CN115438571A - Ground wave radar wave field calculation method and device based on machine learning - Google Patents
Ground wave radar wave field calculation method and device based on machine learning Download PDFInfo
- Publication number
- CN115438571A CN115438571A CN202210931135.4A CN202210931135A CN115438571A CN 115438571 A CN115438571 A CN 115438571A CN 202210931135 A CN202210931135 A CN 202210931135A CN 115438571 A CN115438571 A CN 115438571A
- Authority
- CN
- China
- Prior art keywords
- data
- training
- marine environment
- radar echo
- neural network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Medical Informatics (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Geometry (AREA)
- Computer Hardware Design (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention discloses a method and a device for calculating a wave field of a ground wave radar based on machine learning, wherein the method comprises the following steps: the method comprises the steps of obtaining radar echo data and marine environment data, and conducting data preprocessing on the radar echo data and the marine environment data to obtain training data; determining first radar echo data, first sea surface wind field data and first marine environment data according to the training data; performing feature extraction on the first marine environment data through a first neural network to obtain spatial feature information; training the first radar echo data, the first sea surface wind field data and the spatial feature information through a second neural network to generate a wave field inversion model; and performing wave inversion calculation according to the wave field inversion model. The method realizes the extraction of the spatial characteristic information by combining the marine environment data of the environmental influence factors, improves the inversion precision of the wave field, and can be widely applied to the technical field of artificial intelligence.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a method and a device for calculating a wave field of a ground wave radar based on machine learning.
Background
The early detection of the sea waves mainly depends on manual visual inspection, and with the development of science and technology, the current sea wave buoy detection method is the most widely used method in national ocean stations. However, buoys are easily affected by severe sea conditions, and some inherent problems are difficult to fundamentally solve, so that the detection of sea wave parameters gradually develops to digitization and intellectualization. The inversion of the wave height of the ground wave radar is mainly based on the extraction of wave information from echo data, and because the relationship between echo parameters and the wave height is difficult to be expressed by a simple analytical expression, the inversion method is continuously improved.
At present, a high-frequency ground wave radar wave field forming method based on a first-order peak ratio exists, a first-order peak area in echo parameters of a high-frequency ground wave radar is selected, the arrival angle of a first-order peak and a second-order peak ratio in the area is calculated, and a model for solving the wave height by using the power ratio is established by combining buoy wave height data of corresponding points.
However, most of the existing inversion algorithms are based on the electronic information theory, and the wave field and the echo data are not in a simple analytical relationship, so that the obtained result is easy to generate errors due to the fact that the formulated analytical formula lacks some influence factors.
Disclosure of Invention
In view of this, the embodiment of the invention provides a method and a device for calculating a ground wave radar wave field based on machine learning, which have high inversion accuracy.
In a first aspect, an embodiment of the present invention provides a method for calculating a wave field of a ground wave radar based on machine learning, including:
acquiring radar echo data and marine environment data, and performing data preprocessing on the radar echo data and the marine environment data to obtain training data; the radar echo data comprise radar first-order peak data and radar second-order peak data, and the marine environment data comprise sea surface wind field data and underwater terrain elevation data; the training data comprises a training set and a test set;
determining first radar echo data, first sea surface wind field data and first marine environment data according to the training data; the first radar echo data are radar echo data of a target calculation point, the first sea surface wind field data are sea surface wind field data of the target calculation point, and the first marine environment data are first marine environment data;
performing feature extraction on the first marine environment data through a first neural network to obtain spatial feature information;
training the first radar echo data, the first sea surface wind field data and the spatial feature information through a second neural network to generate a wave field inversion model;
and performing wave inversion calculation according to the wave field inversion model.
Optionally, the preprocessing the radar echo data and the marine environment data to obtain training data includes:
performing normalization processing on the radar echo data and the marine environment data to obtain training data;
the expression of the normalization processing is as follows:
wherein, X i Represents the normalized result; x original Representing the raw data in the input item; x max A maximum value representing the original data; x min Representing the minimum of the raw data.
Optionally, the performing data preprocessing on the radar echo data and the marine environment data to obtain training data further includes:
randomly distributing training data; wherein, 80% of the training data is used as a training set, and the rest 20% of the training data is used as a test set.
Optionally, the performing feature extraction on the first marine environment data through a first neural network to obtain spatial feature information includes:
carrying out convolution processing on the first sea surface wind field data and the underwater topography elevation data through a convolution layer to obtain a first characteristic;
cooling the first characteristic through the pooling layer to obtain a second characteristic;
leveling the second characteristic to obtain spatial characteristic information; the spatial feature information is a one-dimensional array;
wherein the first neural network comprises two convolutional layers and one pooling layer.
Optionally, the performing feature extraction on the first marine environment data through a first neural network to obtain spatial feature information further includes:
determining the number and the size of convolution kernels of the convolutional layer according to the training set;
and performing boundary filling processing on the first sea surface wind field data and the underwater terrain elevation value data in a same padding mode.
Optionally, the training processing is performed on the first radar echo data, the first sea surface wind field data, and the spatial feature information through a second neural network, so as to generate a wave field inversion model, including:
inputting the first radar echo data, the first sea surface wind field data and the spatial feature information into an input layer of the second neural network to serve as input item data;
selecting a Tansig function as a transfer function from an input layer to a hidden layer, and using a Purelin function as a transfer function from the hidden layer to an output layer;
respectively using a Trainlm function, a Trainbr function and a Traincg function to train according to the input item data, and selecting a target training function of which the training result and the test result both meet preset requirements;
calculating the neuron number of the hidden layer;
correcting weights and thresholds among layers of the second neural network through forward propagation or error backward propagation in the second neural network according to the target training function until a target minimum error is reached;
and training the second neural network according to the input item data, determining the network structure of the second neural network, and generating a wave field inversion model.
Optionally, the calculation formula of the neuron number of the hidden layer is as follows:
wherein, p represents the number of neurons of the hidden layer; m represents the number of neurons in the input layer; n represents the number of neurons in the output layer.
In a second aspect, an embodiment of the present invention provides a ground wave radar wave field computing apparatus based on machine learning, including:
the system comprises a first module, a second module and a third module, wherein the first module is used for acquiring radar echo data and marine environment data and preprocessing the radar echo data and the marine environment data to obtain training data; the radar echo data comprise radar first-order peak data and radar second-order peak data, and the marine environment data comprise sea surface wind field data and underwater terrain elevation data; the training data comprises a training set and a test set;
a second module for determining first radar echo data and first sea surface wind field data and first marine environment data from the training data; the first radar echo data are radar echo data of a target calculation point, the first sea surface wind field data are sea surface wind field data of the target calculation point, and the first marine environment data are first marine environment data;
the third module is used for extracting the characteristics of the first marine environment data through a first neural network to obtain spatial characteristic information;
the fourth module is used for training the first radar echo data, the first sea surface wind field data and the spatial characteristic information through a second neural network to generate a wave field inversion model;
and the fifth module is used for carrying out wave inversion calculation according to the wave field inversion model.
In a third aspect, an embodiment of the present invention provides an electronic device, including a processor and a memory;
the memory is used for storing programs;
the processor executes the program to implement the method as described above.
In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, where a program is stored, and the program is executed by a processor to implement the method as described above.
The embodiment of the invention also discloses a computer program product or a computer program, which comprises computer instructions, and the computer instructions are stored in a computer readable storage medium. The computer instructions may be read by a processor of a computer device from a computer-readable storage medium, and the computer instructions executed by the processor cause the computer device to perform the foregoing method.
The embodiment of the invention firstly acquires radar echo data and marine environment data, and carries out data preprocessing on the radar echo data and the marine environment data to obtain training data; subsequently determining first radar echo data and first sea surface wind field data and first marine environment data from the training data; then, carrying out feature extraction on the first marine environment data through a first neural network to obtain spatial feature information; then training the first radar echo data, the first sea surface wind field data and the spatial characteristic information through a second neural network to generate a wave field inversion model; and finally, performing wave inversion calculation according to the wave field inversion model. The method realizes the extraction of the spatial characteristic information by combining the marine environment data of the environmental influence factors, and improves the inversion precision of the wave field.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the description below are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
FIG. 1 is a flowchart illustrating the overall steps provided by an embodiment of the present invention;
FIG. 2 is a flowchart illustrating the principle steps provided by an embodiment of the present invention;
FIG. 3 is a diagram of a model framework of a first neural network provided by an embodiment of the present invention;
FIG. 4 is a model framework diagram of a second neural network provided by an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
In view of the problems in the prior art, in a first aspect, the present invention provides a method for calculating a wave field of a ground wave radar based on machine learning, as shown in fig. 1, the method includes the following steps:
acquiring radar echo data and marine environment data, and performing data preprocessing on the radar echo data and the marine environment data to obtain training data; the radar echo data comprise radar first-order peak data and radar second-order peak data, and the marine environment data comprise sea surface wind field data and underwater terrain elevation data; the training data comprises a training set and a test set;
determining first radar echo data, first sea surface wind field data and first marine environment data according to the training data; the first radar echo data are radar echo data of a target calculation point, the first sea surface wind field data are sea surface wind field data of the target calculation point, and the first marine environment data are first marine environment data;
performing feature extraction on the first marine environment data through a first neural network to obtain spatial feature information;
training and processing the first radar echo data, the first sea surface wind field data and the spatial characteristic information through a second neural network to generate a wave field inversion model;
and performing wave inversion calculation according to the wave field inversion model.
Optionally, the performing data preprocessing on the radar echo data and the marine environment data to obtain training data includes:
performing normalization processing on the radar echo data and the marine environment data to obtain training data;
the expression of the normalization processing is as follows:
wherein X i Represents the normalized result; x original Representing the raw data in the input item; x max A maximum value representing the original data; x min Representing the minimum of the raw data.
Optionally, the preprocessing the radar echo data and the marine environment data to obtain training data further includes:
randomly distributing training data; wherein 80% of the training data is used as a training set, and the rest 20% of the training data is used as a test set.
Optionally, the performing feature extraction on the first marine environment data through a first neural network to obtain spatial feature information includes:
carrying out convolution processing on the first sea surface wind field data and the underwater topography elevation data through a convolution layer to obtain a first characteristic;
cooling the first characteristic through the pooling layer to obtain a second characteristic;
leveling the second characteristic to obtain spatial characteristic information; the spatial feature information is a one-dimensional array;
wherein the first neural network includes two convolutional layers and one pooling layer.
Optionally, the performing feature extraction on the first marine environment data through a first neural network to obtain spatial feature information further includes:
determining the number and the size of convolution kernels of the convolution layer according to the training set;
and performing boundary filling processing on the first sea surface wind field data and the underwater terrain elevation data in a same padding mode.
Optionally, the training processing is performed on the first radar echo data, the first sea surface wind field data, and the spatial feature information through a second neural network, so as to generate a wave field inversion model, including:
inputting the first radar echo data, the first sea surface wind field data and the spatial feature information into an input layer of the second neural network to serve as input item data;
selecting a Tansig function as a transfer function from an input layer to a hidden layer, and using a Purelin function as a transfer function from the hidden layer to an output layer;
respectively using a Trainlm function, a Trainbr function and a Traincg function to train according to the input item data, and selecting a target training function of which the training result and the test result both meet preset requirements;
calculating the neuron number of the hidden layer;
correcting weights and thresholds among layers of the second neural network through forward propagation or error backward propagation in the second neural network according to the target training function until a target minimum error is reached;
and training the second neural network according to the input item data, determining the network structure of the second neural network, and generating a wave field inversion model.
Optionally, the calculation formula of the neuron number of the hidden layer is as follows:
wherein, p represents the number of neurons of the hidden layer; m represents the number of neurons of the input layer; n represents the number of neurons in the output layer.
In a second aspect, an embodiment of the present invention provides a device for calculating a ground wave radar wave field based on machine learning, including:
the system comprises a first module, a second module and a third module, wherein the first module is used for acquiring radar echo data and marine environment data and preprocessing the radar echo data and the marine environment data to obtain training data; the radar echo data comprise radar first-order peak data and radar second-order peak data, and the marine environment data comprise sea surface wind field data and underwater terrain elevation data; the training data comprises a training set and a test set;
a second module to determine first radar echo data and first sea surface wind field data and first marine environment data from the training data; the first radar echo data are radar echo data of a target calculation point, the first sea surface wind field data are sea surface wind field data of the target calculation point, and the first marine environment data are first marine environment data;
the third module is used for extracting the characteristics of the first marine environment data through a first neural network to obtain spatial characteristic information;
the fourth module is used for training the first radar echo data, the first sea surface wind field data and the spatial characteristic information through a second neural network to generate a wave field inversion model;
and the fifth module is used for carrying out wave inversion calculation according to the wave field inversion model.
The content of the method embodiment of the invention is all applicable to the device embodiment, the functions specifically realized by the device embodiment are the same as those of the method embodiment, and the beneficial effects achieved by the device embodiment are also the same as those achieved by the method.
Another aspect of the embodiments of the present invention further provides an electronic device, which includes a processor and a memory;
the memory is used for storing programs;
the processor executes the program to implement the method as described above.
The contents of the embodiment of the method of the invention are all applicable to the embodiment of the electronic device, the functions specifically realized by the embodiment of the electronic device are the same as those of the embodiment of the method, and the beneficial effects achieved by the embodiment of the electronic device are also the same as those achieved by the method.
Yet another aspect of the embodiments of the present invention provides a computer-readable storage medium, which stores a program, which is executed by a processor to implement the method as described above.
The contents of the embodiment of the method of the present invention are all applicable to the embodiment of the computer-readable storage medium, the functions specifically implemented by the embodiment of the computer-readable storage medium are the same as those of the embodiment of the method described above, and the advantageous effects achieved by the embodiment of the computer-readable storage medium are also the same as those achieved by the method described above.
Embodiments of the present invention also disclose a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The computer instructions may be read by a processor of a computer device from a computer-readable storage medium, and the computer instructions executed by the processor cause the computer device to perform the foregoing method.
The implementation principle of the ground wave radar flow field calculation method of the invention is described in detail as follows:
in order to fully explain the implementation principle of the present invention, the related technical contents are explained first:
BP neural network: the BP neural network is a multi-layer feedforward network trained according to error back propagation and comprises an input layer, a hidden layer and an output layer. If the actual output of the forward propagation does not match the expected output, the error is propagated backwards and the weights and offsets in the network are adjusted to change the loss function value.
A convolutional neural network: convolutional Neural Networks (CNNs) are a class of feed-forward Neural Networks that include convolution calculations and have a deep structure, and are one of the representative algorithms for deep learning. The convolutional neural network has the characteristic learning ability and can carry out translation invariant classification on input information according to the hierarchical structure of the convolutional neural network.
Ground wave radar: ground wave radar is a main sea detection means. The detection principle is that high-frequency electric waves are emitted by utilizing the characteristic of small diffraction propagation attenuation on the surface of the conductive ocean, so that a target beyond 300 kilometers can be detected by breaking through the horizon, and the detection precision is high.
Aiming at the problems in the prior art, the invention aims to solve the problem that physical ocean factors, echo signals of a ground wave radar and spatial information characteristics are added simultaneously in machine learning to obtain a more accurate and smooth ocean wave field.
It should be noted that the wave field is smooth, and the wave height at a certain point is not only the comprehensive representation of the various influencing factors at the point, but also influenced by the surrounding area.
Therefore, the present invention is directed to a technique for inverting a wave field by using spatial information in combination with radar echo data and actual marine environment information. The invention has the following characteristics:
1. in order to improve the inversion calculation method of ocean currents, the method simultaneously uses the ground wave radar echo data and the actual marine environment data.
2. Different from the traditional method for inverting the wave field through the information of the first-order peak, the second-order spectrum or the first-second-order peak of the ground wave radar, the technology adopts a machine learning method, and combines a BP neural network and a convolution neural network CNN to extract the spatial information influencing the wave field, so that the inversion accuracy is improved.
Specifically, as shown in fig. 2, the present invention comprises the steps of:
the method comprises the following steps: the data used by the training model are long-time sea surface wind field data, underwater terrain elevation value data, and first-order peak and second-order peak data of the radar. These data were normalized:
wherein X i Represents the normalized result; x original Representing the raw data in the input item; x max A maximum value representing the raw data; x min Representing the minimum of the raw data.
The training data is randomly divided into two parts, 80% and 20%, which are used as a training set and a test set respectively for training network parameters and testing network performance.
Step two: as shown in fig. 3, spatial information is extracted using a CNN convolutional neural network, and the specific refinement steps are as follows:
step1: and carrying out convolution on the wind field data and the underwater terrain elevation data of the area around the calculation point. The number of convolution kernels is usually chosen to be a power exponent of 2, and the training set is experimented with to determine the appropriate number of convolution kernels.
Step2: the convolution kernel size is determined. The sizes of the convolution kernels are 1 × 1, 3 × 3, 5 × 5 and 7 × 7, and the sizes of the convolution kernels are determined according to the sizes of input dimensions. The convolution kernel size of the present invention is chosen to be 3 x 3.
Step3: in order to keep the generated feature map size the same as the input size and improve the utilization of the input grid peripheral data, 0 padding is performed on the input boundary in a same padding manner. The convolution step is set to 1 and two convolutions are performed.
Step4: the model is downsampled (pooled) to reduce the dimensions of the features. Selecting an average pooling method.
Step5: and leveling the obtained characteristic graph into a one-dimensional array, so that the characteristic graph can be conveniently input into the BP neural network in the third step.
Step three: as shown in fig. 4, the radar first-order peak, the second-order peak data, the wind field data of the calculated point and the spatial feature information of the surrounding area extracted by using the CNN in the second step are given to a BP neural network together, a wave field inversion model is trained, and the specific refining steps are as follows:
firstly, it should be noted that, in modeling of the BP neural network, the BP algorithm includes two processes of forward propagation and back propagation of errors: we promise to use
Representing the connection weight of the kth neuron in the (L-1) th layer of the network to the jth neuron in the L-th layer
To indicate deviation, by
The activation function output of the kth neuron in the (L-1) th layer of the network is represented, f represents the activation function, and therefore, the calculation mode of the jth neuron in the L-1 th layer can be represented as
Generally, the forward propagation process is calculated as follows:
x [l] ＝f(w [l] x [l-1] +b [l] )
wherein, w [l] Connection weight, x, representing the forward propagation process [l-1] Representing the activation function output of network layer (L-1) neurons, b [l] The deviation is indicated.
The loss function is:
wherein: j. k represents the neuron numbers of the L-th and L-1-th layers, respectively, y j The output value of the neuron is represented,
mean of all neurons is indicated. If the actual output of the forward propagation does not match the expected output, the error is propagated backwards, which can be done by the following 4 sets of equations.
Equation 1 output layer error:
wherein, delta [l] Representing the output layer error, f representing the activation function, z [l] An output value indicating the output layer, indicates the inner product. And adding perturbation to the jth neuron at the L level
The loss function is reduced and the best perturbation is found to minimize the loss function value.
Equation 2 hidden layer error:
δ [l] ＝w [l+1]T x [l+1] ⊙f(z [l] )
wherein, delta [l] Indicating the hidden layer error, w [l+1] Denotes the connection weight of the l +1 layer, T denotes the matrix transpose, x [l+1] Representing the function output of the l +1 layer, f representing the activation function, z [l] Representing the output value of the output layer.
Equation 3 rate of change of parameter:
wherein the content of the first and second substances,
which represents the partial differential of the L and,
denotes b [l] Partial differential of (d), δ [l] Indicating the error in the hidden layer or layers,
partial differential, alpha, representing the link weight [l+1] Denotes the learning rate, T denotes the matrix transposition, α [l+1]T Matrix transpose representing learning rate
Equation 4 parameter update rule:
where α is the learning rate, which specifies the step size of the gradient descent during back propagation. The back propagation of the error uses a gradient descent method, where the gradient represents the direction in which the function increases the fastest, so that the minimum of the function can be found faster in the opposite direction. In the back propagation process, the output error is reversely transmitted to the input layer by layer through the hidden layer, the error is distributed to each unit, the loss function value is changed by adjusting the weight w and the deviation b in the network, and the minimum value of the function L (y) is obtained. The above process completes the modeling of the BP basic framework.
Step1: the input data comprises first-order peak data and second-order peak data of the ground wave radar, an underwater terrain elevation value, a wind field and characteristic information obtained in the second step.
Step2: selecting a Tansig function as a transfer function from an input layer to a hidden layer, and using a Purelin function as a transfer function from the hidden layer to an output layer.
Step3: and (3) selecting a training function of the network, respectively carrying out experiments on the Trainlm function, the Trainbr function and the Trainscg function, and selecting a function with ideal training and testing results.
Step4: the neuron number of the hidden layer is determined, and the empirical formula can be used as follows:
wherein p, m and n are the numbers of neurons in the hidden layer, the input layer and the output layer respectively, and q is a constant of 1-10. And finding the best state that the performance of the training set and the test set is close to the performance from large to small.
Step5: after the training function is determined, the neural network continuously corrects the weight and the threshold value among all layers of the network in a mode of forward propagation and error backward propagation of signals in the network until a target minimum error is reached.
Step6: after the steps, a final network structure is determined, all input training data are used as a training set to train a neural network, and when the target minimum error is achieved and the relative stability is kept, the training is finished, so that a wave field inversion model is generated.
Step7: evaluation indexes are as follows: mean Absolute Error (MAE), root Mean Square Error (RMSE), index of Agent (IA), and interpretive Variance Score (Var) were selected. Wherein n represents the number of samples, x i Which is representative of the observed value(s),
representing the predicted value.
Step four: and performing wave inversion by using the model obtained by training.
In summary, compared with the prior art, the invention has the following outstanding advantages:
1. high accuracy: the invention uses a machine learning method, which is beneficial to simulating a nonlinear process and obtaining a high-precision result. The seawater is continuous, adding the extraction of spatial features helps to improve accuracy and smooth the data results.
2. High coverage rate: the data production of radar is limited by technology, and most of the available data is the central area range of radar echo. According to the method, a plurality of influence factors influencing the inversion result are increased, the dependence on radar echo data in the inversion is reduced, and therefore the invertible sea area of the radar data is enlarged.
In alternative embodiments, the functions/acts noted in the block diagrams may occur out of the order noted in the operational illustrations. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality/acts involved. Furthermore, the embodiments presented and described in the flow charts of the present invention are provided by way of example in order to provide a more thorough understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is changed and in which sub-operations described as part of larger operations are performed independently.
Furthermore, although the present invention is described in the context of functional modules, it should be understood that, unless otherwise stated to the contrary, one or more of the described functions and/or features may be integrated in a single physical device and/or software module, or one or more functions and/or features may be implemented in a separate physical device or software module. It will also be appreciated that a detailed discussion of the actual implementation of each module is not necessary for an understanding of the present invention. Rather, the actual implementation of the various functional modules in the apparatus disclosed herein will be understood within the ordinary skill of an engineer, given the nature, function, and internal relationship of the modules. Accordingly, those of ordinary skill in the art will be able to practice the invention as set forth in the claims without undue experimentation. It is also to be understood that the specific concepts disclosed are merely illustrative of and not intended to limit the scope of the invention, which is to be determined from the appended claims along with their full scope of equivalents.
The functions may be stored in a computer-readable storage medium if they are implemented in the form of software functional units and sold or used as separate products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
The logic and/or steps represented in the flowcharts or otherwise described herein, such as an ordered listing of executable instructions that can be considered to implement logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device.
More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium could even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via for instance optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner if necessary, and then stored in a computer memory.
It should be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, various steps or methods may be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or combination of the following technologies, which are well known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
While embodiments of the invention have been shown and described, it will be understood by those of ordinary skill in the art that: various changes, modifications, substitutions and alterations can be made to the embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.
While the preferred embodiments of the present invention have been illustrated and described, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (10)
1. A method for calculating a wave field of a ground wave radar based on machine learning is characterized by comprising the following steps:
acquiring radar echo data and marine environment data, and performing data preprocessing on the radar echo data and the marine environment data to obtain training data; the radar echo data comprise radar first-order peak data and radar second-order peak data, and the marine environment data comprise sea surface wind field data and underwater terrain elevation data; the training data comprises a training set and a test set;
determining first radar echo data, first sea surface wind field data and first marine environment data according to the training data; the first radar echo data are radar echo data of a target calculation point, the first sea surface wind field data are sea surface wind field data of the target calculation point, and the first marine environment data are first marine environment data;
performing feature extraction on the first marine environment data through a first neural network to obtain spatial feature information;
training the first radar echo data, the first sea surface wind field data and the spatial feature information through a second neural network to generate a wave field inversion model;
and performing wave inversion calculation according to the wave field inversion model.
2. The method for calculating the wave field of the ground wave radar based on the machine learning of claim 1, wherein the preprocessing the radar echo data and the marine environment data to obtain training data comprises:
carrying out normalization processing on the radar echo data and the marine environment data to obtain training data;
the expression of the normalization process is:
wherein X i Represents the normalized result; x original Representing the raw data in the input item; x max A maximum value representing the raw data; x min Representing the minimum of the raw data.
3. The method for calculating the wave field of the ground wave radar based on the machine learning of claim 1, wherein the preprocessing the radar echo data and the marine environment data to obtain training data further comprises:
randomly distributing training data; wherein 80% of the training data is used as a training set, and the rest 20% of the training data is used as a test set.
4. The method according to claim 1, wherein the performing feature extraction on the first marine environment data through a first neural network to obtain spatial feature information comprises:
carrying out convolution processing on the first sea surface wind field data and the underwater topography elevation data through a convolution layer to obtain a first characteristic;
cooling the first characteristic through the pooling layer to obtain a second characteristic;
leveling the second characteristic to obtain spatial characteristic information; the spatial feature information is a one-dimensional array;
wherein the first neural network includes two convolutional layers and one pooling layer.
5. The method according to claim 4, wherein the performing feature extraction on the first marine environment data through a first neural network to obtain spatial feature information further comprises:
determining the number and the size of convolution kernels of the convolutional layer according to the training set;
and performing boundary filling processing on the first sea surface wind field data and the underwater terrain elevation data in a same padding mode.
6. The method according to claim 1, wherein the training processing is performed on the first radar echo data, the first sea surface wind field data and the spatial feature information through a second neural network to generate a wave field inversion model, and the method comprises:
inputting the first radar echo data, the first sea surface wind field data and the spatial feature information into an input layer of the second neural network to serve as input item data;
selecting a Tansig function as a transfer function from an input layer to a hidden layer, and using a Purelin function as a transfer function from the hidden layer to an output layer;
respectively using a Trainlm function, a Trainbr function and a Traincg function to train according to the input item data, and selecting a target training function of which the training result and the test result both meet preset requirements;
calculating the neuron number of the hidden layer;
correcting weights and thresholds among layers of the second neural network through forward propagation or error backward propagation in the second neural network according to the target training function until a target minimum error is reached;
and training the second neural network according to the input item data, determining the network structure of the second neural network, and generating a wave field inversion model.
7. The method of claim 6, wherein the method comprises the steps of,
the calculation formula of the neuron number of the hidden layer is as follows:
wherein, p represents the number of neurons of the hidden layer; m represents the number of neurons of the input layer; n represents the number of neurons in the output layer.
8. A ground wave radar wave field computing device based on machine learning, comprising:
the system comprises a first module, a second module and a third module, wherein the first module is used for acquiring radar echo data and marine environment data and preprocessing the radar echo data and the marine environment data to obtain training data; the radar echo data comprise radar first-order peak data and radar second-order peak data, and the marine environment data comprise sea surface wind field data and underwater terrain elevation data; the training data comprises a training set and a test set;
a second module for determining first radar echo data and first sea surface wind field data and first marine environment data from the training data; the first radar echo data are radar echo data of a target calculation point, the first sea surface wind field data are sea surface wind field data of the target calculation point, and the first marine environment data are first marine environment data;
the third module is used for extracting the characteristics of the first marine environment data through a first neural network to obtain spatial characteristic information;
the fourth module is used for training the first radar echo data, the first sea surface wind field data and the spatial characteristic information through a second neural network to generate a wave field inversion model;
and the fifth module is used for carrying out wave inversion calculation according to the wave field inversion model.
9. An electronic device comprising a processor and a memory;
the memory is used for storing programs;
the processor executing the program implements the method of any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that the storage medium stores a program, which is executed by a processor to implement the method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210931135.4A CN115438571B (en) | 2022-08-04 | 2022-08-04 | Ground wave radar wave field calculation method and device based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210931135.4A CN115438571B (en) | 2022-08-04 | 2022-08-04 | Ground wave radar wave field calculation method and device based on machine learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115438571A true CN115438571A (en) | 2022-12-06 |
| CN115438571B CN115438571B (en) | 2023-06-30 |
Family
ID=84242956
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210931135.4A Active CN115438571B (en) | 2022-08-04 | 2022-08-04 | Ground wave radar wave field calculation method and device based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115438571B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115932838A (en) * | 2022-12-12 | 2023-04-07 | 中山大学 | Ground wave radar and navigation observation data correction method based on neural network |
| CN115983141A (en) * | 2023-03-21 | 2023-04-18 | 国家***北海预报中心((国家***青岛海洋预报台)(国家***青岛海洋环境监测中心站)) | Method, medium and system for inverting wave height based on deep learning |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102103708A (en) * | 2011-01-28 | 2011-06-22 | 哈尔滨工程大学 | Radial basis function neural network-based wave significant wave height inversion model establishment method |
| CN103870676A (en) * | 2014-01-23 | 2014-06-18 | 中国石油大学(华东) | Altimeter sea wave mean wave period inversion method suitable for Chinese offshore area |
| CN109116320A (en) * | 2018-07-13 | 2019-01-01 | 中国人民解放军国防科技大学 | Sea wave characteristic parameter extraction method based on radar echo signals |
| CN109782287A (en) * | 2019-03-12 | 2019-05-21 | 国家***北海预报中心 | A method of it is surged wave height by high-frequency ground wave radar inverting |
| WO2021218424A1 (en) * | 2020-04-30 | 2021-11-04 | 江苏科技大学 | Rbf neural network-based method for sea surface wind speed inversion from marine radar image |
| CN114167076A (en) * | 2021-11-10 | 2022-03-11 | 中国地质大学（武汉） | Method for establishing sea surface wind speed and direction inversion model and wind speed and direction inversion method |
| CN114488143A (en) * | 2021-12-13 | 2022-05-13 | 北京电子工程总体研究所 | Ocean eddy rotating force parameter inversion method and system in SAR image |
-
2022
- 2022-08-04 CN CN202210931135.4A patent/CN115438571B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102103708A (en) * | 2011-01-28 | 2011-06-22 | 哈尔滨工程大学 | Radial basis function neural network-based wave significant wave height inversion model establishment method |
| CN103870676A (en) * | 2014-01-23 | 2014-06-18 | 中国石油大学(华东) | Altimeter sea wave mean wave period inversion method suitable for Chinese offshore area |
| CN109116320A (en) * | 2018-07-13 | 2019-01-01 | 中国人民解放军国防科技大学 | Sea wave characteristic parameter extraction method based on radar echo signals |
| CN109782287A (en) * | 2019-03-12 | 2019-05-21 | 国家***北海预报中心 | A method of it is surged wave height by high-frequency ground wave radar inverting |
| WO2021218424A1 (en) * | 2020-04-30 | 2021-11-04 | 江苏科技大学 | Rbf neural network-based method for sea surface wind speed inversion from marine radar image |
| CN114167076A (en) * | 2021-11-10 | 2022-03-11 | 中国地质大学（武汉） | Method for establishing sea surface wind speed and direction inversion model and wind speed and direction inversion method |
| CN114488143A (en) * | 2021-12-13 | 2022-05-13 | 北京电子工程总体研究所 | Ocean eddy rotating force parameter inversion method and system in SAR image |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115932838A (en) * | 2022-12-12 | 2023-04-07 | 中山大学 | Ground wave radar and navigation observation data correction method based on neural network |
| CN115932838B (en) * | 2022-12-12 | 2023-11-21 | 中山大学 | Data correction method for ground wave radar and navigation observation based on neural network |
| CN115983141A (en) * | 2023-03-21 | 2023-04-18 | 国家***北海预报中心((国家***青岛海洋预报台)(国家***青岛海洋环境监测中心站)) | Method, medium and system for inverting wave height based on deep learning |
| CN115983141B (en) * | 2023-03-21 | 2023-07-14 | 国家***北海预报中心((国家***青岛海洋预报台)(国家***青岛海洋环境监测中心站)) | Deep learning-based method, medium and system for inverting wave height of ocean waves |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115438571B (en) | 2023-06-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115438571A (en) | Ground wave radar wave field calculation method and device based on machine learning | |
| CN105654483B (en) | The full-automatic method for registering of three-dimensional point cloud | |
| CN115204058B (en) | Ground wave radar flow field calculation method and device based on BP neural network | |
| CN108960421B (en) | Improved online forecasting method for speed of unmanned surface vehicle based on BP neural network | |
| CN106599129A (en) | Multi-beam point cloud data denoising method considering terrain characteristics | |
| Kawai et al. | Sea state estimation using monitoring data by convolutional neural network (CNN) | |
| CN103617344A (en) | Single-layer earth surface dielectric parameter and roughness parameter fast inversion combined optimization algorithm based on measured radar back scattering data | |
| CN113553755B (en) | Power system state estimation method, device and equipment | |
| CN112180369B (en) | Depth learning-based sea surface wind speed inversion method for one-dimensional synthetic aperture radiometer | |
| CN107255920A (en) | PID control method and apparatus and system based on network optimization algorithm | |
| CN112069723A (en) | Method and system for evaluating transient stability of power system | |
| Tian et al. | Joint learning model for underwater acoustic target recognition | |
| CN107292039B (en) | UUV bank patrolling profile construction method based on wavelet clustering | |
| CN115859116A (en) | Marine environment field reconstruction method based on radial basis function regression interpolation method | |
| CN109961512A (en) | The airborne data reduction method and device of landform | |
| CN113657595A (en) | Neural network real-time pruning method and system and neural network accelerator | |
| CN117037450A (en) | Environment early warning method, device, equipment and medium of intelligent piling ship | |
| Noever-Castelos et al. | Model updating of a wind turbine blade finite element Timoshenko beam model with invertible neural networks | |
| Zhang et al. | Root quantization: a self-adaptive supplement STE | |
| CN115994481A (en) | Ship motion attitude prediction method based on multiple combinations | |
| CN112434467A (en) | Low-target-strength shape design method for underwater rotating body target | |
| CN115687953B (en) | Target clustering method, medium and device based on Hilbert coding | |
| CN115688004B (en) | Target attribute determining method, medium and device based on Hilbert coding | |
| CN115619687B (en) | Tunnel lining void radar signal identification method, equipment and storage medium | |
| CN117610364A (en) | Super-surface design method, property prediction model training method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |