Ship-borne radar sea wave perception domain extension method and system based on deep learning

Abstract

The invention belongs to the technical field of marine environment expansion, and discloses a ship-borne radar sea wave perception domain expansion method and system based on deep learning, and the method comprises the steps: building a sea wave perception domain step-by-step expansion model, and achieving the spatial feature modeling and limited range expansion of local observation data through a sea wave perception domain expansion sub-module, sea wave field information in a larger range is deduced step by step in a step-by-step expansion and residual connection mode; meanwhile, a sea wave continuity constraint mechanism is introduced, continuity constraint is applied to the junction of the expanded sea area and the original observation sea area, and it is ensured that the whole wave field keeps high precision in a large range and meanwhile has space continuity. Through the ship-borne radar sea wave sensing domain extension method based on deep learning, the real-time sensing ability of the ship-borne radar to a large-range sea area in a local environment can be remarkably improved, independent operation of a ship in a marine environment is supported, and then the ship navigation safety and the real-time sea wave sensing ability of ocean engineering application are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of marine environment expansion, and particularly relates to a shipborne radar sea wave perception domain expansion method and system based on deep learning. BACKGROUND

[0002] With the continuous development of global maritime transportation, fishery production and marine engineering activities, real-time and accurate acquisition of wave field information in a large range of sea area has become an important basis for ensuring the safety of ship navigation, developing route planning and carrying out marine operations. Wave is a key environmental factor affecting ship motion and structural safety, and its spatial distribution characteristics are directly related to navigation safety, operation window selection and disaster warning. Therefore, being able to timely grasp the wave field in a large range is not only helpful to improve the safety and efficiency of autonomous ship navigation, but also has important significance for marine engineering and disaster prevention and mitigation.

[0003] At present, shipborne radar as a common observation means can detect the wave field within a certain range near the ship, and extract key parameters such as effective wave height, spectral peak period, and main wave direction. Such data are very useful for local situation awareness of the ship, but its coverage is limited, and it can only reflect the local wave conditions within tens of kilometers near the ship. Therefore, how to calculate and obtain the wave field information in a larger range on the basis of limited local observation data has become a key technical problem to be solved for the safety of ship navigation and marine engineering application.

[0004] The existing methods for expanding the range of sea wave perception mainly include the following categories: (1) Numerical model calculation and data assimilation: This kind of method uses large-scale sea wave numerical models such as WAVEWATCH III and SWAN, and combines with buoy, satellite or ship observation data for assimilation correction to obtain large-scale wave field information.

[0005] (2) Sea area interpolation based on measured data: This kind of method extends the local wave information measured by shipborne radar to the adjacent area through spatial interpolation or statistical model, so as to obtain the wave estimation in a certain range.

[0006] Through the above analysis, the problems and defects of the existing technology are: (1) Lack of real-time performance: The existing large-scale wave field acquisition mainly relies on numerical model and data assimilation method, which has large calculation amount and complex process, and often needs several hours or even longer calculation time, which is difficult to meet the demand of real-time wave field information in the process of ship navigation.

[0007] (2) Lack of local independence: The results of numerical model usually depend on the generation of shore-based supercomputing platform, and the ship needs to obtain the data through satellite communication or other remote methods, which has the problems of data transmission dependence and delay, and it is difficult to realize the local autonomous operation of single ship in the marine environment.

[0008] (3) Limited calculation accuracy: some methods extend local observation data to adjacent areas through interpolation or statistical models, but the results are highly dependent on the number and distribution of observation points, making it difficult to accurately reflect the wave propagation and nonlinear evolution characteristics, and the prediction accuracy is limited. SUMMARY

[0009] To overcome the problems in the related art, the embodiment of the present application provides a shipborne radar sea wave perception domain expansion method and system based on deep learning, and the technical solution is as follows: The present application is implemented as follows: the shipborne radar sea wave perception domain expansion method based on deep learning comprises the following steps: S1, model data set construction: obtaining wave field data observed by a shipborne radar, and wave field analysis data covering and exceeding the observation range of the shipborne radar, generating a data set matching the shipborne radar observed wave field and the target expanded wave field through space-time matching and sea area range interception operation; S2, sea wave perception domain step-by-step expansion model construction and training: constructing a sea wave perception domain step-by-step expansion model, expanding the local wave field in a limited range through a sea wave perception domain expansion submodule, inputting a visual converter network for spatial feature extraction and modeling; expanding the local wave field step by step through the sea wave perception domain step-by-step expansion module, and providing residual connection between the submodules; constructing a sea wave continuity constraint loss function, which comprehensively considers the overall accuracy of the expansion result and the physical continuity at the junction of the expansion area and the original observation area; training the model using the obtained model data set; S3, wave expansion and verification: using the trained sea wave perception domain step-by-step expansion model to expand the local wave field observed by the shipborne radar, and verifying the reliability of the expansion result using the accuracy index.

[0010] In step S1, the model data set construction comprises: Obtaining shipborne radar measured wave field data and regional significant wave height analysis data, matching the wave height analysis data corresponding to the time of the shipborne radar measured wave field data; intercepting the sea area corresponding to the wave height analysis data, the sea area center point is consistent with the shipborne radar measured sea area, and the model data set construction is completed; Wherein, the wave field data is based on X-band navigation radar data inversion to obtain latitude and longitude gridding regional significant wave height data; the wave height analysis data adopts 1 / 12° resolution significant wave height analysis data of the Copernicus Marine Environment Monitoring Service (CMEMS), and the space-time matching is to perform space-time interpolation on the significant wave height analysis data to form analysis data with uniform spatial resolution covering and exceeding the radar observation range, and at the same time, the time stamp of the radar measured wave height data is matched.

[0011] Further, the length and width of the shipborne radar measured wave field data are The length and width of the wave height analysis data at the corresponding time are Through the construction of the model data set, the length of the shipborne radar sea wave perception domain expansion target is times the original, and the area expansion target is times the original.

[0012] In step S2, the construction of the sea wave perception domain expansion submodule includes: For the first layer sea wave perception domain expansion submodule in the step-by-step expansion model, the boundary expansion function is used to supplement the outer edge of the local sea wave field data observed by the shipborne radar, so as to generate an expanded area in width outside the original observation range, and the sea area grid length is expanded from to The expanded wave field data is input into the visual converter network for spatial feature coding and modeling. In the visual converter network, the ViT of each sea wave perception domain expansion submodule uses 4 layers of encoder, the number of attention heads is 4, and the hidden layer latitude is 256; so as to obtain the wave field feature representation containing local observation and extension area at the same time, providing input basis for subsequent step-by-step expansion; the expression is: ; In the formula, is the feature vector output by the sea wave perception domain expansion submodule, is the visual converter network, is the grid point row and column position index, is the length and width grid number of the shipborne radar measured wave field data, is the data of the grid point in the shipborne radar observation wave height input, is the interpolation weight of the boundary expansion function of the first layer sea wave perception domain expansion submodule, is the visual converter network parameter of the first layer sea wave perception domain expansion submodule.

[0013] Further, the sea wave perception domain step-by-step expansion module is composed of sea wave perception domain expansion submodules in series, which is used to realize the step-by-step gradual expansion of the wave field data from the initial range to the target range.

[0014] Further, the construction of the sea wave perception domain step-by-step expansion module includes: The input shipborne radar local observation wave field data is input into the first sea wave perception domain expansion submodule, the boundary expansion in a limited range is performed, and the spatial features are extracted, so as to obtain the wave field with a grid length of larger than the original range. The extended result is used as the input to the next wave sensing domain extension submodule, and the process is repeated iteratively. Next, the initial small-scale wave field is gradually pushed outward to a grid with a side length of... This module covers a wide range of sea areas; it also adds residual links between sub-modules, and the module ultimately outputs feature vectors. ; ; In the formula, For the first The feature vector output by the extended submodule of the layered wave sensing domain. For shipborne radar observation wave height input Grid point data, For the number of iterations, For the first In the feature vector output by the extended submodule of the wave sensing domain Grid point data, For the first Interpolation weights for the boundary extension function of the extended submodule of the layered wave sensing domain. For the first The feature vector output by the extended submodule of the layered wave sensing domain. To expand the number of submodules in the wave sensing domain, For the first Visual converter network parameters of the extended submodule of the layered wave perception domain. Input for shipborne radar observation wave height.

[0015] In step S2, the construction of the wave continuity constraint loss function includes: Error term of the step-by-step expansion module: Step-by-step expansion of the wave sensing domain module After this, the output feature vector Corresponding actual observed wave field Error calculations are performed across the entire sea area to constrain the overall prediction results of the wave sensing domain progressively expanding modules; Continuity constraint error term: for the boundary between the extended region and the original observation region. The first derivative of the wave field output by the wave sensing domain progressively expanded model and the observed wave field at the boundary is calculated. With the second derivative ; Calculate the first derivative of the observed wave field at the inner boundary. With the second derivative and calculate and ,as well as and To minimize errors between them and ensure the continuity of the wave field at the boundary; The total loss function is formed by weighting and combining the step-by-step expansion module error and the continuity constraint error; ; In the formula, is the total loss, are weighting coefficients of the first derivative and the second derivative loss in the step-by-step expansion module error term and the continuity constraint error term of the sea wave perception domain, is the sample number, is the number of the i-th sample, is the number of the i-th sample, are model output data and real observed wave field data of the i-th sample, are the first derivative and the second derivative of the wave field output by the i-th sample model and the observed wave field at the junction, are the first derivative and the second derivative of the observed wave field of the i-th sample at the inner boundary. The training of the sea wave perception domain step-by-step expansion model in step S3 includes: The wave field data observed by the shipborne radar is input into the sea wave perception domain step-by-step expansion model for training, and the model is trained for 100 rounds, and 64 groups of data are input in each batch. Meanwhile, the model introduces a cosine annealing algorithm to dynamically schedule the learning rate, and the initial learning rate is

[0016] , and finally decays to ; ; ; In the formula, is the learning rate of the i-th step, are the minimum value and the maximum value of the learning rate, is the current iteration step number, is the total iteration step number.

[0017] In step S3, the wave expansion further includes wave expansion result accuracy verification; The root mean square error RMSE and the correlation coefficient are used as accuracy evaluation indexes to evaluate the sea wave perception domain expansion effect of the sea wave perception domain step-by-step expansion model.

[0018] Another purpose of the present application is to provide a shipborne radar sea wave perception domain expansion system based on deep learning, which is realized by the shipborne radar sea wave perception domain expansion method based on deep learning, and the system comprises: ​​​The model data set construction module is configured to respectively acquire wave field data observed by a shipborne radar and wave field analysis data covering and exceeding the observation range of the shipborne radar, and generate a data set matching the wave field observed by the shipborne radar and the target extended wave field through space-time matching and sea area range interception operations; The sea wave perception domain step-by-step expansion model construction and training module is configured to construct a sea wave perception domain step-by-step expansion model, expand the local wave field in a limited range through a sea wave perception domain expansion submodule, and input the visual converter network for spatial feature extraction and modeling;The local wave field is expanded step by step through the sea wave perception domain step-by-step expansion module, and residual connections are provided between the submodules;A sea wave continuity constraint loss function is constructed, which comprehensively considers the overall accuracy of the expansion result and the physical continuity at the junction of the expansion area and the original observation area, and the model is trained using the constructed data set; The wave expansion and verification module is configured to expand the local wave field observed by the shipborne radar using the trained sea wave perception domain step-by-step expansion model, and verify the reliability of the expansion result using accuracy indicators; Any one or more of the model data set construction module, the sea wave perception domain step-by-step expansion model construction and training module, and the wave expansion and verification module are implemented by a processor executing computer program instructions stored in a memory.

[0019] Another object of the present application is to provide a shipborne radar sea wave perception domain expansion system based on deep learning, which comprises: The model data set construction module is configured to respectively acquire wave field data observed by a shipborne radar and wave field analysis data covering and exceeding the observation range of the shipborne radar, and generate a data set matching the wave field observed by the shipborne radar and the target extended wave field through space-time matching and sea area range interception operations; The sea wave perception domain step-by-step expansion model construction and training module is configured to construct a sea wave perception domain step-by-step expansion model, expand the local wave field in a limited range through a sea wave perception domain expansion submodule, and input the visual converter network for spatial feature extraction and modeling;The local wave field is expanded step by step through the sea wave perception domain step-by-step expansion module, and residual connections are provided between the submodules;A sea wave continuity constraint loss function is constructed, which comprehensively considers the overall accuracy of the expansion result and the physical continuity at the junction of the expansion area and the original observation area, and the model is trained using the constructed data set; The wave expansion and verification module is configured to expand the local wave field observed by the shipborne radar using the trained sea wave perception domain step-by-step expansion model, and verify the reliability of the expansion result using accuracy indicators; Any one or more of the model data set construction module, the sea wave perception domain step-by-step expansion model construction and training module, and the wave expansion and verification module are implemented by a processor executing computer program instructions stored in a memory.

[0020] In combination with all the above technical solutions, the present application has the following beneficial effects: Firstly, the present application aims to overcome the problems of limited sea wave perception range of existing shipborne radars, insufficient real-time performance of numerical models, and limited accuracy of interpolation methods, and proposes a shipborne radar sea wave perception domain expansion method based on deep learning. This method constructs a sea wave perception domain step-by-step expansion model, uses a sea wave perception domain expansion submodule to model and expand the spatial features of local observation data, and through step-by-step expansion operations, avoids the model from directly learning the complex mapping from local to global in a single calculation, reduces the difficulty of model training, and realizes the gradual high-precision deduction of larger range sea wave field information. At the same time, the present application introduces a sea wave continuity constraint mechanism, introduces multi-order differential continuity constraints at the boundary between the expanded sea area and the original observation sea area, the first-order derivative ensures the smooth connection of the wave surface slope at the boundary, and the second-order derivative ensures the natural extension of the wave surface curvature, thereby avoiding sudden changes or polyline patterns, and improving the physical credibility of the expansion results. Finally, through the weighting of the overall error term and the continuity error term, the balance between overall accuracy and physical continuity is achieved, and the expansion accuracy and the physical reasonableness of the expanded wave field are improved. Through the method of the present application, the real-time perception ability of shipborne radars in the local environment for large-scale sea areas can be significantly improved, supporting the independent operation of ships in the marine environment, thereby improving the safety of navigation and the efficiency of marine engineering applications.

[0021] Secondly, the present application improves the sea wave field perception range and accuracy of shipborne radars. Through the sea wave perception domain step-by-step expansion module, the local wave field data observed by the shipborne radar is gradually expanded to a larger range of sea areas, and the continuity of the wave field at the boundary is ensured by combining the sea wave continuity constraint loss function, so that the expanded wave field features cover a large range of sea areas and maintain high-precision spatial features, significantly improving the real-time sea wave perception ability of ship navigation safety and marine engineering applications.

[0022] Thirdly, the present application supports the localized and rapid operation of shipborne radars. Unlike traditional methods that rely on large-scale numerical models and remote data assimilation, the present application performs wave field extrapolation calculations based on deep learning models, which can quickly complete the wave field feature deduction of large-scale sea areas in the local environment of the ship, without relying on shore-based supercomputing platforms or satellite communication, realizing independent real-time sea wave perception and application of ships.

[0023] Fourth, the existing method mainly relies on large-scale numerical model and data assimilation or interpolation-based extension means, which is difficult to realize real-time, independent large-scale sea wave field perception in the local environment of the ship. The application first realizes the local rapid high-precision extension of shipborne radar local observation data to large-scale sea wave field, breaking through the limitations of traditional numerical model and interpolation method in real-time and accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0024] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and serve to explain the principles of the present disclosure, together with the description; Figure 1 is a flow chart of a shipborne radar sea wave perception domain extension method based on deep learning provided by an embodiment of the application; Figure 2 is a hierarchical expansion model structure diagram of sea wave perception domain provided by an embodiment of the application; Figure 3 is a comparison diagram of the expansion effects of different models provided by an embodiment of the application; Figure 4 is a sea area wave height expansion RMSE error diagram of a comparison model one provided by an embodiment of the application; Figure 5 is a sea area wave height expansion RMSE error diagram of a comparison model two provided by an embodiment of the application; Figure 6 is a sea area wave height expansion RMSE error diagram of a comparison model three provided by an embodiment of the application; Figure 7 is a sea area wave height expansion RMSE error diagram of a model provided by an embodiment of the application. DETAILED DESCRIPTION

[0025] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application will be described in detail below. In the following description, a lot of specific details are set forth in order to facilitate a full understanding of the present application. However, the present application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without departing from the spirit of the present application, so the present application is not limited to the specific implementation disclosed below.

[0026] The application proposes a shipborne radar sea wave perception domain expansion method based on deep learning, which models and expands the local wave observation through a visual converter network, realizes high-precision deduction from local to large range by combining a step-by-step expansion module and a residual connection, and introduces multi-order differential continuity constraints to ensure the spatial continuity and physical rationality of the expansion boundary, thereby realizing local, rapid, high-precision and continuous expansion of shipborne radar observation data to a large-scale sea wave field. On the basis of the traditional error constraint, the application proposes a sea wave continuity constraint loss to solve the discontinuity problem at the junction of the expansion area and the observation area. The first-order derivative ensures the smooth connection of the wave surface slope at the boundary, and the second-order derivative ensures the natural extension of the wave surface curvature, thereby avoiding sudden changes or polyline patterns and improving the physical credibility of the expansion result.

[0027] The innovation points of the application are embodied in the following aspects: (1) Sea wave perception domain step-by-step expansion module. The application proposes a multi-stage expansion strategy to expand the local wave data observed by the shipborne radar to a larger sea area through multi-stage extrapolation, reduces the difficulty of learning the global mapping at one time, and realizes high-precision step-by-step extrapolation.

[0028] (2) Sea wave continuity constraint mechanism. Multi-order differential continuity constraints are introduced at the junction of the expanded sea area and the original observation sea area, and error calculation of continuity is added to the loss function to ensure the continuity of the junction and the internal sea area and solve the spatial fault problem that may occur during the expansion process.

[0029] (3) Rapid expansion of shipborne data based on a deep learning model. The application uses a boundary expansion function and a visual converter network to perform spatial expansion and spatial feature coding and modeling on the shipborne radar measured wave field data, realizes efficient expansion of local observation information to a large-scale sea area, and significantly improves the wave field deduction speed based on the deep learning method while ensuring the prediction accuracy compared with the traditional numerical mode method.

[0030] Embodiment 1, as shown in Figure 1 The shipborne radar sea wave perception domain expansion method based on deep learning provided by the application embodiment includes the following steps: S1, model data set construction: obtain the wave field data observed by the shipborne radar, and the wave field analysis data covering and exceeding the observation range of the shipborne radar, generate a data set matched with the target expanded wave field through time and space matching and sea area range interception operation; Obtain the shipborne radar measured wave field data and the regional significant wave height analysis data, match the wave height analysis data at the corresponding time of the shipborne radar measured wave field data; intercept the sea area of the corresponding time wave height analysis data, the sea area center point is consistent with the shipborne radar measured sea area, and the model data set construction is completed; The wave field data is based on X-band navigation radar data inversion to obtain the latitude and longitude grid region effective wave height data; the wave height analysis data adopts the 1 / 12° resolution effective wave height analysis data of the Copernicus Marine Environment Monitoring Service (CMEMS), and the space-time matching is to perform space-time interpolation on the effective wave height analysis data to form analysis data with uniform spatial resolution covering and exceeding the radar observation range, and to make the time stamp of the radar measured wave height data correspond. The space-time matching and sea area range interception operation is to align the numerical simulation wave height data with the sea area center point of the radar measured wave height data, and set the sea area size with a side length of times the radar measured wave height data as the interception range.

[0031] The shipborne radar measured wave field data and the regional effective wave height analysis data are obtained, the wave height analysis data corresponding to the time of the shipborne radar measured wave field data is matched, the length and width of the shipborne radar measured wave field data are , and the length and width of the wave height analysis data corresponding to the time are . Through the construction of the model data set, the shipborne radar sea wave perception domain edge length expansion target is times the original, and the area expansion target is times the original.

[0032] S2, construction and training of the sea wave perception domain step-by-step expansion model: the sea wave perception domain step-by-step expansion model is constructed, the local wave field is expanded in a limited range by the sea wave perception domain expansion submodule, and the spatial feature extraction and modeling are performed by inputting the visual converter network; the local wave field is expanded step by step by the sea wave perception domain step-by-step expansion module, and residual connections are provided between the submodules; a sea wave continuity constraint loss function is constructed, which comprehensively considers the overall accuracy of the expansion result and the physical continuity at the junction of the expansion area and the original observation area; the model is trained using the obtained model data set; As shown in Figure 2 , using the constructed data set, first, the local wave field is expanded in a limited range by the sea wave perception domain expansion submodule, and the spatial feature encoding and modeling are performed by inputting the visual converter network, to obtain the expanded sea wave field feature vector; second, the local wave field is expanded step by step to a larger range of sea area by the step-by-step expansion module, and residual connections are added between the submodules; finally, the sea wave continuity constraint loss function is constructed to constrain the performance of the model in overall accuracy and continuity at the junction of the sea area expansion, to form a high-precision large-range sea wave field feature. Specifically, it includes: (1) constructing a sea wave perception domain expansion submodule; taking the first layer sea wave perception domain expansion submodule in the model as an example, first, the local sea wave field data observed by the shipborne radar is supplemented with a boundary expansion function to generate an expansion area with a width outside the original observation range, and the sea area grid side length is extended to ; then, the extended wave field data is input into a vision transformer network (Vision Transformer) for spatial feature encoding and modeling, so as to obtain a wave field feature representation containing local observation and extended area, thereby providing an input basis for subsequent step-by-step expansion.

[0033] ; In the formula, is a feature vector output by the sea wave perception domain expansion submodule, is a vision transformer network, is a grid point row-column position index, is the number of long and wide grids of the ship-borne radar measured wave field data, is the data of the grid point in the ship-borne radar observed wave height input, is an interpolation weight of the boundary expansion function of the first layer sea wave perception domain expansion submodule, is a vision transformer network parameter of the first layer sea wave perception domain expansion submodule.

[0034] (2) Construct a sea wave perception domain step-by-step expansion module; the sea wave perception domain step-by-step expansion module is composed of sea wave perception domain expansion submodules in series. First, the input ship-borne radar local observation wave field data is input into the first sea wave perception domain expansion module for boundary expansion in a limited range and extraction of spatial features, so as to obtain a wave field with a grid side length of larger than the original range; then, the expansion result is taken as the input of the next sea wave perception domain expansion module, and the iteration is performed step by step times, so as to avoid that the model directly learns a complex mapping from local to global in a single calculation, reduce the training difficulty, and gradually extrapolate the initial small-range wave field to a larger range of sea area with a grid side length of . At the same time, the module adds a residual link between the submodules to ensure stable convergence of the model. The module finally outputs a feature vector .

[0035] ; In the formula, is a feature vector output by the layer sea wave perception domain expansion submodule, is the data of the grid point in the ship-borne radar observed wave height input, is the number of iterations, is a feature vector output by the layer sea wave perception domain expansion submodule, is the data of the grid point in the ship-borne radar observed wave height input, ​data of the grid points, the first layer sea wave perception domain expansion sub-module, the first layer sea wave perception domain expansion sub-module output feature vector, the number of sea wave perception domain expansion sub-modules, the first layer sea wave perception domain expansion sub-module visual converter network parameter, the ship-borne radar observed wave height input.

[0036] (3) Construct a sea wave continuity constraint loss function; In the embodiment, the calculation of the sea wave continuity constraint loss function includes two parts: Step-by-step expansion module error term: for the sea wave perception domain step-by-step expansion module After output, the feature vector is compared with the corresponding real observed wave field Error calculation is performed in the entire sea area range to constrain the overall prediction result of the expansion module.

[0037] Continuity constraint error term: for the expansion area and the original observation area , the first-order derivative and the second-order derivative of the sea wave perception domain step-by-step expansion model output wave field and the observed wave field at the junction are calculated , the first-order derivative and the second-order derivative of the observed wave field at the inner boundary are calculated , and the errors between and are calculated to ensure the continuity of the wave field at the junction; The step-by-step expansion module error and the continuity constraint error are weighted and combined to form the total loss function; ; In the formula, is the total loss, are the weighting coefficients of the first-order derivative and the second-order derivative loss in the step-by-step expansion module error term and the continuity constraint error term, is the sample number, is the number of the first sample, are the model output data and the real observed wave field data of the first sample, respectively, are the first-order derivative and the second-order derivative of the wave field and the observed wave field at the junction of the model output of the first sample, respectively, The first The first and second derivatives of the observed wave field at the inner boundary of each sample.

[0038] S3, conduct accuracy verification of wave extension results.

[0039] After model training is complete, the local wave field observed by shipborne radar is input into the model to expand it, and a cosine annealing algorithm is introduced to dynamically adjust the learning rate to improve training performance; the root mean square error (RMSE) and correlation coefficient are used to measure the learning rate. Indicators such as [list of indicators] are used to compare the extended results with actual observation data, verifying the accuracy and reliability of the model's wave field extension over a large area of ​​sea. Specifically, these include: (1) Training the wave sensing domain progressive expansion model; Wave field data observed by shipborne radar were input into the wave sensing domain progressive expansion model for training. The model was trained for a total of 100 rounds, with 64 sets of data input in each batch. At the same time, the model introduced a cosine annealing algorithm to dynamically schedule the learning rate. The initial learning rate was Eventually decayed to ; ; In the formula, For training the first The learning rate of the step. These are the minimum and maximum values ​​of the learning rate, respectively. This represents the current iteration step. This represents the total number of iterations.

[0040] (2) Verify the wave propagation effect of the model; This invention uses root mean square error (RMSE) and correlation coefficient ( Error evaluation indexes are used to calculate the error between the true value of the sea area and the value of the sea area expansion.

[0041] ; ; In the formula, For the number of test samples, and The first True and extended values ​​of wave height for each sample This represents the average of the true wave height values.

[0042] Example 2: The deep learning-based shipborne radar wave perception domain extension system provided in this embodiment of the invention includes: The model data set construction module is configured to respectively acquire wave field data observed by a ship-borne radar and wave field analysis data covering and exceeding the observation range of the ship-borne radar, and generate a data set matched with the target extended wave field through space-time matching and sea area range interception operation. The sea wave perception domain step-by-step expansion model construction and training module is configured to construct a sea wave perception domain step-by-step expansion model, expand the local wave field in a limited range through a sea wave perception domain expansion submodule, and input the visual converter network for spatial feature extraction and modeling; the local wave field is expanded step by step through the sea wave perception domain step-by-step expansion module, and residual connections are provided between the submodules; a sea wave continuity constraint loss function is constructed, which comprehensively considers the overall accuracy of the expansion result and the physical continuity at the junction between the expansion area and the original observation area, and the model is trained using the constructed data set. The wave expansion and verification module is configured to expand the local wave field observed by the ship-borne radar using the trained sea wave perception domain step-by-step expansion model, and verify the reliability of the expansion result using accuracy indicators. Any one or more of the model data set construction module, the sea wave perception domain step-by-step expansion model construction and training module, and the wave expansion and verification module are implemented by a processor executing computer program instructions stored in a memory.

[0043] To further prove the positive effects of the above embodiments, the present application based on the above technical solutions carries out the following experiments: the model of the present application is compared with a bilinear edge extrapolation interpolation model (comparison model one), the model of the present application without using a step-by-step expansion strategy (comparison model two), and the model of the present application without adding a continuity constraint error term (comparison model three), the sea wave field expansion effects of different models are compared, see Figure 3 .

[0044] The experimental results are shown in Table 1, and the model of the present application shows lower error and higher correlation in the comparison experiment. The step-by-step expansion strategy is added to the model of the present application compared with the comparison model two, which proves that the strategy effectively reduces the training difficulty and stability of the model, and realizes high-precision expansion of the wave field; the continuity constraint error term is added to the model of the present application compared with the comparison model three, which proves that the continuity constraint error term accurately represents the wave continuity feature of the model of the present application, and further improves the accuracy and reliability of the model in expanding the perceived sea area. The sea area wave height expansion RMSE error graph of different models is shown in Figures 4-7 .

[0045] Table 1 Comparison of reconstruction error of different models

[0046] The above merely describes the preferred specific embodiments of the present application, but the protection scope of the present application is not limited thereto, and any modification, equivalent replacement and improvement within the technical range disclosed by the present application and within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for extending the wave perception domain of shipborne radar based on deep learning, characterized in that, The method includes the following steps: S1, Model Dataset Construction: Obtain wave field data observed by shipborne radar, as well as wave field analysis data that covers and exceeds the observation range of shipborne radar. Through spatiotemporal matching and sea area range interception operations, generate a dataset that matches the shipborne radar observed wave field with the target extended wave field. S2, Construction and Training of a Stepwise Expanding Wave Perception Domain Model: A stepwise expanding wave perception domain model is constructed. A local wave field is expanded within a limited range using a wave perception domain expansion submodule, and the expanded field is input into a visual converter network for spatial feature extraction and modeling. The local wave field is gradually expanded in stages using the stepwise expanding wave perception domain module, with residual connections between each submodule. A wave continuity constraint loss function is constructed, which comprehensively considers the overall accuracy of the expansion results and the physical continuity at the boundary between the expanded region and the original observation region. The model is then trained using the obtained model dataset. S3, Wave Extension and Validation: The local wave field observed by shipborne radar is extended using the trained wave perception domain stepwise extension model, and the reliability of the extension results is verified using accuracy indicators.

2. The deep learning-based method for extending the wave perception domain of shipborne radar according to claim 1, characterized in that, In step S1, the model dataset construction includes: Acquire the measured wave field data and regional effective wave height analysis data of the shipborne radar, and match the wave height analysis data of the shipborne radar at the corresponding time. Extract the sea area of ​​the wave height analysis data at the corresponding time, and make the center point of the sea area consistent with the sea area measured by the shipborne radar to complete the construction of the model dataset. Among them, the wave field data is based on the inversion of X-band navigation radar data to obtain the significant wave height data of the latitude and longitude gridded area; the wave height analysis data adopts the 1 / 12° resolution significant wave height analysis data of Copernicus Marine Environment Observation Service (CMEMS); the spatiotemporal matching is to perform spatiotemporal interpolation on the significant wave height analysis data to form analysis data with a unified spatial resolution that covers and exceeds the radar observation range, while making it correspond to the timestamp of the radar measured wave height data. The spatiotemporal matching and sea area interception operations align the center point of the sea area between the numerically simulated wave height data and the radar-measured wave height data, and set the side length to the radar-measured wave height data. Set the intercept range to a size equal to a multiple of the sea area.

3. The deep learning-based method for extending the wave perception domain of shipborne radar according to claim 2, characterized in that, The length and width of the wave field data measured by shipborne radar are both The length and width of the wave height analysis data at the corresponding time are both By constructing a model dataset, the side length of the shipborne radar wave perception domain is expanded to the original value. The area expansion target is [number] times the original [area]. times.

4. The deep learning-based method for extending the wave perception domain of shipborne radar according to claim 1, characterized in that, In step S2, the construction of the wave sensing domain extension submodule includes: For the first-layer wave sensing domain expansion submodule in the hierarchical expansion model of the wave sensing domain, the local wave field data obtained from shipborne radar observations are processed using a boundary expansion function to supplement the outer edge, thereby generating an expanded region with a wider width outside the original observation range, and increasing the side length of the sea area grid from... Expand to The expanded wave field data is input into a visual converter network for spatial feature encoding and modeling. In this network, each wave perception domain extension submodule uses a 4-layer encoder (ViT), with 4 attention heads and 256 hidden layer dimensions. This yields a wave field feature representation that simultaneously includes local observations and extended regions, providing the input basis for subsequent progressive expansion. The expression is: ； In the formula, The feature vector output by the wave sensing domain extension submodule. For visual converter networks, For the row and column positions of the grid points, The number of grid squares in the length and width of the measured wave field data from shipborne radar. For shipborne radar observation wave height input Grid point data, The interpolation weights are used for the boundary expansion function of the first-layer wave sensing domain extension submodule. These are the network parameters for the visual converter of the first-layer wave perception domain extension submodule.

5. The deep learning-based method for extending the wave perception domain of shipborne radar according to claim 4, characterized in that, The wave sensing domain expansion module consists of... It consists of a series of wave sensing domain extension sub-modules, used to realize the gradual expansion of wave field data from the initial range to the target range.

6. The deep learning-based method for extending the wave perception domain of shipborne radar according to claim 5, characterized in that, The construction of the wave sensing domain progressively expanding module includes: The input wave field data from the shipborne radar's local observations is fed into the first wave sensing domain extension submodule. This submodule performs a limited-range boundary extension and extracts spatial features, resulting in a grid with a side length larger than the original range. wave field The extended result is used as the input to the next wave sensing domain extension submodule, and the process is repeated iteratively. Next, the initial small-scale wave field is gradually pushed outward to a grid with a side length of... This module covers a wide range of sea areas; it also adds residual links between sub-modules, and the module ultimately outputs feature vectors. ; ； In the formula, For the first The feature vector output by the extended submodule of the layered wave sensing domain. For shipborne radar observation wave height input Grid point data, For the number of iterations, For the first In the feature vector output by the extended submodule of the wave sensing domain Grid point data, For the first Interpolation weights for the boundary extension function of the extended submodule of the layered wave sensing domain. For the first The feature vector output by the extended submodule of the layered wave sensing domain. To expand the number of submodules in the wave sensing domain, For the first Visual converter network parameters of the extended submodule of the layered wave perception domain. Input for shipborne radar observation wave height.

7. The deep learning-based method for extending the wave perception domain of shipborne radar according to claim 1, characterized in that, In step S2, the construction of the wave continuity constraint loss function includes: Error term of the step-by-step expansion module: Step-by-step expansion of the wave sensing domain module After this, the output feature vector Corresponding actual observed wave field Error calculations are performed across the entire sea area to constrain the overall prediction results of the wave sensing domain progressively expanding modules; Continuity constraint error term: for the boundary between the extended region and the original observation region. The first derivative of the wave field output by the wave sensing domain progressively expanded model and the observed wave field at the boundary is calculated. With the second derivative ; Calculate the first derivative of the observed wave field at the inner boundary. With the second derivative and calculate and ,as well as and To minimize errors between them and ensure the continuity of the wave field at the boundary; The error of the progressively expanding module and the error of the continuity constraint are weighted and combined to form the total loss function; ； In the formula, For the total loss, These are all weighting coefficients for the first and second derivative losses in the error terms of the progressively expanding module of the wave sensing domain and the continuity constraint error terms. For the sample size, For the first Number of samples The first Model output data and actual observed wave field data for each sample. The first The first and second derivatives of the wave field output by the sample model and the observed wave field at the boundary. The first The first and second derivatives of the observed wave field at the inner boundary of each sample.

8. The deep learning-based method for extending the wave perception domain of shipborne radar according to claim 1, characterized in that, In step S3, the training of the wave perception domain progressively expanding model includes: Wave field data observed by shipborne radar is input into the wave sensing domain progressively expanding model for training. The model is trained for a total of 100 rounds, with 64 sets of data input in each batch. Simultaneously, a cosine annealing algorithm is introduced to dynamically schedule the learning rate, with an initial learning rate of... Eventually decayed to ; ； In the formula, For training the first The learning rate of the step. These are the minimum and maximum values ​​of the learning rate, respectively. This represents the current iteration step. This represents the total number of iterations.

9. The deep learning-based method for extending the wave perception domain of shipborne radar according to claim 1, characterized in that, In step S3, the wave expansion is followed by wave expansion result accuracy verification; Using root mean square error (RMSE) and correlation coefficient As an accuracy evaluation index, the wave perception domain expansion effect of the wave perception domain progressive expansion model is evaluated.

10. A deep learning-based shipborne radar wave perception domain extension system, characterized in that, The system is implemented using the deep learning-based shipborne radar wave perception domain extension method as described in any one of claims 1-9, and the system includes: The model dataset construction module is configured to acquire wave field data observed by shipborne radar and wave field analysis data that covers and exceeds the observation range of the shipborne radar, and generate a dataset that matches the shipborne radar observed wave field with the target extended wave field through spatiotemporal matching and sea area truncation operations. The wave perception domain progressive expansion model construction and training module is configured to construct a wave perception domain progressive expansion model. It expands the local wave field to a limited range through a wave perception domain expansion sub-module and inputs the data into a visual converter network for spatial feature extraction and modeling. The local wave field is progressively expanded in stages through the wave perception domain progressive expansion module, with residual connections between each sub-module. A wave continuity constraint loss function is constructed, which comprehensively considers the overall accuracy of the expansion result and the physical continuity at the boundary between the expanded region and the original observation region. The model is then trained using the constructed dataset. The wave extension and verification module is configured to extend the local wave field observed by shipborne radar using a trained wave perception domain hierarchical extension model, and to verify the reliability of the extension results using accuracy indicators. Among them, any one or more of the model dataset construction module, the wave perception domain progressively expanded model construction and training module, and the wave expansion and verification module are implemented by the processor executing computer program instructions stored in the memory.

Images