second and third feature vectors.

[0020] Optionally, training the siltation prediction model using the first feature vector and the second feature vector includes:

[0021] The first feature vector and the second feature vector are fused through an implicit spatial coding network to generate a fourth feature vector;

[0022] Obtain the regression mapping network;

[0023] The regression mapping network is trained based on the fourth feature vector to obtain the trained regression mapping network.

[0024] Obtain the sedimentation prediction and labeling results data;

[0025] Features are extracted from the siltation prediction labeling result data to obtain the fifth feature vector;

[0026] The siltation prediction model is trained using the fourth and fifth feature vectors.

[0027] Optionally, training the siltation prediction model using the fourth and fifth feature vectors includes:

[0028] The siltation prediction model and erosion prediction model are pre-trained multiple times using the fourth feature vector.

[0029] After completing multiple rounds of pre-training, the parameters of the fusion module are frozen according to a preset step size;

[0030] The fourth and fifth feature vectors are compared to train the siltation prediction model, thereby obtaining the trained siltation prediction model.

[0031] Optionally, the step of generating fusion features based on the current shoreline coordinate information, current water level data, current flow rate data, current rainfall data, current seepage pressure data, current seepage data, and current geological data includes:

[0032] The current shoreline coordinates, current water level, current flow rate, current rainfall, current seepage pressure, current seepage flow, and current geological data are fused using a regression mapping network to obtain fused features.

[0033] Optionally, before training the sedimentation prediction model and the erosion prediction model, the river shoreline morphology prediction and early warning method based on the BP algorithm further includes:

[0034] The acquired historical data on water level, flow rate, rainfall, seepage pressure, seepage, and geological conditions were denoised.

[0035] Optionally, the filtering and noise reduction processing of the acquired historical water level data, historical flow data, historical rainfall data, historical seepage pressure data, historical seepage data, and historical geological data adopts the following formula:

[0036] ;

[0037] n is the sample index, The sampled signal is h[nk], which is the value of the filter's impulse response at time nk, N is the length of x[n], and M is the length of the filter's impulse response.

[0038] Optionally, the siltation prediction model and / or erosion prediction model is a BP neural network.

[0039] This application also provides a river shoreline morphology prediction device, the river shoreline morphology prediction device comprising:

[0040] The data acquisition module is used to acquire historical coordinate information of shoreline changes, historical water level data, historical flow data, historical rainfall data, historical seepage pressure data, historical seepage data, and historical geological data.

[0041] A model acquisition module, which is used to acquire a siltation prediction model and an erosion prediction model;

[0042] The training module is used to train the sedimentation prediction model and the erosion prediction model respectively using historical coordinate information of shoreline changes, historical water level data, historical flow data, historical rainfall data, historical seepage pressure data, historical seepage data and geological historical data, so as to obtain the trained sedimentation prediction model and the trained erosion prediction model.

[0043] The prediction data acquisition module is used to acquire current shoreline coordinate information, current water level data, current flow rate data, current rainfall data, current seepage pressure data, current seepage data, and current geological data.

[0044] The feature fusion module is used to generate fused features based on the current shoreline coordinate information, current water level data, current flow rate data, current rainfall data, current seepage pressure data, current seepage data, and current geological data.

[0045] A prediction module is used to input fused features into the sedimentation prediction model and / or the erosion prediction model to obtain sedimentation prediction information and / or erosion prediction information.

[0046] Beneficial effects:

[0047] The river shoreline morphology prediction method in this application combines neural networks with historical data on shoreline erosion and sedimentation to generate corresponding erosion and sedimentation prediction models. These models then provide predictions of river shoreline morphology under specific time, water level, rainfall, and engineering conditions. This process not only utilizes historical data and real-time monitoring information but also optimizes model performance through deep learning algorithms, thereby improving the accuracy and reliability of predictions and providing a scientific basis for water conservancy management and decision-making. Attached Figure Description

[0048] Figure 1This is a flowchart illustrating a method for predicting river shoreline morphology according to an embodiment of this application.

[0049] Figure 2 This is a schematic diagram of an electronic device used to implement a river shoreline morphology prediction method according to an embodiment of this application.

[0050] Figure 3 This is a schematic diagram of sample element information processing.

[0051] Figure 4 This is a diagram illustrating the prediction results. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be described in more detail below with reference to the accompanying drawings. In the drawings, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The described embodiments are some, but not all, embodiments of this application. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. The embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0053] like Figure 1 The methods for predicting river shoreline morphology shown include:

[0054] Acquire historical coordinate information on shoreline changes, historical water level data, historical flow data, historical rainfall data, historical seepage pressure data, historical seepage data, and historical geological data;

[0055] Obtain siltation prediction models and erosion prediction models; in this embodiment, there are two types of shoreline morphology changes: one is shoreline deformation after siltation (predicted by the siltation prediction model), and the other is shoreline deformation after being washed away (predicted by the erosion prediction model).

[0056] The sedimentation prediction model and erosion prediction model are trained by using historical coordinate information of shoreline changes, historical water level data, historical flow data, historical rainfall data, historical seepage pressure data, historical seepage data, and historical geological data, respectively, so as to obtain trained sedimentation prediction models and erosion prediction models.

[0057] Obtain current shoreline coordinates, current water level, current flow rate, current rainfall, current seepage pressure, current seepage flow, and current geological data;

[0058] A fusion feature is generated based on the current shoreline coordinates, current water level, current flow rate, current rainfall, current seepage pressure, current seepage flow, and current geological data.

[0059] The fused features are input into the sedimentation prediction model and / or the erosion prediction model to obtain sedimentation prediction information and / or erosion prediction information.

[0060] The river shoreline morphology prediction and early warning method proposed in this application combines neural networks with historical data on shoreline erosion and sedimentation to generate corresponding erosion and sedimentation prediction models. These models then provide predictions of river shoreline morphology under specific time, water level, rainfall, and engineering conditions. This process not only utilizes historical data and real-time monitoring information but also optimizes model performance through deep learning algorithms, thereby improving the accuracy and reliability of predictions and providing a scientific basis for water conservancy management and decision-making.

[0061] In this embodiment, the sedimentation prediction model and erosion prediction model are trained using historical shoreline change coordinate information, historical water level data, historical flow data, historical rainfall data, historical seepage pressure data, historical seepage data, and historical geological data, respectively, to obtain the trained sedimentation prediction model and erosion prediction model, which include:

[0062] Obtain historical coordinate information of shoreline changes for training siltation prediction models and historical coordinate information of shoreline changes for training erosion prediction models;

[0063] Acquire historical data on water level, flow rate, rainfall, seepage pressure, seepage, and geological conditions for training purposes.

[0064] The historical coordinate information of shoreline changes used to train the siltation prediction model is preprocessed to obtain a first feature vector.

[0065] The historical data of water level, flow rate, rainfall, seepage pressure, seepage flow, and geological history used in the training are preprocessed to obtain the second feature vector.

[0066] The historical coordinate information of shoreline changes used to train the erosion prediction model is subjected to a third preprocessing step to obtain a third feature vector.

[0067] The siltation prediction model is trained using the first feature vector and the second feature vector.

[0068] The erosion prediction model is trained using the second and third feature vectors.

[0069] In this embodiment, the difference between the sedimentation prediction model and the erosion prediction model lies in the different ways of acquiring historical coordinate information of shoreline changes. When acquiring historical coordinate information of shoreline changes, the erosion model obtains it through remote sensing data, oblique photography data, geographic data (shoreline), historical water level data, historical flow data, historical rainfall data, historical seepage pressure data (optional), historical seepage data (optional), and historical geological data. The sedimentation model, on the other hand, adds underwater photography data compared to the erosion model.

[0070] On the other hand, the specific training data used for the historical coordinate information of shoreline changes differs. For example, if the historical coordinate information of shoreline changes is based on one year of data, it includes... Figure 3 The data shown includes data used in erosion models, such as the coastline evolving from a1-a7 to b1-b7 (hereinafter referred to as coastline reduction data), as well as data used in siltation models, such as the coastline evolving from b1-b7 to a1-a7 (hereinafter referred to as coastline increase data).

[0071] For example, if there are 100 historical coordinate data points of shoreline change used for training, including 50 data points of shoreline decrease and 50 data points of shoreline increase, then the 50 data points of shoreline decrease are used to train the erosion model, and the 50 data points of shoreline increase are used to train the siltation model.

[0072] The remaining data, such as historical water level data, historical flow data, historical rainfall data, historical seepage pressure data, historical seepage data, and historical geological data, correspond to the data on the day the shoreline decreased or the data on the day the shoreline increased.

[0073] For example, if the data from March 1, 2024, is added to the shoreline and used to train the sedimentation model, then the historical water level data, historical flow data, historical rainfall data, historical seepage pressure data, historical seepage data, and historical geological data from March 1, 2024, are also used to train the sedimentation model.

[0074] In this embodiment, the sedimentation prediction model and erosion prediction model are trained using historical shoreline change coordinate information, historical water level data, historical flow data, historical rainfall data, historical seepage pressure data, historical seepage data, and historical geological data, respectively, to obtain the trained sedimentation prediction model and erosion prediction model, which include:

[0075] Obtain historical coordinate information of shoreline changes for training siltation prediction models and historical coordinate information of shoreline changes for training erosion prediction models;

[0076] Acquire historical data on water level, flow rate, rainfall, seepage pressure, seepage, and geological conditions for training purposes.

[0077] The historical coordinate information of shoreline changes used to train the siltation prediction model is preprocessed to obtain a first feature vector.

[0078] The historical data of water level, flow rate, rainfall, seepage pressure, seepage flow, and geological history used in the training are preprocessed to obtain the second feature vector.

[0079] The historical coordinate information of shoreline changes used to train the erosion prediction model is subjected to a third preprocessing step to obtain a third feature vector.

[0080] The siltation prediction model is trained using the first feature vector and the second feature vector.

[0081] The erosion prediction model is trained using the second and third feature vectors.

[0082] In this embodiment, training the siltation prediction model using the first feature vector and the second feature vector includes:

[0083] The first feature vector and the second feature vector are fused through an implicit spatial coding network to generate a fourth feature vector;

[0084] Obtain the regression mapping network;

[0085] The regression mapping network is trained based on the fourth feature vector to obtain the trained regression mapping network.

[0086] Obtain the sedimentation prediction and labeling results data;

[0087] Features are extracted from the siltation prediction labeling result data to obtain the fifth feature vector;

[0088] The siltation prediction model is trained using the fourth and fifth feature vectors.

[0089] In this embodiment, training the siltation prediction model using the fourth and fifth feature vectors includes:

[0090] The siltation prediction model and erosion prediction model are pre-trained multiple times using the fourth feature vector.

[0091] After completing multiple rounds of pre-training, the parameters of the fusion module are frozen according to a preset step size;

[0092] The fourth and fifth feature vectors are compared to train the siltation prediction model, thereby obtaining the trained siltation prediction model.

[0093] In this embodiment, the training method for the erosion prediction model is the same as that for the sedimentation prediction model described above, except that the data is different, which will not be repeated here.

[0094] In this embodiment, the step of generating fusion features based on the current shoreline coordinates, current water level data, current flow rate data, current rainfall data, current seepage pressure data, current seepage data, and current geological data includes:

[0095] The current shoreline coordinates, current water level, current flow rate, current rainfall, current seepage pressure, current seepage flow, and current geological data are fused using a regression mapping network to obtain fused features.

[0096] In this embodiment, obtaining historical coordinate information of shoreline changes includes:

[0097] Acquire comprehensive remote sensing data and oblique photography data, and extract shoreline data using remote sensing image recognition algorithms;

[0098] Using data from deformation displacement sensors and map data, the shoreline morphology data is normalized to GIS and transformed into GIS coordinate system data that correspond to each other in time and geospatial space.

[0099] The sedimentation change elements are extracted, and the data of sedimentation graphics are extracted by integrating remote sensing data, oblique photography and underwater topographic data through image recognition algorithms. The data fusion method is then used to transform the data into three-dimensional GIS data in terms of time and geospatial dimensions.

[0100] For the process of shoreline changes from a1, a2…an to b1, b2…bn, the GIS coordinate data is converted into relative displacement data.

[0101] It is understood that the aforementioned historical coordinate information on shoreline changes is based on existing technology and will not be elaborated upon here.

[0102] In this embodiment, before training the sedimentation prediction model and the erosion prediction model, the river shoreline morphology prediction and early warning method based on the BP algorithm further includes:

[0103] The acquired historical data on water level, flow rate, rainfall, seepage pressure, seepage, and geological conditions were denoised.

[0104] In this embodiment, the filtering and noise reduction processing of the acquired historical water level data, historical flow data, historical rainfall data, historical seepage pressure data, historical seepage data, and historical geological data adopts the following formula:

[0105] ;

[0106] n is the sample index, The sampled signal is h[nk], which is the value of the filter's impulse response at time nk, N is the length of x[n], and M is the length of the filter's impulse response.

[0107] In this embodiment, the sedimentation prediction information and / or erosion prediction information is a BP neural network.

[0108] The following examples further illustrate this application in detail. It is understood that these examples do not constitute any limitation on this application.

[0109] By comprehensively collecting and preserving relevant data on riverbank morphology, including data from the smart water conservancy monitoring system, remote sensing data, geological data, and geographic data, this structured, semi-structured, and unstructured data is processed, cleaned, and fused using models from the element extraction model platform and machine learning and data analysis algorithms. This yields spatiotemporal element data related to shoreline erosion and siltation, which is then stored in a data warehouse. The prediction model platform uses the spatiotemporal data from the data warehouse, combined with a backpropagation (BP) artificial neural network algorithm, to perform in-depth calculations on shoreline erosion and siltation data, thereby obtaining prediction results under specific time, water conditions, rainfall, and engineering conditions. Specifically, sensor data from the water conservancy platform's database is structured, while image data and online weather forecasts are unstructured. The element extraction model uses existing image or remote sensing extraction algorithms to obtain reference object coordinate information. Machine learning and data analysis algorithms are primarily used for ETL (Extract, Transform, Load) to clean erroneous data.

[0110] The Water Conservancy Digital Base is based on a data warehouse and uses classified storage to efficiently integrate and manage real-time data from the smart water conservancy monitoring system. This data includes: water level data, flow data, rainfall data, seepage pressure data, seepage data, deformation data, as well as remote sensing data, oblique photography, underwater topography, geographic information data, map data, and geological data. The Water Conservancy Digital Base can achieve efficient storage and management of various types of data.

[0111] The data processing and analysis platform, based on big data technology, has built a model platform and an algorithm platform capable of batch calculations on relevant structured, semi-structured, and unstructured data. This platform is primarily used to acquire shoreline morphology and sedimentation data, and to store this data in a data warehouse according to time and spatial dimensions.

[0112] The feature extraction model platform extracts shoreline morphological features by integrating remote sensing data and oblique photography data, and using remote sensing image recognition algorithms to extract shoreline data. Using data from deformation displacement sensors and map data, the platform performs GIS normalization processing on the shoreline morphological data, transforming it into GIS coordinate system data that corresponds to both time and geospatial coordinates. For siltation change features, it integrates remote sensing data, oblique photography, and underwater topography data, extracting siltation graphic data through image recognition algorithms, and using data fusion methods to transform it into three-dimensional GIS data that corresponds to time and geospatial coordinates. For the shoreline change process from a1, a2…an to b1, b2…bn, the platform converts the GIS coordinate data into relative displacement data: the distance moved along the X-axis and the distance moved along the Y-axis. For example, a shoreline is a line, but the line itself cannot be calculated. Line A can be considered as a connection of points a1, a2…an, where a1 is the reference point, and the value of a1 is the corresponding GIS coordinate, which corresponds to the x and y axis values ​​in the reference coordinate system. At time T, the shoreline line is A; at time T+1, the shoreline line is B; the corresponding reference point a1… It then moved to b1.

[0113] For real-time data from sensors such as water level, flow rate, rainfall, seepage pressure, and seepage, standardized data of water level, flow rate, rainfall, seepage pressure, and seepage are generated after data quality processing; the data is aggregated and transformed into time-dimension element data based on the sampling frequency of the shoreline; and geological data is transformed into standardized data items in the time dimension using data normalization methods.

[0114] The prediction model platform utilizes data on riverbank morphology and sedimentation features, and applies a prediction model generated by a BP artificial neural network algorithm to calculate and produce predictions for a future period based on specific input information. These predictions include bank erosion predictions and sedimentation predictions. Sample data processing is as follows. Figure 3 As shown; it is known that the shoreline pattern for month A is composed of line segments a1, a2…a8; the shoreline pattern for month B is composed of line segments b1, b2…b8; the shoreline pattern for month C is composed of line segments c1, c2…c8; the date of month C is later than that of month B, and the date of month B is later than that of month A;

[0115] Divide the shoreline in month A into equal sections a1…an according to the monitoring points, and divide month B into equal sections b1…bn according to month A. Assume that the shoreline change in section a1-a2 corresponds to the shoreline change in section b1-b2; similarly, the shoreline change in section b1-b2 corresponds to the shoreline change in section c1-c2. The shoreline coordinates of section a1-a2 include (a11, a12, a13…a1n), and the shoreline coordinates of section b1-b2 include (b11, b12, b13…b1n). The relative X-axis displacement of b11 relative to a11 is (ab11x, ab12x…ab1nx), and the relative X-axis displacement of b11 relative to a11 is (ab11y, ab12y…ab1ny). Water level data is m1, and flow rate data is m2. Rainfall data m3, seepage pressure data m4, seepage data m5, geological data m6. Using the BP algorithm: input values ​​are a11, a12….a1n, m1, m2, m3, m4, m5, m6; validation results are ab11x, ab12x … ab1nx; input values ​​are a11, a12….a1n, m1, m2, m3, m4, m5, m6; validation results are ab11y, ab12y … ab1ny.

[0116] Similarly: the shoreline coordinate array for segment b1-b2 includes (b11, b12, b13….b1n), and the shoreline coordinate data for segment c1-c2 includes (c11, c12, c13….c1m, water level data m1, flow rate data m2, rainfall data m3, seepage pressure data m4, seepage data m5, and geological data m6). Using the BP algorithm: input values ​​are b11, b12, b13…b1n, m1, m2, m3, m4, m5, m6; verification results are bc11x, bc12x…bc1nx; input values ​​are b11, b12, b13….b1n, m1, m2, m3, m4, m5, m6; verification results are bc11y, bc12y…bc1ny; the shoreline a1-a2… The segments that change from b1 to b2 and the segments that change from b1 to b2 to c1 to c2 are considered as the same set of samples, and are used to learn and generate sedimentation prediction models and erosion prediction models.

[0117] Sedimentation and erosion prediction models are generated using sample groups of shoreline segments a1-a2 changing to b1-b2 and b1-b2 changing to c1-c2. For segment c1-c2, user-defined values ​​for water level (m1), flow rate (m2), rainfall (m3), seepage pressure (m4), seepage flow (m5), and geological data (m6) are used to predict shoreline data for segment d1-d2. Users input m1…m6 and known values ​​[c11, c12, c13], [c12, c13, c14]… [c1n-2, c1n-1, c1n] to predict the relative positions of corresponding shoreline points [d11x, d11y], [d12x, d12y]… [d1nx, d1ny].

[0118] Finally, the relative position data is converted into GIS coordinate points [d1, d11, d12…d2…d3…dn], such as... Figure 4 As shown.

[0119] This application has the following advantages:

[0120] This process integrates multi-source data, including water conservancy monitoring data, remote sensing data, geographic imagery data, and geological data. It involves cleaning and consolidating these structured, semi-structured, and unstructured data, and applying advanced machine learning and data analysis algorithms to extract spatiotemporal elements related to shoreline morphology. Through these extracted data on water conditions, rainfall, engineering conditions, remote sensing, map images, and geology, predictions can be made under different conditions such as heavy rain and floods, providing comprehensive data support for understanding the development trends of shoreline erosion and siltation.

[0121] For smart water conservancy, data on water level, flow, rainfall, seepage pressure, seepage, deformation, as well as various related data such as remote sensing data, oblique photography, geographic information data, map data, and geological data, the BP artificial neural network algorithm is used to generate corresponding erosion and sedimentation prediction models. This enables the prediction and analysis of shoreline erosion and sedimentation trends in rivers and lakes, thereby allowing for early understanding of potential changes and guiding the planning of water conservancy projects, flood control facilities, and ecological protection measures.

[0122] This application also provides a riverbank morphology prediction device, which includes a data acquisition module, a model acquisition module, a training module, a prediction data acquisition module, a feature fusion module, and a prediction module, wherein...

[0123] The data acquisition module is used to acquire historical coordinate information of shoreline changes, historical water level data, historical flow data, historical rainfall data, historical seepage pressure data, historical seepage data, and historical geological data.

[0124] The model acquisition module is used to acquire sedimentation prediction models and erosion prediction models;

[0125] The training module is used to train the sedimentation prediction model and the erosion prediction model using historical coordinate information of shoreline changes, historical water level data, historical flow data, historical rainfall data, historical seepage pressure data, historical seepage data, and historical geological data, respectively, so as to obtain the trained sedimentation prediction model and the trained erosion prediction model.

[0126] The prediction data acquisition module is used to acquire current shoreline coordinates, current water level, current flow rate, current rainfall, current seepage pressure, current seepage flow, and current geological data.

[0127] The feature fusion module is used to generate fused features based on the current shoreline coordinate information, current water level data, current flow rate data, current rainfall data, current seepage pressure data, current seepage data, and current geological data;

[0128] The prediction module is used to input the fused features into the sedimentation prediction model and / or the erosion prediction model, thereby obtaining sedimentation prediction information and / or erosion prediction information.

[0129] It should be noted that the foregoing explanation of the method embodiments also applies to the system of this embodiment, and will not be repeated here.

[0130] This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the computer program to implement the above-described riverbank morphology prediction method.

[0131] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the above-described method for predicting riverbank morphology.

[0132] Figure 2 This is an exemplary structural diagram of an electronic device capable of implementing the riverbank morphology prediction method provided in one embodiment of this application.

[0133] like Figure 2As shown, the electronic device includes an input device 501, an input interface 502, a central processing unit 503, a memory 504, an output interface 505, and an output device 506. The input interface 502, central processing unit 503, memory 504, and output interface 505 are interconnected via a bus 507. The input device 501 and output device 506 are connected to the bus 507 via the input interface 502 and output interface 505, respectively, and thus connected to other components of the electronic device. Specifically, the input device 501 receives input information from the outside and transmits it to the central processing unit 503 via the input interface 502. The central processing unit 503 processes the input information based on computer-executable instructions stored in the memory 504 to generate output information, temporarily or permanently storing the output information in the memory 504, and then transmitting the output information to the output device 506 via the output interface 505. The output device 506 outputs the output information to the outside of the electronic device for user use.

[0134] In other words, Figure 2 The illustrated electronic device may also be implemented as including: a memory storing computer-executable instructions; and one or more processors, which can be coupled when executing the computer-executable instructions. Figure 1 A method for predicting river shoreline morphology.

[0135] In one embodiment, Figure 2 The electronic device shown can be implemented as including: a memory 504 configured to store executable program code; and one or more processors configured to run the executable program code stored in the memory 504 to perform the riverbank morphology prediction method in the above embodiments.

[0136] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0137] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0138] Computer-readable media include both permanent and non-permanent, removable and non-removable media, and information storage can be achieved by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, DVD or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0139] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code, which includes one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutively marked blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or the overall flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0140] In this embodiment, the processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0141] Memory can be used to store computer programs and / or modules. The processor implements various functions of the system / terminal device by running or executing the computer programs and / or modules stored in the memory, and by accessing data stored in the memory. Memory can mainly include a program storage area and a data storage area. The program storage area can store the operating system, application programs required for at least one function (such as sound playback, image playback, etc.), etc.; the data storage area can store data created based on the use of the mobile phone (such as audio data, phonebook, etc.). In addition, memory can include high-speed random access memory, and can also include non-volatile memory, such as hard disks, RAM, plug-in hard disks, smart media cards (SMC), secure digital cards (SD cards), flash cards, at least one disk storage device, flash memory device, or other volatile solid-state storage devices.

[0142] In this embodiment, if the modules / units integrated into the system / terminal device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or system capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. Although this application discloses preferred embodiments as described above, it is not intended to limit this application. Any person skilled in the art can make possible changes and modifications without departing from the spirit and scope of this application. Therefore, the scope of protection of this application should be determined by the scope defined in the claims of this application.

[0143] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0144] Furthermore, it is clear that the word "comprising" does not exclude other units or steps. Multiple units, modules, or systems stated in the system claims may also be implemented by a single unit or overall system through software or hardware.

[0145] Although the present invention has been described in detail above with general descriptions and specific embodiments, modifications or improvements can be made to it, which will be obvious to those skilled in the art. Therefore, all such modifications or improvements made without departing from the spirit of the present invention fall within the scope of protection claimed by the present invention.

Claims

1. A method of predicting a river bank shape, characterized by, The river shoreline morphology prediction method comprises: obtaining shoreline change historical coordinate information, water level historical data, flow historical data, rainfall historical data, seepage pressure historical data, seepage historical data and geological historical data; obtaining a deposition prediction model and an erosion prediction model; training the deposition prediction model and the erosion prediction model respectively based on the shoreline change historical coordinate information, the water level historical data, the flow historical data, the rainfall historical data, the seepage pressure historical data, the seepage historical data and the geological historical data, so as to obtain a trained deposition prediction model and a trained erosion prediction model; obtaining current shoreline coordinate information, current water level data, current flow data, current rainfall data, current seepage pressure data, current seepage data and current geological data; generating fusion features based on the current shoreline coordinate information, the current water level data, the current flow data, the current rainfall data, the current seepage pressure data, the current seepage data and the current geological data; inputting the fusion features into the deposition prediction model and / or the erosion prediction model, so as to obtain deposition prediction information and / or erosion prediction information; training the deposition prediction model and the erosion prediction model respectively based on the shoreline change historical coordinate information, the water level historical data, the flow historical data, the rainfall historical data, the seepage pressure historical data, the seepage historical data and the geological historical data, so as to obtain a trained deposition prediction model and a trained erosion prediction model, which comprises: obtaining shoreline change historical coordinate information for training a deposition prediction model and shoreline change historical coordinate information for training an erosion prediction model; obtaining water level historical data, flow historical data, rainfall historical data, seepage pressure historical data, seepage historical data and geological historical data for training; firstly preprocessing the shoreline change historical coordinate information for training a deposition prediction model, so as to obtain a first feature vector; preprocessing the water level historical data, the flow historical data, the rainfall historical data, the seepage pressure historical data, the seepage historical data and the geological historical data for training, so as to obtain a second feature vector; thirdly preprocessing the shoreline change historical coordinate information for training an erosion prediction model, so as to obtain a third feature vector; training the deposition prediction model based on the first feature vector and the second feature vector; training the erosion prediction model based on the second feature vector and the third feature vector; the training of the deposition prediction model based on the first feature vector and the second feature vector comprises: fusing the first feature vector and the second feature vector through an implicit space coding network, so as to generate a fourth feature vector; obtaining a regression mapping network; training the regression mapping network based on the fourth feature vector, so as to obtain a trained regression mapping network; obtaining deposition prediction label result data; extracting features in the deposition prediction label result data, so as to obtain a fifth feature vector; training the deposition prediction model based on the fourth feature vector and the fifth feature vector; the training of the deposition prediction model based on the fourth feature vector and the fifth feature vector comprises: The siltation prediction model and the erosion prediction model are pre-trained for multiple rounds through the fourth feature vector; After the pre-training for multiple rounds is completed, the parameters of the fusion module are frozen according to a preset step; The siltation prediction model is trained by comparing the fourth feature vector and the fifth feature vector, so as to obtain the trained siltation prediction model.

2. The river bank shape prediction method according to claim 1, wherein The fusion feature is generated according to the current shoreline coordinate information, the current water level data, the current flow data, the current rainfall data, the current seepage pressure data, the current seepage flow data and the current geological data. The current shoreline coordinate information, the current water level data, the current flow data, the current rainfall data, the current seepage pressure data, the current seepage flow data and the current geological data are subjected to feature fusion by using a regression mapping network, so as to obtain the fusion feature.

3. The river bank shape prediction method according to claim 2, wherein Before the siltation prediction model and the erosion prediction model are trained, the river shoreline morphology prediction and early warning method based on the BP algorithm further comprises: The obtained water level historical data, flow historical data, rainfall historical data, seepage pressure historical data, seepage flow historical data and geological historical data are respectively subjected to denoising processing.

4. The river bank shape prediction method according to claim 3, wherein The obtained water level historical data, flow historical data, rainfall historical data, seepage pressure historical data, seepage flow historical data and geological historical data are respectively subjected to filtering denoising processing by using the following formula: ； n is a sample index, is a sampling signal, h[n-k] is a value of an impulse response of the filter at time n-k, N is a length of x[n], and M is a length of an impulse response of the filter.

5. The river bank shape prediction method according to claim 4, wherein The siltation prediction model and / or the erosion prediction model is a BP neural network.

6. A river shoreline form prediction device characterized by comprising: a river shoreline form prediction program that predicts a river shoreline form based on a river flow rate and a river bed material. The river shoreline morphology prediction device comprises: A data acquisition module, configured to acquire shoreline change historical coordinate information, water level historical data, flow historical data, rainfall historical data, seepage pressure historical data, seepage flow historical data and geological historical data; A model acquisition module, configured to acquire a siltation prediction model and an erosion prediction model; A training module, configured to train the siltation prediction model and the erosion prediction model respectively through the shoreline change historical coordinate information, the water level historical data, the flow historical data, the rainfall historical data, the seepage pressure historical data, the seepage flow historical data and the geological historical data, so as to obtain a trained siltation prediction model and a trained erosion prediction model; A prediction data acquisition module, configured to acquire current shoreline coordinate information, current water level data, current flow data, current rainfall data, current seepage pressure data, current seepage flow data and current geological data; A feature fusion module, configured to generate a fusion feature according to the current shoreline coordinate information, the current water level data, the current flow data, the current rainfall data, the current seepage pressure data, the current seepage flow data and the current geological data; A prediction module, configured to input the fusion feature into the siltation prediction model and / or the erosion prediction model, so as to obtain siltation prediction information and / or erosion prediction information; The siltation prediction model and the erosion prediction model are trained respectively by shoreline change historical coordinate information, water level historical data, flow historical data, rainfall historical data, seepage pressure historical data, seepage historical data and geological historical data, so as to obtain the trained siltation prediction model and the trained erosion prediction model, including: obtaining shoreline change historical coordinate information for training the siltation prediction model, and obtaining shoreline change historical coordinate information for training the erosion prediction model; obtaining water level historical data, flow historical data, rainfall historical data, seepage pressure historical data, seepage historical data and geological historical data for training; firstly preprocessing the shoreline change historical coordinate information for training the siltation prediction model, so as to obtain a first feature vector; preprocessing the water level historical data, flow historical data, rainfall historical data, seepage pressure historical data, seepage historical data and geological historical data for training, so as to obtain a second feature vector; thirdly preprocessing the shoreline change historical coordinate information for training the erosion prediction model, so as to obtain a third feature vector; training the siltation prediction model by the first feature vector and the second feature vector; training the erosion prediction model by the second feature vector and the third feature vector; the training of the siltation prediction model by the first feature vector and the second feature vector includes: fusing the first feature vector and the second feature vector through an implicit space coding network, so as to generate a fourth feature vector; obtaining a regression mapping network; training the regression mapping network according to the fourth feature vector, so as to obtain a trained regression mapping network; obtaining siltation prediction label result data; extracting features in the siltation prediction label result data, so as to obtain a fifth feature vector; training the siltation prediction model by the fourth feature vector and the fifth feature vector; the training of the siltation prediction model by the fourth feature vector and the fifth feature vector includes: pre-training the siltation prediction model and the erosion prediction model by the fourth feature vector for multiple rounds; freezing parameters of a fusion module according to a preset step length after the pre-training for multiple rounds is completed; comparing the fourth feature vector and the fifth feature vector to train the siltation prediction model, so as to obtain a trained siltation prediction model.

Images