Method, device and equipment for predicting drifting of suspended sediment at coastal estuary

Abstract

The invention provides a coastal estuary suspended sediment drift prediction method, device and equipment, and relates to the technical field of target tracking, and the method comprises the steps: obtaining a sediment image of a target monitoring area through employing an unmanned plane to carry a synthetic aperture radar; preprocessing the sediment image, and extracting sediment feature information from the preprocessed sediment image; the sediment characteristic information comprises sediment distribution information, sediment concentration, sediment particle size, sediment drift speed, longitude and latitude of sediment diffusion range, sediment shape and sediment surface area; training a preset sediment drift prediction model based on the sediment characteristic information in a preset period, determining a wind power factor and an ocean current factor which influence sediment drift, and obtaining a target sediment drift prediction model; and inputting the sediment characteristic information at the current moment into the target sediment drift prediction model to obtain sediment drift prediction information. According to the method, airborne real-time monitoring and a coastal estuary sediment prediction model are combined, and the precision of sediment drift prediction is improved.

Description

Technical Field

[0001] The present invention relates to the field of target tracking technology, and in particular to a method, device and equipment for predicting the drift of suspended sediment in a coastal estuary. Background Art

[0002] At present, the prediction method of coastal suspended sediment drift based on airborne radar monitoring is in a rapid development stage. Airborne radar has shown great potential in marine environmental monitoring with its high temporal and spatial resolution, large-scale coverage and ability to penetrate clouds. The data obtained by airborne radar, combined with advanced data processing algorithms and ocean dynamics models, can realize real-time monitoring and prediction of coastal suspended sediment drift. In addition, with the development of computer technology and artificial intelligence, algorithms based on big data and machine learning have been widely used in the prediction model of suspended sediment drift, further improving the accuracy and efficiency of prediction.

[0003] Although the monitoring method based on airborne radar has made significant progress in the prediction of coastal suspended sediment drift, there are still some problems and technical defects. First, the accuracy and resolution of airborne radar monitoring data in complex sea conditions may be affected, resulting in errors in the prediction results. In addition, the complexity of marine environmental factors makes it difficult to accurately simulate and predict the drift process of suspended sediment, especially under the interweaving of multiple factors, the accuracy and stability of the prediction model face challenges. Summary of the invention

[0004] The purpose of the present invention is to solve the problem of insufficient accuracy and stability in the current prediction of coastal suspended sediment drift. The present invention provides a method, device and equipment for predicting suspended sediment drift in coastal estuaries.

[0005] The technical solution of the embodiment of the present application is implemented as follows:

[0006] A first aspect of an embodiment of the present application provides a method for predicting the drift of suspended sediment in a coastal estuary, comprising:

[0007] Use UAVs equipped with synthetic aperture radar to obtain sediment images of the target monitoring area;

[0008] Preprocessing the sediment image, and extracting sediment characteristic information from the preprocessed sediment image; the sediment characteristic information includes sediment distribution information, sediment concentration, sediment particle size, sediment drift speed, longitude and latitude of sediment diffusion range, sediment shape and sediment surface area;

[0009] Based on the sediment characteristic information within a preset period, a preset sediment drift prediction model is trained to determine the wind factor and the current factor that affect the sediment drift, and obtain a target sediment drift prediction model;

[0010] The sediment characteristic information at the current moment is input into the target sediment drift prediction model to obtain sediment drift prediction information.

[0011] Optionally, the preprocessing of the sediment image includes: filtering, denoising and geometric correction of the sediment image.

[0012] Optionally, extracting sediment feature information from the pre-processed sediment image includes:

[0013] Using threshold segmentation, edge detection and region growing techniques, the background image is removed from the sediment image to obtain a sediment region image;

[0014] The influencing features in the sediment area image are enhanced by using a histogram equalization and contrast adjustment algorithm;

[0015] A neural network classifier is used to classify and identify the sediment area image after feature enhancement processing to extract sediment feature information.

[0016] Optionally, the preset sediment drift prediction model includes: a tidal module, a sediment module and a wave module; the tidal module is configured to simulate the water flow movement in the target monitoring area, including the water flow changes caused by tides and river runoff, the sediment module is configured to simulate the convection and turbulent diffusion of the water flow, and the wave module is configured to simulate the action of ocean waves.

[0017] Optionally, the tidal flow module includes a momentum equation, a continuity equation, a temperature equation, a salinity equation and a density equation, which are expressed as:

[0018]

[0019] Where x, y, z are the coordinates of the east, north and vertical directions in the Cartesian coordinate system; u, v, w are the flow velocities in the three directions; T is the temperature; S is the salinity; ρ is the density; P a is the sea surface air pressure; P h is the hydrostatic pressure; q is the non-hydrostatic pressure; f is the Coriolis parameter; g is the gravitational acceleration; K m is the vertical eddy viscosity coefficient; K h is the vertical vortex thermal diffusion coefficient; F u , F v , F T , F S are the horizontal and vertical momentum, heat and salinity diffusion terms, respectively;

[0020] The calculation equation corresponding to the sediment module is as follows:

[0021]

[0022] Where i represents the component of sediment; C i , w i Respectively represent the concentration and sedimentation velocity of the i-th component sediment; A h is the horizontal eddy viscosity coefficient; K h is the vertical eddy viscosity coefficient;

[0023]

[0024] In the formula, Q i is the flushing flux, P b is the bottom porosity, F bi is the bottom friction of sediment component i, τ b is the bottom scour stress, τ ci is the critical scour stress at the bottom of component i, D i is the sedimentation flux, E i is the erosion flux;

[0025] The wave module uses the wave action density spectrum to describe the ocean waves. The control equation of the wave action density spectrum is expressed as:

[0026]

[0027] Where N is the wave action density spectrum; t is time; σ is the relative frequency; θ is the wave direction; C σ and C θ are the propagation speeds of waves in the spectral space (σ, θ); C g is the group velocity; V is the ambient water velocity; is the horizontal divergence operator in geographic space; the first term on the left side of the equation represents the local variation of wave action density in time, the second term represents the propagation of wave action density in space, and the third and fourth terms represent the frequency shift and refraction caused by changes in water depth and flow velocity, respectively. tot is the source-sink term, and its formula is as follows:

[0028] S tot =S in +S nl3 +S nl4 +S ds,w +S ds，b +S ds,br ;

[0029] In the formula, S in is the growth function of wind-generated waves, i.e., wind energy input; S nl / 3 is the nonlinear three-wave interaction; S nl / 4 is the nonlinear four-wave interaction; S ds,w is the wave energy dissipation caused by the breaking of the white hat; S ds,bis the wave energy dissipation caused by bottom friction; S ds,br It is the dissipation of wave energy caused by wave breaking due to changes in water depth.

[0030] Optionally, the preset sediment drift prediction model includes a prediction module; the prediction module is represented as:

[0031]

[0032] V L =β0+β1V c +β2V w

[0033] β2 = λT;

[0034] Where C0 is the initial position of the sediment center, Δt is the sediment drift time, C is the final position of the sediment center, V L V is the drift velocity of suspended sediment in coastal estuaries; c is the sea surface velocity, V w is the wind speed at 10 meters above the sea surface, β1, β2, β3 are current factors, λ is the wind factor, and T is the conversion matrix that introduces the drift deflection angle.

[0035] A second aspect of the embodiment of the present application provides a device for predicting the drift of suspended sediment in coastal estuaries, including a first acquisition module, a feature extraction module, a training module, and a second acquisition module, wherein:

[0036] The first acquisition module is configured to acquire sediment images of the target monitoring area using a synthetic aperture radar mounted on an unmanned aerial vehicle;

[0037] The feature extraction module is configured to preprocess the sediment image and extract sediment feature information from the preprocessed sediment image; the sediment feature information includes sediment distribution information, sediment concentration, sediment particle size, sediment drift speed, longitude and latitude of sediment diffusion range, sediment shape and sediment surface area;

[0038] The training module is configured to train a preset sediment drift prediction model based on the sediment characteristic information within a preset period, determine the wind factor and the current factor that affect the sediment drift, and obtain a target sediment drift prediction model;

[0039] The second acquisition module is configured to input the sediment characteristic information at a current moment into the target sediment drift prediction model to obtain sediment drift prediction information.

[0040] A third aspect of an embodiment of the present application provides an electronic device, including a processor and a memory; the memory stores a computer program, wherein the computer program, when executed by the processor, implements the method for predicting suspended sediment drift in coastal estuaries described in the first aspect.

[0041] A fourth aspect of an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in the first aspect.

[0042] Compared with the prior art, the technical solution provided by this application has the following beneficial effects:

[0043] The present invention provides a method, device and equipment for predicting the drift of suspended sediment in coastal estuaries, which can significantly improve the accuracy and reliability of prediction by optimizing data processing algorithms and building a more accurate drift model. By using advanced machine learning algorithms and ocean dynamics principles, airborne radar data is deeply mined and analyzed to provide a basis for more accurately predicting the drift trajectory and distribution range of coastal suspended sediment. When constructing the drift model, the influence of various marine environmental factors on the drift of suspended sediment is fully considered. Through the fusion and comprehensive analysis of multi-source data, the comprehensiveness and accuracy of the prediction are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A schematic diagram of a flow chart of a method for predicting suspended sediment drift in a coastal estuary provided in an embodiment of the present application;

[0045] Figure 2 A technical roadmap for radar image data processing provided for the embodiments of this application;

[0046] Figure 3 A schematic diagram of the principle of airborne small synthetic aperture radar image positioning provided in an embodiment of the present application;

[0047] Figure 4 A schematic diagram of the radar image positioning principle considering the influence of the earth's curvature provided in an embodiment of the present application;

[0048] Figure 5 A schematic diagram of the structure of a device for predicting the drift of suspended sediment in coastal estuaries provided in an embodiment of the present application;

[0049] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0050] Below, embodiments of the present application will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present application. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present application. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present application.

[0051] The terms used herein are only for describing specific embodiments and are not intended to limit the present application. The terms "include", "comprising", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0052] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0053] Some block diagrams and / or flow charts are shown in the accompanying drawings. It should be understood that some blocks or combinations thereof in the block diagrams and / or flow charts may be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that these instructions, when executed by the processor, may create a device for implementing the functions / operations described in these block diagrams and / or flow charts.

[0054] In some embodiments, see Figure 1 , Figure 1 A schematic diagram of a flow chart of a method for predicting the drift of suspended sediment in coastal estuaries provided in an embodiment of the present application; the method for predicting the drift of suspended sediment in coastal estuaries provided in an embodiment of the present application comprises:

[0055] S110 uses a UAV equipped with a synthetic aperture radar to obtain sediment images of the target monitoring area.

[0056] Here, a small synthetic aperture radar can be mounted on an unmanned aerial vehicle to periodically scan and dynamically monitor the suspended sediment situation in the target monitoring area, and obtain radar echo signals and suspended sediment drift image data in real time. The target monitoring area can be a coastal estuary at a specific location.

[0057] S120, preprocessing the sediment image, and extracting sediment characteristic information from the preprocessed sediment image; the sediment characteristic information includes sediment distribution information, sediment concentration, sediment particle size, sediment drift speed, longitude and latitude of sediment diffusion range, sediment shape and sediment surface area.

[0058] In one example, according to the resolution of the small synthetic aperture radar system and the scope of the monitoring area, a suitable flight altitude is set to ensure the high resolution of the small synthetic aperture radar image. Ensure that the drone collects data in a stable flight state to avoid excessive speed that causes data quality to deteriorate. According to the terrain and tidal conditions of the monitoring area, a reasonable flight route is planned to ensure full coverage of the monitoring area. The L-band and C-band working frequencies of the small synthetic aperture radar for marine environment monitoring are used together to obtain better sea surface scattering characteristics and penetration capabilities. To obtain more ground object information, dual polarization is selected. Select the beam imaging mode to provide higher local resolution. According to the monitoring needs and tidal cycle, the flight plan of the drone is set to regularly scan the coastal estuary to obtain real-time suspended sediment drift image data. The entire small synthetic aperture radar system (including radar equipment and antenna) weighs less than 4kg, which is suitable for small drones to carry, can reduce the load pressure of the aircraft, and improve endurance and flexibility. The system resolution is better than 0.5m. The detection distance is 10km, and the X-band (8-12GHz) is selected as the working frequency band. This band has a higher frequency and can provide better resolution and penetration, which is suitable for the detection of various environments and targets. The positive side view (that is, the radar beam is perpendicular to the flight direction of the drone) helps to obtain a true two-dimensional image of the target area and reduce image distortion caused by angle deviation.

[0059] In some embodiments, S120, pre-processing the sediment image includes: filtering, denoising and geometric correction of the sediment image.

[0060] Here, the sea surface sediment concentration detection technology can be used to analyze the radar backscatter coefficient to identify and extract the distribution information of suspended sediment. Image processing and pattern recognition technology are used to extract sediment distribution information, sediment concentration, sediment particle size, sediment drift speed, longitude and latitude of sediment diffusion range, sediment shape and sediment surface area from radar images. Since the pulse signal emitted by the radar will be affected by various factors during the transmission process, the radiation and geometric characteristics of the target will change, and the real information of the target cannot be truly reflected. The coherence of electromagnetic waves will reduce the accuracy of small synthetic aperture radar image recognition, so filtering, denoising and geometric correction are required before the depth interpretation of the sediment image obtained by the small synthetic aperture radar. Exemplarily, Gaussian filtering, median filtering and wavelet transform can be used to reduce random noise and interference that may come from the radar system itself, atmospheric conditions or sea surface clutter data. Noise suppression method based on statistical model is used to further clean up the data. According to the geometric characteristics of the radar system and factors such as the curvature of the earth, the radar image is geometrically corrected to ensure that the geographic coordinates in the image are accurate.

[0061] In some embodiments, S120, extracting sediment feature information from the pre-processed sediment image includes:

[0062] Threshold segmentation, edge detection and region growing techniques are used to remove the background image from the sediment image to obtain the sediment area image.

[0063] The influencing features in the sediment area image are enhanced through histogram equalization and contrast adjustment algorithms;

[0064] A neural network classifier is used to classify and identify the sediment area images after feature enhancement processing to extract sediment feature information.

[0065] In this embodiment, the model-driven MRF segmentation method can be used to extract all dark spot areas in the sediment image as candidate marine suspended sediment areas. This method assumes that the pixel values ​​in each field in the image are independent and the values ​​are similar or the same. It makes full use of the prior knowledge between the pixel field structures, quantitatively calculates the local structure information in the image, and effectively segments the image. Threshold segmentation, edge detection and regional growth techniques are used to remove the background image from the sediment image to obtain the sediment area image. The visibility of the sediment features in the image is enhanced by histogram equalization and contrast adjustment methods. The statistical features of the dark spot area (candidate area) are extracted, mainly including the shape, area, boundary, physical features, backscatter value and associated features of the target in the image of the candidate area, and then a neural network classifier is used for classification and identification based on these statistical features. The characteristic parameters such as sediment concentration, sediment particle size, and sediment drift velocity of the suspended sediment are extracted from the processed image.

[0066] In one example, the extracted sediment diffusion range is superimposed on the GIS map, its location is marked in the form of longitude and latitude, and its shape and area are calculated. In the process of sediment image positioning, because it is very difficult to obtain effective ground control points in the sea area, a direct positioning method without control points is adopted, which specifically includes the following steps:

[0067] According to the geometric relationship of the positive and side-view imaging of the airborne small synthetic aperture radar (such as Figure 2 As shown, Figure 2 The radar image data processing technology roadmap provided in the embodiment of the present application is used to calculate the distance R0 between the near point P0 of the surveying band and the carrier, and the distance R1 between the far point P1 of the surveying band and the carrier. The platform height is h. According to the Pythagorean theorem, the relationship between the height and the distance can be obtained as follows:

[0068] R0=h2secα0 R1=h2secα1;

[0069] In the formula, α0 is the minimum viewing angle, α1 is the maximum viewing angle, which is calculated by the auxiliary parameters of the small synthetic aperture radar image. h is the height of the carrier from the sea level, which is obtained by subtracting the platform altitude measured by the airborne sensor and the average sea level height at the coordinates of the carrier calculated by the marine environment prediction model.

[0070] In an example, see Figure 3 ; Figure 3 Schematic diagram of the principle of airborne small synthetic aperture radar image positioning provided in the embodiment of the present application; θ0 and θ1 are calculated by the following formula:

[0071]

[0072] Then, using the distances D0 and D1 obtained from the above formula, we can calculate the longitude and latitude P0 (L0, B0) and P1 (L1, B1) of the near point P0 and the far point P1 of the surveying band. The calculation formula is as follows:

[0073]

[0074] B0=sin -1 (sin B cos D0+cos B sin D0 cosβ0)

[0075]

[0076] Among them, L and B are the longitude and latitude of the point O under the aircraft, which are obtained by solving the auxiliary parameters of the small synthetic aperture radar image; β0 and β1 are the azimuths of the near point P0 and the far point P1 of the surveying band relative to the point O, which are obtained by the aircraft platform track angle and the radar side-view mode.

[0077] In an example, see Figure 4 , Figure 4 Schematic diagram of the radar image positioning principle considering the influence of the earth's curvature provided in the embodiment of the present application; when the working range and the width of the mapping band are large and the influence of the earth's curvature needs to be considered, the distances from the near point P0 and the far point P1 of the mapping band to the airborne platform are R0 and R1 respectively, and the platform height is h. Using the Pythagorean theorem, the distances from the point O under the aircraft to the points P0 and P1 are obtained as D0 and D1 according to the following formula:

[0078] D0=rcos -1 θ0 D1=rcos -1 θ1;

[0079] In the formula, r is the average radius of the earth, θ0 is the angle between the lines connecting points O and P0 and the center of the earth, and θ1 is the angle between the lines connecting points O and P1 and the center of the earth.

[0080] Then, the obtained distance and azimuth are error verified in combination with the Vincenty formula. The Vincenty formula is an accurate method for calculating the geodesic distance (i.e., the shortest distance) and azimuth between two points on the earth. It takes into account the ellipsoidal shape of the earth and is therefore more accurate than simple spherical trigonometry methods. The Nelder-Mead optimization method is used to optimize P0 (L0, B0) and P1 (L1, B1). The Nelder-Mead simplex method is a direct search method for finding the minimum value of a function in multidimensional space. Optimization process: Initialize the simplex; evaluate the objective function value for each point in the simplex; update the simplex by operations such as reflection, expansion, contraction, or shrinking the boundary. Repeat the above steps until the stopping criterion is met.

[0081] The map scale of the small synthetic aperture radar image can be obtained from the distances D0 and D1 from the under-aircraft point O to points P0 and P1 and the longitudes and latitudes P0 (L0, B0) and P1 (L1, B1) of the near point P0 and the far point P1 in the surveying band. The longitude and latitude information of any point on the small synthetic aperture radar image can be obtained based on the map scale and the longitudes and latitudes of points P0 and P1.

[0082] S130, training a preset sediment drift prediction model based on sediment characteristic information within a preset period, determining wind factors and ocean current factors that affect sediment drift, and obtaining a target sediment drift prediction model.

[0083] Combined with the sediment characteristic information obtained in the above steps and the principles of ocean dynamics, the preset sediment drift prediction model parameters are updated. In this process, the influence of marine environmental factors such as wind field, flow field, tide, and wave on the drift of suspended sediment is considered, and the relevant environmental parameters are incorporated into the model. After iterative training, the target sediment drift prediction model can be obtained.

[0084] In some embodiments, the preset sediment drift prediction model includes: a tidal module, a sediment module and a wave module; the tidal module is configured to simulate the water movement in the target monitoring area, including the water flow changes caused by tides and river runoff, the sediment module is configured to simulate the convection and turbulent diffusion of the water flow, and the wave module is configured to simulate the action of ocean waves.

[0085] In some embodiments, the tidal flow module includes a momentum equation, a continuity equation, a temperature equation, a salinity equation, and a density equation, expressed as:

[0086]

[0087] Where x, y, z are the coordinates of the east, north and vertical directions in the Cartesian coordinate system; u, v, w are the flow velocities in the three directions; T is the temperature; S is the salinity; ρ is the density; P a is the sea surface air pressure; P his the hydrostatic pressure; q is the non-hydrostatic pressure; f is the Coriolis parameter; g is the gravitational acceleration; K m is the vertical eddy viscosity coefficient; K h is the vertical vortex thermal diffusion coefficient; F u , F v , F T , F S are the horizontal and vertical momentum, heat and salinity diffusion terms, respectively;

[0088] The calculation equation corresponding to the sediment module is as follows:

[0089]

[0090] Where i represents the component of sediment; C i , w i Respectively represent the concentration and sedimentation velocity of the i-th component sediment; A h is the horizontal eddy viscosity coefficient; K h is the vertical eddy viscosity coefficient;

[0091]

[0092] In the formula, Q i is the flushing flux, P b is the bottom porosity, F bi is the bottom friction of sediment component i, τ b is the bottom scour stress, τ ci is the critical scour stress at the bottom of component i, D i is the sedimentation flux, E i is the erosion flux; the movement of sediment mainly depends on the magnitude of the bottom stress. When the bottom stress exceeds the critical scouring stress of the sediment, the sediment will be stirred up by the water flow. The horizontal convection, vertical diffusion and scouring of the water flow will increase the mass of the sediment concentration profile, while the sedimentation will reduce the mass of the sediment concentration profile until a balance is reached. The sedimentation flux is expressed by the deposition flux D i The model uses a flux control format to solve the sedimentation equation and adopts a SLIP limiter to add a diffusion term to the restriction format, which can ensure the second-order spatial accuracy and obtain relatively accurate calculation results even in areas where the sediment concentration changes dramatically.

[0093] The wave module uses the wave action density spectrum to describe ocean waves. The governing equation of the wave action density spectrum is expressed as:

[0094]

[0095] Where N is the wave action density spectrum; t is time; σ is the relative frequency; θ is the wave direction; C σ and Cθ are the propagation speeds of waves in the spectral space (σ, θ); C g is the group velocity; V is the ambient water velocity; is the horizontal divergence operator in geographic space; the first term on the left side of the equation represents the local variation of wave action density in time, the second term represents the propagation of wave action density in space, and the third and fourth terms represent the frequency shift and refraction caused by changes in water depth and flow velocity, respectively. tot is the source-sink term, and its formula is as follows:

[0096] S tot =S in +S nl3 +S nl4 +S ds,w +S ds，b +S ds，br ;

[0097] In the formula, S in is the growth function of wind-generated waves, i.e., wind energy input; S nl / 3 is the nonlinear three-wave interaction; S nl / 4 is the nonlinear four-wave interaction; S ds,w is the wave energy dissipation caused by the breaking of the white hat; S ds,b is the wave energy dissipation caused by bottom friction; S ds,br It is the wave energy dissipation caused by wave breaking due to water depth changes. In the horizontal direction, the wave module is discretized using an unstructured grid method, which increases the adaptability of the wave module to complex coastlines. The model introduces the wave breaking mechanism caused by three-wave interaction and water depth changes and adopts a fully implicit finite difference algorithm, so that the wave module is no longer constrained by the CFL condition in the selection of time step, effectively solving the time step limitation problem in nearshore high-resolution calculations.

[0098] In some embodiments, the preset sediment drift prediction model includes a prediction module, and the prediction module is represented as:

[0099]

[0100] Where C0 is the initial position of the sediment center, Δt is the sediment drift time, C is the final position of the sediment center, V L V is the drift velocity of suspended sediment in coastal estuaries; c is the sea surface velocity, V w is the wind speed at 10 meters above the sea surface, β1, β2, β3 are current factors, λ is the wind factor, and T is the conversion matrix that introduces the drift deflection angle.

[0101] In this embodiment, the turbulent diffusion stage is simulated by using the sediment particle method, and the expanded sediment drift range is divided into a finite number of small units according to the volume, and each small unit is a sediment particle, that is, each sediment particle is a part of the sediment drift volume. According to the latitude and longitude information of the sediment drift range obtained dynamically, the longitude and latitude of the location of each sediment particle is determined. The number of sediment particles is determined according to the computer configuration, calculation time, calculation accuracy and other requirements for numerical simulation, and the method of adding volume parameters is used to realize the simulation of the characteristics of sediment particles.

[0102] In an optional embodiment, based on the sediment characteristic information, the variation law of the suspended sediment characteristic parameters over time is analyzed, a characteristic parameter time series is established, and a preset sediment drift prediction model is calibrated and updated in real time for training to determine the wind factor λ and the current factors β1 and β2 that affect sediment drift, as follows:

[0103] According to the drift position information in the sediment characteristic information, the actual drift speed of suspended sediment is calculated as follows:

[0104]

[0105] Among them, C0 n is the initial position of the center of suspended sediment obtained based on sediment image processing, C n is the position of the center of suspended sediment after a period of time obtained based on sediment image processing, Δt is the time difference between two adjacent sediment images, and V L n is the actual drift rate of suspended sediment.

[0106] The actual drift rate V of n (n is a positive integer greater than 1) groups of sediment calculated based on the latest sediment image L n and the wind speed V predicted by the ocean environment w and flow rate V c , the multivariate linear regression model is used to calibrate the wind factor λ and the current factors β1 and β2, as follows:

[0107] V L n Affected by two non-random factors V w and V c , and the influence of random factor ε, and have the following linear relationship:

[0108]

[0109] In the formula, β0, β1, β2 are three unknown parameters, ε is an unobservable random error, and it is assumed that ε∈N(0,σ2), σ2 represents the variance.

[0110] Based on the n sets of actual drift rates of suspended sediment calculated from the latest sediment images and the n sets of wind speed and current speed (V Li n , V ci n , V wi n ), i = 1, 2, ..., n satisfies:

[0111]

[0112] Among them, ε1, ε2, …, ε n Independent of each other and all obey N(0,σ2). Order:

[0113]

[0114] In the formula, Y = Xβ + ε. The sum of squared errors is:

[0115]

[0116] Q(β) is a non-negative quadratic function about β0, β1, β2, and must have a minimum value, which is obtained by the unconditional extreme value:

[0117]

[0118] That is: X T (Y-Xβ)=0;

[0119] That is: β=(X T X) -1 X T Y;

[0120] Thus, β0, β1, and β2 are calculated, and β2 is then substituted into the formula β2=λT to obtain the latest wind factor λ. Through real-time calibration, more accurate wind factor λ and current factors β1 and β2 can be dynamically obtained, so that the sediment drift module can track the drift trajectory of the suspended sediment mass center more precisely.

[0121] S140, inputting the sediment characteristic information at the current moment into the target sediment drift prediction model to obtain sediment drift prediction information.

[0122] The prediction process is as follows: Constructing the initial conditions specifically involves assigning a value to the initial position C0 of the suspended sediment using the latest sediment characteristic information obtained dynamically, converting the longitude and latitude coordinates of the suspended sediment range into a format recognizable by the model, such as a polygon vertex list or grid data, and using this as the initial calculation condition for the tidal module, sediment module, and wave module. Using the real-time calibrated wind factor λ and current factors β1 and β2 to calculate V L =β0+β1V c+β2V w Assign parameters and use them as initial conditions for the calculation of the tidal module, sediment module and wave module. Obtain wind field data in the current and predicted time period, including wind speed, wind direction, etc. Obtain ocean current data in the current and predicted time period, including flow velocity, flow direction, etc. Obtain tide data in the current and predicted time period, and consider the potential impact of tides on the diffusion of suspended sediment as the initial conditions driven by the model. According to the dynamically acquired longitude and latitude information of suspended sediment in coastal estuaries, determine the longitude and latitude of the location and range of sediment particles, and use this as the initial condition of the sediment turbulence diffusion module. Set the simulation time range and step size to observe the drift and diffusion process of sediment over time.

[0123] In one example, the time corresponding to the latest sediment image is used as the starting point of dynamic prediction, and the suspended sediment drift is calculated and predicted for 8 hours based on the latest initial conditions that have been constructed. If there is a new sediment image sent back by the small synthetic aperture radar within the predicted 8 hours, the newly sent sediment image is used to recalibrate the parameters and construct the initial conditions, and the sediment drift is calculated and predicted for 8 hours based on the reconstructed initial conditions.

[0124] The embodiments of the present invention can significantly improve the accuracy and reliability of predictions by optimizing data processing algorithms and building more accurate drift models. Advanced machine learning algorithms and ocean dynamics principles are used to deeply mine and analyze airborne radar data, providing a basis for more accurately predicting the drift trajectory and distribution range of coastal suspended sediment. When building the drift model, the influence of various marine environmental factors on the drift of suspended sediment is fully considered. Through the fusion and comprehensive analysis of multi-source data, the comprehensiveness and accuracy of the prediction are improved.

[0125] In some embodiments, see Figure 5 , Figure 5 A schematic diagram of the structure of a coastal estuary suspended sediment drift prediction device provided in an embodiment of the present application; The present application embodiment provides a coastal estuary suspended sediment drift prediction device 500, including: a first acquisition module 510, a feature extraction module 520, a training module 530 and a second acquisition module 540, wherein:

[0126] The first acquisition module 510 is configured to acquire sediment images of the target monitoring area using a synthetic aperture radar mounted on an unmanned aerial vehicle;

[0127] The feature extraction module 520 is configured to pre-process the sediment image and extract sediment feature information from the pre-processed sediment image; the sediment feature information includes sediment distribution information, sediment concentration, sediment particle size, sediment drift speed, longitude and latitude of the sediment diffusion range, sediment shape and sediment surface area;

[0128] The training module 530 is configured to train a preset sediment drift prediction model based on sediment characteristic information within a preset period, determine wind factors and current factors that affect sediment drift, and obtain a target sediment drift prediction model;

[0129] The second acquisition module 540 is configured to input the sediment characteristic information at the current moment into the target sediment drift prediction model to obtain sediment drift prediction information.

[0130] In some embodiments, the feature extraction module 520 is specifically configured to perform filtering, denoising and geometric correction on the sediment image.

[0131] In some embodiments, the feature extraction module 520 is specifically configured as follows:

[0132] Threshold segmentation, edge detection and region growing techniques are used to remove the background image from the sediment image to obtain the sediment area image.

[0133] The influencing features in the sediment area image are enhanced through histogram equalization and contrast adjustment algorithms;

[0134] A neural network classifier is used to classify and identify the sediment area images after feature enhancement processing to extract sediment feature information.

[0135] In some embodiments, the preset sediment drift prediction model includes: a tidal module, a sediment module and a wave module; the tidal module is configured to simulate the water movement in the target monitoring area, including the water flow changes caused by tides and river runoff, the sediment module is configured to simulate the convection and turbulent diffusion of the water flow, and the wave module is configured to simulate the action of ocean waves.

[0136] In some embodiments, the tidal flow module includes a momentum equation, a continuity equation, a temperature equation, a salinity equation, and a density equation, expressed as:

[0137]

[0138] Where x, y, z are the coordinates of the east, north and vertical directions in the Cartesian coordinate system; u, v, w are the flow velocities in the three directions; T is the temperature; S is the salinity; ρ is the density; P a is the sea surface air pressure; P h is the hydrostatic pressure; q is the non-hydrostatic pressure; f is the Coriolis parameter; g is the gravitational acceleration; K m is the vertical eddy viscosity coefficient; K h is the vertical vortex thermal diffusion coefficient; F u , F v , F T , F S are the horizontal and vertical momentum, heat and salinity diffusion terms, respectively;

[0139] The calculation equation corresponding to the sediment module is as follows:

[0140]

[0141] Where i represents the component of sediment; C i , w i Respectively represent the concentration and sedimentation velocity of the i-th component sediment; A h is the horizontal eddy viscosity coefficient; K h is the vertical eddy viscosity coefficient;

[0142]

[0143] In the formula, Q i is the flushing flux, P b is the bottom porosity, F bi is the bottom friction of sediment component i, τ b is the bottom scour stress, τ ci is the critical scour stress at the bottom of component i, D i is the sedimentation flux, E i is the erosion flux;

[0144] The wave module uses the wave action density spectrum to describe ocean waves. The governing equation of the wave action density spectrum is expressed as:

[0145]

[0146] Where N is the wave action density spectrum; t is time; σ is the relative frequency; θ is the wave direction; C σ and C θ are the propagation speeds of waves in the spectral space (σ, θ); C g is the group velocity; V is the ambient water velocity; is the horizontal divergence operator in geographic space; the first term on the left side of the equation represents the local variation of wave action density in time, the second term represents the propagation of wave action density in space, and the third and fourth terms represent the frequency shift and refraction caused by changes in water depth and flow velocity, respectively. tot is the source-sink term, and its formula is as follows:

[0147] S tot =S in +S nl3 +S nl4 +S ds，w +S ds，b +S ds,br ;

[0148] In the formula, S in is the growth function of wind-generated waves, i.e., wind energy input; S nl / 3is the nonlinear three-wave interaction; S nl / 4 is the nonlinear four-wave interaction; S ds,w is the wave energy dissipation caused by the breaking of the white hat; S ds,b is the wave energy dissipation caused by bottom friction; S ds,br It is the dissipation of wave energy caused by wave breaking due to changes in water depth.

[0149] In some embodiments, the preset sediment drift prediction model includes a prediction module; the prediction module is represented as:

[0150]

[0151] V L =β0+β1V c +β2V w

[0152] β2 = λT;

[0153] Where C0 is the initial position of the sediment center, Δt is the sediment drift time, C is the final position of the sediment center, V L V is the drift velocity of suspended sediment in coastal estuaries; c is the sea surface velocity, V w is the wind speed at 10 meters above the sea surface, β1, β2, β3 are current factors, λ is the wind factor, and T is the conversion matrix that introduces the drift deflection angle.

[0154] The device for predicting the drift of suspended sediment in coastal estuaries provided in the embodiment of the present application can implement each process in the embodiment corresponding to the above-mentioned method for predicting the drift of suspended sediment in coastal estuaries. To avoid repetition, they will not be described again here.

[0155] It should be noted that the coastal estuary suspended sediment drift prediction device provided in the embodiment of the present application and the coastal estuary suspended sediment drift prediction method provided in the embodiment of the present application are based on the same application concept. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned coastal estuary suspended sediment drift prediction method, and the repeated parts will not be repeated.

[0156] In some embodiments, see Figure 6 , Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. An electronic device 600 provided in an embodiment of the present application includes a processor 610 and a memory 620; the memory 620 stores a computer program, wherein the computer program implements the above-mentioned coastal estuary suspended sediment drift prediction method when executed by the processor.

[0157] Specifically, the processor 610 may include, for example, a general-purpose microprocessor, an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 610 may also include an onboard memory for cache purposes. The processor 610 may be a single processing unit or multiple processing units for executing different actions of the method flow according to the embodiment of the present application.

[0158] The memory 620 may be any medium capable of containing, storing, conveying, propagating or transmitting instructions. For example, the memory 620 may include, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared or semiconductor system, device, component or propagation medium. Specific examples of the memory 620 include: a magnetic storage device, such as a magnetic tape or a hard disk (HDD); an optical storage device, such as a compact disk (CD-ROM); a random access memory (RAM) or flash memory; and / or a wired / wireless communication link.

[0159] The present application also provides a computer-readable medium on which a computer program is stored. When the program is executed by a processor, the above-mentioned method for predicting the drift of suspended sediment in coastal estuaries is implemented. The computer-readable medium may be included in the device / apparatus / system described in the above-mentioned embodiments; or it may exist independently without being assembled into the device / apparatus / system. The above-mentioned computer-readable medium carries one or more programs. When the above-mentioned one or more programs are executed, the method according to the embodiment of the present application is implemented.

[0160] According to an embodiment of the present application, a computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium or any combination of the two. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, wherein a computer-readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wired, optical cable, radio frequency signal, etc., or any suitable combination of the above.

[0161] Those skilled in the art will appreciate that the features described in the various embodiments and / or claims of the present application may be combined and / or combined in a variety of ways, even if such combinations or combinations are not explicitly described in the present application. In particular, without departing from the spirit and teachings of the present application, the features described in the various embodiments and / or claims of the present application may be combined and / or combined in a variety of ways. All of these combinations and / or combinations fall within the scope of the present application. Therefore, the scope of the present application should not be limited to the above-described embodiments, but should be determined not only by the appended claims, but also by the equivalents of the appended claims.

Claims

1. A method for predicting the drift of suspended sediment in coastal estuaries, characterized in that: include: Use UAVs equipped with synthetic aperture radar to obtain sediment images of the target monitoring area; Preprocessing the sediment image, and extracting sediment characteristic information from the preprocessed sediment image; the sediment characteristic information includes sediment distribution information, sediment concentration, sediment particle size, sediment drift speed, longitude and latitude of sediment diffusion range, sediment shape and sediment surface area; Based on the sediment characteristic information within a preset period, a preset sediment drift prediction model is trained to determine the wind factor and the current factor that affect the sediment drift, and obtain a target sediment drift prediction model; The sediment characteristic information at the current moment is input into the target sediment drift prediction model to obtain sediment drift prediction information.

2. The method for predicting suspended sediment drift in coastal estuaries according to claim 1, characterized in that: The preprocessing of the sediment image includes: The sediment image is subjected to filtering, denoising and geometric correction processing.

3. The method for predicting suspended sediment drift in coastal estuaries according to claim 1, characterized in that: The step of extracting sediment feature information from the pre-processed sediment image includes: Using threshold segmentation, edge detection and region growing techniques, the background image is removed from the sediment image to obtain a sediment region image; The influencing features in the sediment area image are enhanced by using a histogram equalization and contrast adjustment algorithm; A neural network classifier is used to classify and identify the sediment area image after feature enhancement processing to extract sediment feature information.

4. The method for predicting suspended sediment drift in coastal estuaries according to claim 1, characterized in that: The preset sediment drift prediction model includes: a tidal module, a sediment module and a wave module; the tidal module is configured to simulate the water flow movement in the target monitoring area, including the water flow changes caused by tides and river runoff, the sediment module is configured to simulate the convection and turbulent diffusion of the water flow, and the wave module is configured to simulate the action of ocean waves.

5. The method for predicting the drift of suspended sediment in coastal estuaries according to claim 4, characterized in that: The tidal current module includes momentum equation, continuity equation, temperature equation, salinity equation and density equation, which can be expressed as: ρ = ρ(T, S, p); Where x, y, z are the coordinates of the east, north and vertical directions in the Cartesian coordinate system; u, v, w are the flow velocities in three directions; T is the temperature; S is the salinity; ρ is the density; P a is the sea surface air pressure; P h is the hydrostatic pressure; q is the non-hydrostatic pressure; f is the Coriolis parameter; g is the gravitational acceleration; K m is the vertical eddy viscosity coefficient; K h is the vertical vortex thermal diffusion coefficient; F u , F v , F T , F S are the horizontal and vertical momentum, heat and salinity diffusion terms, respectively; The calculation equation corresponding to the sediment module is as follows: Where i represents the component of sediment; C i , w i Respectively represent the concentration and sedimentation velocity of the i-th component sediment; A h is the horizontal eddy viscosity coefficient; K h is the vertical eddy viscosity coefficient; In the formula, Q i is the flushing flux, P b is the bottom porosity, F bi is the bottom friction of sediment component i, τ b is the bottom scour stress, τ ci is the critical scour stress at the bottom of component i, D i is the sedimentation flux, E i is the erosion flux; The wave module uses the wave action density spectrum to describe the ocean waves. The control equation of the wave action density spectrum is expressed as: Where N is the wave action density spectrum; t is time; σ is the relative frequency; θ is the wave direction; C σ and C θ are the propagation speeds of waves in the spectral space (σ, θ); C g is the group velocity; V is the ambient water velocity; is the horizontal divergence operator in geographic space; The first term on the left side of the equation represents the local change of wave action density in time, the second term represents the propagation of wave action density in space, and the third and fourth terms represent the frequency shift and refraction caused by changes in water depth and flow velocity, respectively. tot is the source-sink term, and its formula is as follows: S tot =S in +S nl3 +S nl4 +S ds，w +S ds，b +S ds，br ; In the formula, Si n is the growth function of wind-generated waves, i.e., wind energy input; Sn l / 3 is the nonlinear three-wave interaction; Sn l / 4 is the nonlinear four-wave interaction; Sd s ,w is the wave energy dissipation caused by the white hat breaking; Sd s , b is the wave energy dissipation caused by bottom friction; Sd s , br It is the dissipation of wave energy caused by wave breaking due to changes in water depth.

6. The method for predicting the drift of suspended sediment in coastal estuaries according to claim 1, characterized in that: The preset sediment drift prediction model includes a prediction module; the prediction module is represented as: V L =β0+β1V c +β2V w β2 = λT; Where C0 is the initial position of the sediment center, Δt is the sediment drift time, C is the final position of the sediment center, V L V is the drift velocity of suspended sediment in coastal estuaries; c is the sea surface velocity, V w is the wind speed at 10 meters above the sea surface, β1, β2, β3 are current factors, λ is the wind factor, and T is the conversion matrix that introduces the drift deflection angle.

7. A device for predicting the drift of suspended sediment in coastal estuaries, characterized in that: include: A first acquisition module, a feature extraction module, a training module, and a second acquisition module, wherein: The first acquisition module is configured to acquire sediment images of the target monitoring area using a synthetic aperture radar mounted on an unmanned aerial vehicle; The feature extraction module is configured to preprocess the sediment image and extract sediment feature information from the preprocessed sediment image; the sediment feature information includes sediment distribution information, sediment concentration, sediment particle size, sediment drift speed, longitude and latitude of sediment diffusion range, sediment shape and sediment surface area; The training module is configured to train a preset sediment drift prediction model based on the sediment characteristic information within a preset period, determine the wind factor and the current factor that affect the sediment drift, and obtain a target sediment drift prediction model; The second acquisition module is configured to input the sediment characteristic information at a current moment into the target sediment drift prediction model to obtain sediment drift prediction information.

8. An electronic device comprising a processor and a memory; the memory stores a computer program, wherein: When the computer program is executed by the processor, the computer program implements the method for predicting the drift of suspended sediment in coastal estuaries according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, wherein when the program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images