Wave spectrum estimation method and device based on dynamic dispersion factor

Abstract

The invention provides a wave spectrum estimation method and device based on a dynamic dispersion factor, and the method comprises the following steps: obtaining an ADCP speed sequence of continuous time, obtaining a vertical speed power spectrum based on the ADCP speed sequence, and obtaining complex value frequencies, wave numbers and depths of different frequency points; taking the complex value frequency, the wave number and the depth corresponding to the same frequency point as a group of input vectors, and inputting a plurality of groups of input vectors into a pre-trained dynamic dispersion factor prediction model in a batch input mode to obtain dynamic dispersion factors; and correcting the wave number of each frequency point by using the dynamic dispersion factor to obtain a corrected wave number, and obtaining the wave spectrum of the ADCP speed sequence based on the corrected wave number. According to the scheme, the dynamic dispersion factor is predicted by using complex value Transform network modeling, so that the frequency spectrum estimation precision and the wave parameter stability are improved under the condition of water depth change, and better technical support is provided for wave observation of complex water areas such as coastal areas and shoals.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of hydrological measurement, in particular to a wave spectrum estimation method and device based on a dynamic dispersion factor. BACKGROUND

[0002] As the core equipment for measuring water flow and wave parameters in oceans, lakes and nearshore areas, acoustic Doppler current profilers (ADCP) have been widely used in the field of hydrological measurement and ocean observation. The wave parameter estimation of ADCP usually relies on the frequency domain analysis of the vertical velocity time series, and the wave characteristics are quantified by deriving the wave height spectrum and related parameters. However, the traditional wave spectrum inversion method generally processes different water depths in segments. In deep water, one formula is used to calculate the dispersion, while in shallow water, another formula is used to calculate the dispersion. This segmented processing method is difficult to adapt to the actual complex water environment.

[0003] However, in the transition area between deep and shallow water such as nearshore, sandbar and shoal, the dynamic change of water depth and the irregularity of seabed topography will cause significant changes in wave number during wave propagation, and the nonlinear dispersion effect is more prominent. When the traditional method uses a fixed dispersion relationship to calculate the wave number, it cannot respond to the dynamic correlation between water depth and frequency in real time, resulting in spectrum shift, increased wave height estimation error and other problems. Since the ADCP measurement data needs to be obtained by frequency domain analysis (such as FFT) to obtain the wave frequency and then invert the wave number, the limitations of the fixed formula further amplify the errors caused by the changes of water depth and wavelength in actual sea conditions. Therefore, there is an urgent need for a dispersion relationship correction algorithm that can adapt to different water depth conditions to improve the accuracy and scene adaptability of wave spectrum estimation. SUMMARY

[0004] The embodiments of the present application provide a wave spectrum estimation method and device based on a dynamic dispersion factor. By using a complex-valued Transformer network modeling to predict the dynamic dispersion factor, the spectrum estimation accuracy and wave parameter stability are improved under the condition of water depth change, providing better technical support for wave observation in complex water areas such as nearshore and shoal.

[0005] In a first aspect, the embodiments of the present application provide a wave spectrum estimation method based on a dynamic dispersion factor, which comprises: Obtaining an ADCP velocity sequence of continuous time, obtaining a vertical velocity power spectrum based on the ADCP velocity sequence, obtaining a complex frequency of different frequency points based on the vertical velocity power spectrum, and obtaining a wave number and depth corresponding to each frequency point, and performing complex-valued processing on the wave number and depth; The complex-valued frequency, wave number and depth corresponding to the same frequency point are taken as an input vector, and multiple input vectors are input into a pre-trained dynamic dispersion factor prediction model in a batch input manner, the dynamic dispersion factor prediction model learns the correlation between the complex frequency, wave number and depth in each input vector based on a multi-head self-attention mechanism to obtain a dynamic dispersion factor; The wave number of each frequency point is corrected using the dynamic dispersion factor to obtain a corrected wave number, and a wave spectrum of an ADCP velocity sequence is obtained based on the corrected wave number, the wave spectrum including a wave height spectrum, an effective wave height and a main frequency period.

[0006] In a second aspect, an embodiment of the present application provides a wave spectrum estimation device based on a dynamic dispersion factor, including: An obtaining module is configured to obtain an ADCP velocity sequence of continuous time, obtain a vertical velocity power spectrum based on the ADCP velocity sequence, obtain a complex-valued frequency of different frequency points based on the vertical velocity power spectrum, further obtain a wave number and a depth corresponding to each frequency point, and perform complex-valued processing on the wave number and the depth; A prediction module takes the complex-valued frequency, wave number and depth corresponding to the same frequency point as an input vector, and inputs multiple input vectors into a pre-trained dynamic dispersion factor prediction model in a batch input manner, the dynamic dispersion factor prediction model learns the correlation between the complex frequency, wave number and depth in each input vector based on a multi-head self-attention mechanism to obtain a dynamic dispersion factor; An estimation module uses the dynamic dispersion factor to correct the wave number of each frequency point to obtain a corrected wave number, and obtains a wave spectrum of an ADCP velocity sequence based on the corrected wave number, the wave spectrum including a wave height spectrum, an effective wave height and a main frequency period.

[0007] In a third aspect, an embodiment of the present application provides an electronic device including a memory and a processor, the memory storing a computer program, and the processor being configured to run the computer program to perform a wave spectrum estimation method based on a dynamic dispersion factor.

[0008] In a fourth aspect, an embodiment of the present application provides a readable storage medium storing a computer program, the computer program including program code for controlling a process to perform a process, the process including a wave spectrum estimation method based on a dynamic dispersion factor.

[0009] The main contributions and innovations of the present application are as follows: The scheme effectively strips off interference signals and reduces spectral leakage by mean removal, band-pass filtering and window function weighting on ADCP velocity sequence, thereby laying a precise data foundation for subsequent analysis; the scheme uses complex-valued Transformer model to build a complex parameter model, completely retains the amplitude and phase correlation of wave signals, breaks through the information fragmentation limitation of traditional real-valued models, and improves the completeness of dispersion relation description; the scheme deeply mines the dynamic correlation between complex frequency, wave number and depth through multi-head self-attention mechanism, accurately captures the nonlinear dispersion effect in the shallow and deep water transition zone, and enhances the adaptability to complex water environment; the scheme uses dynamic dispersion factor to adaptively correct the wave number, optimizes the calculation accuracy of wave height spectrum, significant wave height and main frequency period, and solves the error amplification problem of traditional segmented dispersion formula under dynamic water depth conditions.

[0010] The details of one or more embodiments of the present application are presented in the following drawings and description to make other features, objects and advantages of the present application more apparent. BRIEF DESCRIPTION OF DRAWINGS

[0011] The drawings described herein are intended to provide further understanding of the present application, and constitute a part of the present application. The illustrative embodiments of the present application and their description serve to explain the present application, and do not constitute an improper limitation on the present application. In the drawings: Figure 1 is a flow chart of a wave spectrum estimation method based on a dynamic dispersion factor according to an embodiment of the present application; Figure 2 is a structural block diagram of a wave spectrum estimation device based on a dynamic dispersion factor according to an embodiment of the present application; Figure 3 is a hardware structure schematic diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0012] The exemplary embodiments will be described in detail herein with reference to the attached drawings. The following description is with reference to the drawings, in which like numerals represent like elements, unless otherwise specified. The embodiments described in the following exemplary embodiments do not represent all the embodiments consistent with one or more embodiments of the present specification. Rather, they are merely examples of devices and methods consistent with some aspects of one or more embodiments of the present specification, as detailed in the appended claims.

[0013] It should be noted that the steps of the corresponding method are not necessarily performed in the order shown and described in the specification in other embodiments. In some other embodiments, the steps included in the method thereof can be more or less than described in the specification. In addition, a single step described in the specification may, in other embodiments, be broken down into multiple steps for description; and multiple steps described in the specification may, in other embodiments, be combined into a single step for description.

[0014] Embodiment one The embodiment of the present application provides a wave spectrum estimation method based on a dynamic dispersion factor, which predicts the dynamic dispersion factor by using a complex-valued Transformer network model, thereby improving the spectrum estimation accuracy and wave parameter stability under the condition of water depth change, and providing better technical support for wave observation in complex water areas such as nearshore and shoal. Specifically, referring to Figure 1 , the method comprises: obtaining an ADCP velocity sequence of continuous time, obtaining a vertical velocity power spectrum based on the ADCP velocity sequence, obtaining a complex-valued frequency of different frequency points based on the vertical velocity power spectrum, and then obtaining a wave number and a depth corresponding to each frequency point, and performing complex-valued processing on the wave number and the depth; using the complex-valued frequency, the wave number and the depth corresponding to the same frequency point as an input vector, inputting multiple input vectors in a batch input manner into a pre-trained dynamic dispersion factor prediction model, and the dynamic dispersion factor prediction model learns the association relationship between the complex frequency, the wave number and the depth in each input vector based on a multi-head self-attention mechanism to obtain a dynamic dispersion factor; using the dynamic dispersion factor to correct the wave number of each frequency point to obtain a corrected wave number, and obtaining a wave spectrum of the ADCP velocity sequence based on the corrected wave number, wherein the wave spectrum comprises a wave height spectrum, an effective wave height and a main frequency period.

[0015] In some embodiments, an ADCP velocity sequence of continuous time is obtained from an acoustic Doppler profiler (ADCP), and in the process of obtaining the ADCP velocity sequence, a continuous vertical velocity time sequence is collected from an upward beam as the ADCP velocity sequence, and the sampling frequency in the collection process is not less than 1Hz and the sampling time is not less than 10 minutes.

[0016] In some embodiments, the ADCP velocity sequence is preprocessed, and then a vertical velocity power spectrum is obtained based on the preprocessed ADCP velocity sequence, and the preprocessing comprises mean removal operation, band-pass filtering operation and window function weighting operation.

[0017] Further, the purpose of the mean removal operation is to remove the direct current component (mean flow) in the ADCP velocity sequence and retain the oscillation part of the wave, and it is assumed that To obtain the ADCP velocity sequence, the mean value in the ADCP velocity sequence is obtained. The mean-removal operation is represented by the following formula:

[0018] in, This is the ADCP velocity sequence after the mean removal operation.

[0019] Furthermore, since the wave signal velocity oscillates around zero, the acoustic Doppler profiler will simultaneously measure low-frequency components such as tidal current and residual current when acquiring acoustic signals. These low-frequency components will affect the spectrum analysis. Therefore, this scheme uses bandpass filtering to retain the signal within the wave frequency range and filter out low-frequency (tidal current, etc.) and high-frequency (noise) components.

[0020] Specifically, in the bandpass filtering process, the ADCP velocity sequence after the mean removal operation is first subjected to a discrete Fourier transform, expressed by the following formula:

[0021] in, It is the Fourier transform result, which is the complex value representation of the k-th frequency point of the ADCP velocity sequence after the mean removal operation. It contains amplitude and phase information, and N is the total number of samples.

[0022] Specifically, let the sampling frequency be... Low cutoff frequency High cutoff frequency filter order Construct the frequency domain mask using the following formula:

[0023] in, For the center frequency, For bandwidth, Each frequency component in the corresponding spectrum , This is an intermediate value of the frequency domain mask, because in actual calculations, the frequency domain mask... Therefore, by The frequency domain mask can be obtained.

[0024] Then based on the obtained frequency domain mask Fourier transform results Processing is performed to obtain intermediate quantities. Then... The filtered result is obtained by performing an inverse Fourier transform, expressed by the following formula:

[0025] in, is a filtered result after inverse Fourier transform, is a time domain signal, k is a frequency point, and N is a total number of samples, is an intermediate quantity.

[0026] Specifically, since a typical sea wave period is 2-25 s, the corresponding frequency range is about 0.04-0.5 Hz, so the present scheme retains the time domain signal in the frequency range of 0.04-0.5 Hz through a band-pass filtering operation.

[0027] Further, the purpose of the window function weighting in the present scheme is to reduce the sidelobe leakage problem in the Fourier transform, thereby improving the spectral resolution. In the present scheme, since the Hann window has low spectral leakage, the Hann window is used to weight the filtered result. The formula of the window function is as follows:

[0028] where N is the number of samples, and n is the sample.

[0029] Specifically, the formula for weighting the filtered result using the window function is as follows:

[0030] wherein, is the result after window function weighting, i.e., the preprocessing result after preprocessing the ADCP velocity sequence, is the window function, is the filtered result.

[0031] In some embodiments, the ADCP velocity sequence is segmented, each segment corresponding to a frequency point, and the vertical velocity power spectrum of each frequency point is calculated by the following formula:

[0032] wherein, denotes the vertical velocity power spectrum of the kth frequency point, L is the total number of segments, l is the segment index, and FFT denotes the Fourier transform, is the ADCP velocity sequence of the lth segment, and U is the window energy correction factor.

[0033] Specifically, , is the number of samples in each segment, used to correct the attenuation of the power by the window function.

[0034] Further, the complex-valued frequency of each frequency point is obtained based on the vertical velocity power spectrum. Since the vertical velocity power spectrum is obtained by the Fourier transform, the vertical velocity power spectrum is in the form of complex values, and the complex-valued frequency obtained from the vertical velocity power spectrum is also in the form of complex values.

[0035] In other words, this scheme uses FFT to process the vertical power spectrum to obtain the complex frequency of each frequency point.

[0036] In some specific embodiments, the depth corresponding to each frequency point is directly obtained from the ADCP velocity sequence, and the wavenumber of each frequency point is obtained based on the dispersion relationship between the complex-valued frequency and the wavenumber at different frequencies. In wave propagation theory, the dispersion relationship describes the wavenumber. The relationship between the complex-valued frequency ω and the wavenumber dispersion when the water depth is greater than or equal to the threshold λ is as follows:

[0037] Where g is the gravitational acceleration constant, and ω is the complex frequency. For wave number.

[0038] When the water depth is less than the threshold λ, the dispersion relationship between the complex frequency and the wavenumber is as follows:

[0039] Where g is the gravitational acceleration constant, and ω is the complex frequency. For wave number, For depth.

[0040] Furthermore, wavenumbers are represented by complex linear mapping. And the depth h is complexified, expanding the real-valued input to... ,in This represents the wave number or water depth. Subsequently, a complex weight matrix is ​​constructed. ,in For trainable parameters, the complex embedding result is defined as follows: This mapping can automatically learn the nonlinear relationship between wavenumber, water depth, and complex frequency, and during training, the parameters... Joint optimization is performed through error backpropagation. This method allows for the optimization of real-valued inputs. The complex-valued vector form is converted to maintain consistency with the complex frequency features obtained by the Fourier transform, thereby ensuring that the complex-valued Transformer model performs a unified feature representation and modeling in the complex domain.

[0041] (1) Specifically, the definition of a complex-valued linear mapping is as follows: For input scalar Construct the extended vector

[0042] Here, a constant 1 is added as a bias input.

[0043] Define the complex weight matrix:

[0044] Then the complex embedding is:

[0045] Here denotes the complex number field.

[0046] (2) Decompose into real / imaginary parts In implementation, complex numbers are usually decomposed into real and imaginary parts:

[0047] Here denotes the real part operation, denotes the imaginary part operation.

[0048] (3) Complexify wave number : get complex embedding .

[0049] Complexify water depth : get complex embedding .

[0050] Concatenate with (complex frequency obtained by FFT):

[0051] And then project to the input dimension of Transformer through a layer of complex linear mapping .

[0052] Specifically, the benefits of complexifying wave number and depth are: 1) Form uniformity — all input features are in the complex number field, facilitating homogeneous processing in complex-valued attention and complex-valued feedforward layers.

[0053] (2) Parameters are learnable — the mapping matrix parameters are completely optimized by training, which can automatically adapt to different water depths and wave number intervals.

[0054] (3) Strong expressive power — complex linear layers can learn the coupling relationship between amplitude and phase, which helps to model the phase delay effect in the dispersion relationship.

[0055] In some embodiments, the complex-valued frequency and wave number dispersion relationship is segmented based on different water depths, i.e., the dispersion relationship is different when the water depth is greater than or equal to a threshold λ and when the water depth is less than the threshold λ. In actual situations, the depth of water is often smoothly transitioned, so the accuracy is affected when the above formula is used in the "deep and shallow water transition zone". Therefore, the present scheme predicts a dynamic dispersion factor by using a preset dynamic dispersion factor prediction model, and uses the dynamic dispersion factor to correct the dispersion relationship, so as to more accurately describe the dispersion relationship between the complex-valued frequency and the wave number.

[0056] Specifically, the dynamic dispersion factor in the present scheme is a correction function for correcting the relationship between the complex-valued frequency and the wave number.

[0057] In some embodiments, before the multi-head self-attention mechanism is performed, an input embedding layer is constructed, and the normalized wave number and depth are positionally encoded in the input embedding layer in a sine-cosine position encoding manner.

[0058] Specifically, the formula of the sine-cosine position encoding is as follows:

[0059] wherein p is the normalized wave number and depth, i is the encoding dimension index, d is the encoding dimension, and PE is the position encoding value, that is, when the encoding dimension is even, the sine encoding is performed, and when the encoding dimension is odd, the cosine encoding is performed.

[0060] In some embodiments, a complex-valued Transformer model is used as the dynamic dispersion factor prediction model, and all trainable parameters in the dynamic dispersion factor prediction model are in complex-valued form, i.e., each group of input vectors is also in complex-valued form.

[0061] Further, in each attention head of the multi-head self-attention mechanism, a complex-valued weight is constructed to linearly transform the input vector to obtain a complex-valued representation of the query vector, the key vector, and the value vector, and then attention calculation is performed based on the query vector, the key vector, and the value vector to obtain an attention result, which is expressed by the formula:

[0062] wherein, is the correlation of the i th query vector and the j th key vector, is the i th query vector, is the conjugate transpose of the j th key vector, is the dimension of the attention head, is used to take the real part of the complex number, is the normalization result of , and is the attention result, is the jth value.

[0063] Specifically, the formula for linearly transforming the input vector using complex-valued weights is:

[0064] wherein, is the real part of the complex-valued weight, is the imaginary part of the complex-valued weight, is the real part of the input vector, is the imaginary part of the input vector, is the real part of the complex-valued bias, is the imaginary part of the complex-valued bias, is the real part of the linearly transformed result, is the imaginary part of the linearly transformed result.

[0065] The linearly transformed result is then processed using the preset query matrix, key matrix, and value matrix to obtain a query vector, a key vector, and a value vector.

[0066] In some embodiments, the attention results of each attention head in the multi-head self-attention mechanism are integrated, and after residual connection, a LayerNorm layer, and a feedforward network, the output result of the encoder is obtained, wherein the feedforward network uses a modReLU activation function for output, and the formula of the modReLU activation function is as follows:

[0067] wherein, is the input of the activation function, is the modulus of z, , is the real part, is the imaginary part, and b is a learnable bias.

[0068] Specifically, the residual connection, LayerNorm layer, and feedforward network are all conventional results in the Transformer, and in the present scheme, the parameters in the residual connection, LayerNorm layer, and feedforward network are all in complex-valued form. Specifically, since the intermediate layers of the model involve complex-valued features, the modReLU activation function is used after the feedforward network to nonlinearly adjust the amplitude while preserving the phase information.

[0069] Specifically, the output head is used to output the output result of the encoder to obtain a dynamic dispersion factor , which is also in complex-valued form, wherein the real part of the dynamic dispersion factor is , and the imaginary part is .

[0070] Specifically, the dynamic dispersion factor prediction model explicitly retains complex domain information, effectively expresses the complex frequency spectrum variation law of waves in the deep and shallow water transition area by learning the real part and the imaginary part of the frequency response in parallel, and realizes adaptive correction of the traditional shallow water dispersion factor. It is superior to the real value model in terms of energy spectrum main frequency identification accuracy and frequency spectrum stability, and is especially suitable for complex frequency data modeling scenarios obtained by flow measuring instruments such as ADCP.

[0071] Specifically, the input signal of the dynamic dispersion factor prediction model is derived from the velocity sequence collected by the underwater Doppler flow measuring instrument (ADCP), and after frequency domain transformation, it is in complex form, containing amplitude and phase information. The phase feature plays an important role in wave number inversion, dispersion relation correction and wave propagation direction identification. If the traditional real-valued deep learning model is used, the complex signal needs to be split into real and imaginary parts for independent modeling, which destroys the internal coupling relationship between amplitude and phase, resulting in a decrease in the estimation accuracy of the dispersion factor.

[0072] Specificly, when performing multi-head self-attention calculation, the attention weight is calculated using complex conjugate inner product, and complex value parameters are used in multi-head attention and feedforward network to realize synchronous modeling and learning of amplitude and phase information. Compared with the traditional transformer, the scheme directly performs linear and nonlinear operations in the complex domain, which can completely retain the physical correlation between amplitude and phase, and the complex value operation is consistent with the frequency domain representation of the wave signal, avoiding the information loss caused by real value encoding / decoding. The phase shift of the wave signal corresponds to the phase rotation in the complex plane, and the complex value network can naturally adapt to this change, thereby improving the generalization performance under different propagation directions, and effectively suppressing the interference of phase noise and improving the stability of the dispersion correction factor.

[0073] In some embodiments, the loss function of the dynamic dispersion factor prediction model is:

[0074] wherein, is the total loss of the dynamic dispersion factor prediction model, is the dynamic dispersion factor, is the physical label, MSE is the mean square error, is the complex frequency, g is the acceleration of gravity, is the wave number, is the weight decay system, , , is an adjustable weight coefficient.

[0075] Specifically, when calculating the physical label, if k cannot be directly obtained, the linear dispersion inverse solution (in the form of tanh(kh)) is used as the initial value of k.

[0076] Specifically, the dynamic dispersion factor prediction model uses AdamW as the optimizer, and the learning rate is set to 1e 4~1e 3, and the learning rate is attenuated by 3% per round.

[0077] In some specific embodiments, the formula for correcting the wave number of each frequency point using the dynamic dispersion factor is as follows:

[0078] wherein, is the corrected wave number, is the complex frequency, is the dynamic dispersion factor, and g is the acceleration of gravity.

[0079] Based on the corrected wave number and the vertical velocity power spectrum, the wave height spectrum is obtained, and the formula is as follows:

[0080] wherein, is the wave height spectrum, is the vertical velocity power spectrum, is the hyperbolic sine function, h is the depth, is the frequency.

[0081] The formula for calculating the significant wave height is:

[0082] wherein, is the significant wave height, is the spectral energy zero matrix, is the wave height spectrum, is the frequency, and d is the integral symbol.

[0083] The formula for the peak period is:

[0084] wherein, is the peak frequency, is the peak period.

[0085] Specifically, all complex forms mentioned in the present scheme are complex numbers, that is, the parameters are represented as complex numbers z=a+ib.

[0086] Embodiment Two Based on the same idea, referring to Figure 2 The present application also provides a wave spectrum estimation device based on a dynamic dispersion factor, comprising: The acquisition module is configured to acquire an ADCP velocity sequence in continuous time, acquire a vertical velocity power spectrum based on the ADCP velocity sequence, acquire complex-valued frequencies of different frequency points based on the vertical velocity power spectrum, acquire wave numbers and depths corresponding to each frequency point, and perform complex-valued processing on the wave numbers and depths. The prediction module is configured to input multiple groups of input vectors into a pre-trained dynamic dispersion factor prediction model in a batch input manner, with the complex-valued frequencies, the wave numbers, and the depths corresponding to the same frequency point as an input vector in each group, and with the dynamic dispersion factor prediction model learning the correlation among the complex-valued frequencies, the wave numbers, and the depths in each group of input vectors based on a multi-head self-attention mechanism to obtain a dynamic dispersion factor. The estimation module is configured to correct the wave numbers of each frequency point to obtain corrected wave numbers using the dynamic dispersion factor, and acquire a wave spectrum of the ADCP velocity sequence based on the corrected wave numbers, with the wave spectrum including a wave height spectrum, an effective wave height, and a main frequency period.

[0087] Embodiment Three This embodiment also provides an electronic device, which refers to Figure 3 The electronic device includes a memory 404 and a processor 402, the memory 404 stores a computer program, and the processor 402 is configured to execute the computer program to perform the steps in any of the above method embodiments.

[0088] Specifically, the processor 402 can include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0089] The memory 404 can include a mass storage that stores data or instructions. For example, and without limitation, the memory 404 can include a Hard Disk Drive (HDD), a floppy disk drive, a Solid State Drive (SSD), a flash drive, a Compact Disc Read Only Memory (CD-ROM), a magneto-optical disk, a magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. The memory 404 can be removable and / or non-removable (or fixed) as appropriate. The memory 404 can be internal or external as appropriate. In particular embodiments, the memory 404 is a Non-Volatile memory. In particular embodiments, the memory 404 includes a Read-Only Memory (ROM) and a Random Access Memory (RAM). The ROM can be a mask-programmed ROM, a Programmable ROM (PROM), an Erasable PROM (EPROM), an Electrically EPROM (EEPROM), an Electrically Alterable ROM (EAROM), or a FLASH memory, or a combination of two or more of these, as appropriate. The RAM can be a Static Random-Access Memory (SRAM) or a Dynamic Random Access Memory (DRAM), which can be a Fast Page Mode Dynamic Random Access Memory (FPMDRAM), an Extended Data Output Dynamic Random Access Memory (EDODRAM), a Synchronous Dynamic Random-Access Memory (SDRAM), or the like, as appropriate.

[0090] The memory 404 can be used to store or cache various data files needed for processing and / or communication, and possible computer program instructions executed by the processor 402.

[0091] The processor 402 implements any one of the wave spectrum estimation methods based on the dynamic dispersion factor in the above embodiments by reading and executing the computer program instructions stored in the memory 404.

[0092] Optionally, the electronic device described above can further include a transmission device 406 connected to the processor 402 and an input / output device 408 connected to the processor 402.

[0093] The transmission device 406 can be used to receive or send data via a network. Specific examples of the network can include wired or wireless networks provided by a communication provider of the electronic device. In one example, the transmission device includes a network adapter (NIC) that can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 406 can be a radio frequency (RF) module for communicating with the Internet in a wireless manner.

[0094] The input / output device 408 is used to input or output information. In the present embodiment, the input information can be a continuous-time ADCP velocity sequence, etc., and the output information can be an equal-height spectrum, an effective wave height, and a main frequency period, etc.

[0095] Optionally, in the present embodiment, the processor 402 can be configured to perform the following steps by computer program: obtain a continuous-time ADCP velocity sequence, obtain a vertical velocity power spectrum based on the ADCP velocity sequence, obtain complex-valued frequencies of different frequency points based on the vertical velocity power spectrum, and further obtain wave numbers and depths corresponding to each frequency point, and perform complex-valued processing on the wave numbers and depths; use the complex-valued frequencies, wave numbers, and depths corresponding to the same frequency point as a group of input vectors, input multiple groups of input vectors in a batch input manner into a pre-trained dynamic dispersion factor prediction model, and the dynamic dispersion factor prediction model learns the correlation between the complex frequencies, wave numbers, and depths in each group of input vectors based on a multi-head self-attention mechanism to obtain a dynamic dispersion factor; use the dynamic dispersion factor to correct the wave numbers of each frequency point to obtain corrected wave numbers, and obtain a wave spectrum of the ADCP velocity sequence based on the corrected wave numbers, the wave spectrum including a wave height spectrum, an effective wave height, and a main frequency period.

[0096] It should be noted that the specific examples in the present embodiment can refer to the examples described in the above embodiments and optional implementation manners, and the present embodiment will not be described here again.

[0097] Generally, the various embodiments can be implemented in hardware or special-purpose circuits, software, logic or any combination thereof. Some aspects of the application can be implemented in hardware, while other aspects can be implemented in firmware or software which can be executed by a controller, microprocessor or other computing device, but the application is not limited thereto. While various aspects of the application can be illustrated and described as block diagrams, flow charts, or using some other pictorial representation, it is well understood that these blocks, apparatus, systems, techniques or methods described herein can be implemented in, as non-limiting examples, hardware, software, firmware, special-purpose circuits or logic, general purpose hardware or controler or other computing devices, or some combination thereof.

[0098] Embodiments of the application can be implemented by computer software executable by a data processor of the mobile device such as in the processor entity, or by hardware, or by a combination of software and hardware. Computer software or program, also called program product, including software routines, applets and / or macros, can be stored in any apparatus-readable data storage medium and they include program instructions to implement specific tasks. The program product can include one or more computer-executable components such as the procedures described herein. The one or more computer-executable components can be one or more Figure 3 Any block in the logical flow of the method described herein can represent a module, segment, or portion of code which comprises one or more executable instructions for implementing the specified logical function(s). In some examples, the method or process can be embodied in machine-executable code, which can be used to program computational machines such as computers and the like. In other examples, the method or process can be embodied in software instructions, code or the like, which can be executed by computational machines. When the software instructions are executed, they create a machine that implements the essential function of the embodiments described herein. The software instructions can be stored in any suitable storage media that is accessible by the computational machines.

[0099] It should be understood by those skilled in the art that any combination of the technical features of the above embodiments can be made, and in order to make the description concise, not all possible combinations of the technical features in the above embodiments are described, however, as long as the combination of the technical features does not exist, it should be considered as the scope of the description.

[0100] The above embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but it should not be understood as a limitation on the scope of the present application. It should be pointed out that for those skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are all within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A wave spectrum estimation method based on dynamic dispersion factor, characterized in that, Includes the following steps: A continuous-time ADCP velocity sequence is obtained, a vertical velocity power spectrum is obtained based on the ADCP velocity sequence, a complex-valued frequency at different frequency points is obtained based on the vertical velocity power spectrum, and then the wave number and depth corresponding to each frequency point are obtained, and the wave number and depth are processed into complex values. Using the complex-valued frequency, wavenumber, and depth corresponding to the same frequency point as a set of input vectors, multiple sets of input vectors are input into a pre-trained dynamic dispersion factor prediction model in a batch input manner. The dynamic dispersion factor prediction model learns the correlation between the complex frequency, wavenumber, and depth in each set of input vectors based on a multi-head self-attention mechanism to obtain the dynamic dispersion factor. The wavenumber at each frequency point is corrected using the dynamic dispersion factor to obtain the corrected wavenumber. The wave spectrum of the ADCP velocity sequence is obtained based on the corrected wavenumber. The wave spectrum includes the wave height spectrum, the effective wave height, and the dominant frequency period.

2. The wave spectrum estimation method based on dynamic dispersion factor according to claim 1, characterized in that, The ADCP velocity sequence is preprocessed, and then the vertical velocity power spectrum is obtained based on the preprocessed ADCP velocity sequence. The preprocessing includes: mean removal, bandpass filtering, and window function weighting.

3. The wave spectrum estimation method based on dynamic dispersion factor according to claim 1, characterized in that, The ADCP velocity sequence is segmented, with each segment corresponding to a frequency point. The vertical velocity power spectrum at each frequency point is calculated using the following formula: ； in, Let L represent the vertical velocity power spectrum at the k-th frequency point, where L is the total number of segments, l is the segment index, and FFT represents the Fourier transform. Let l be the ADCP velocity sequence of the l-th segment, and U be the window energy correction factor.

4. The wave spectrum estimation method based on dynamic dispersion factor according to claim 1, characterized in that, The depth corresponding to each frequency point is directly obtained from the ADCP velocity sequence, and the wave number of each frequency point is obtained based on the dispersion relationship between the complex-valued frequency and the wave number of different frequency points.

5. The wave spectrum estimation method based on dynamic dispersion factor according to claim 1, characterized in that, A complex-valued Transformer model is used as the dynamic dispersion factor prediction model, and all trainable parameters in the dynamic dispersion factor prediction model are in complex-valued form.

6. The wave spectrum estimation method based on dynamic dispersion factor according to claim 1, characterized in that, In each attention head of the multi-head self-attention mechanism, complex-valued weights are constructed to perform a linear transformation on the input vector to obtain the complex-valued query vector, key vector, and value vector. Then, attention is calculated based on the query vector, key vector, and value vector to obtain the attention result, expressed by the formula: ； in, Let i be the relevance between the i-th query vector and the j-th key vector. For the i-th query vector, Let be the conjugate transpose of the j-th key vector. For the dimension of attention head, Used to extract the real part of a complex number. To The normalization result, For attention results, This is the j-th value.

7. The wave spectrum estimation method based on dynamic dispersion factor according to claim 1, characterized in that, The formula for correcting the wavenumber at each frequency point using the dynamic dispersion factor is expressed as follows: ； in, To correct the wavenumber, For complex-valued frequencies, Where g is the dynamic dispersion factor and g is the gravitational acceleration; The wave height spectrum is obtained based on the corrected wave number and the vertical velocity power spectrum, and the formula is expressed as follows: ； in, For the wave high spectrum, The vertical velocity power spectrum, Let h be a hyperbolic sine function, and h be the depth. For frequency; The formula for calculating the significant wave height is expressed as follows: ； in, Let be the effective wave height, and be the spectral energy zero matrix. For the wave high spectrum, d represents the frequency, and d represents the integral sign. The formula for the peak period is expressed as: ； in, Peak frequency, This is the peak period.

8. A wave spectrum estimation device based on dynamic dispersion factor, characterized in that, include: The acquisition module is used to acquire a continuous-time ADCP velocity sequence, acquire a vertical velocity power spectrum based on the ADCP velocity sequence, acquire complex-valued frequencies at different frequency points based on the vertical velocity power spectrum, acquire the wave number and depth corresponding to each frequency point, and perform complex-valued processing on the wave number and depth. The prediction module uses the complex-valued frequency, wavenumber, and depth corresponding to the same frequency point as a set of input vectors, and inputs multiple sets of input vectors into the pre-trained dynamic dispersion factor prediction model in a batch input manner. The dynamic dispersion factor prediction model learns the correlation between the complex frequency, wavenumber, and depth in each set of input vectors based on a multi-head self-attention mechanism to obtain the dynamic dispersion factor. The estimation module uses the dynamic dispersion factor to correct the wavenumber at each frequency point to obtain the corrected wavenumber, and obtains the wave spectrum of the ADCP velocity sequence based on the corrected wavenumber. The wave spectrum includes the wave height spectrum, the effective wave height, and the dominant frequency period.

9. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to run the computer program to perform a wave spectrum estimation method based on a dynamic dispersion factor as described in any one of claims 1-7.

10. A readable storage medium, characterized in that, The readable storage medium stores a computer program, which includes program code for controlling a process to execute the process, the process including a wave spectrum estimation method based on a dynamic dispersion factor according to any one of claims 1-7.

Images