Method and system for evaluating the precision of a ratio of multi-set two-dimensional directional frequency energy density spectrum

Abstract

The application provides a kind of multi-group two-dimensional direction frequency energy density spectrum ratio measurement precision evaluation method and system, belong to sea wave monitoring technical field, including respectively extracting the multiple spectral characteristic elements of to-be-measured spectrum sequence and standard spectrum sequence, determine the first similarity result between each spectral characteristic element;With to-be-measured spectrum sequence and standard spectrum sequence, iteratively segmented into multiple subsequences;Determine the second similarity result of each spectral characteristic element between corresponding subsequence;The first similarity result and the second similarity result are fused, and the initial spectral similarity matrix is constructed;Extract the spectral similarity stable component matrix and the spectral similarity change component matrix in initial spectral similarity matrix, superimpose the spectral similarity stable component matrix and the spectral similarity change component matrix, reconstruct to obtain final spectral similarity matrix;According to final spectral similarity matrix, determine the spectral similarity evaluation value between to-be-measured spectrum sequence and standard spectrum sequence, effectively improve the stability and reliability of wave spectrum data contrast analysis.

Description

Technical Field

[0001] This invention relates to the field of wave monitoring technology, and in particular to a method and system for evaluating the comparative accuracy of multiple sets of two-dimensional directional frequency energy density spectra. Background Technology

[0002] In the field of ocean observation, the two-dimensional directional frequency energy density spectrum of ocean waves is a core physical quantity that characterizes the distribution of ocean wave energy in a two-dimensional space of frequency and direction. It can scientifically describe the distribution of the constituent wave energy inside the ocean wave at different frequencies and in different directions, accurately characterize the energy magnitude, spectral peak characteristics, propagation direction and directional diffusion degree of the ocean wave, and is the core basic data for satellite ocean wave remote sensing, ocean numerical forecasting, nearshore engineering design and other fields.

[0003] Currently, comparative analysis of wave spectrum data is usually limited to simple correlation calculations of a few macroscopic parameters such as significant wave height and wave period, which makes the evaluation results susceptible to interference from random factors in the marine environment and lacks stability and reliability. Summary of the Invention

[0004] This invention provides a method and system for evaluating the comparative accuracy of multiple sets of two-dimensional directional frequency energy density spectra, in order to solve the shortcomings of poor stability and reliability in the comparative analysis of wave spectrum data in the prior art.

[0005] This invention provides a method for evaluating the comparative accuracy of multiple sets of two-dimensional directional frequency energy density spectra, comprising: Multiple spectral feature elements are extracted from the spectrum sequence to be tested and the standard spectrum sequence, and the similarity between each spectral feature element is determined to obtain the first similarity result; The test spectrum sequence and the standard spectrum sequence are iteratively divided into multiple sub-sequences; The spectral feature elements of each subsequence are determined, and the similarity of the spectral feature elements between the corresponding subsequences is determined to obtain the second similarity result; The first similarity result and the second similarity result are fused to construct an initial spectral similarity matrix; Extract the stable component matrix and the variable component matrix of spectral similarity from the initial spectral similarity matrix, and superimpose the stable component matrix of spectral similarity and the variable component matrix of spectral similarity to reconstruct the final spectral similarity matrix; Based on the final spectral similarity matrix, the spectral similarity evaluation value between the test spectral sequence and the standard spectral sequence is determined.

[0006] According to the present invention, a method for evaluating the accuracy of comparison of multiple sets of two-dimensional directional frequency energy density spectra is provided. The method involves extracting multiple spectral feature elements from the test spectral sequence and the standard spectral sequence, determining the similarity between each spectral feature element, and obtaining a first similarity result. This includes: Multiple spectral feature elements are extracted from both the spectrum sequence to be tested and the standard spectrum sequence. The spectral feature elements include at least one of the following: wave energy magnitude, frequency distribution, direction distribution, and spectral shape detail features. The similarity between each spectral feature element of the spectrum sequence to be tested and the corresponding spectral feature element in the standard spectrum sequence is calculated, and the calculation result is used as the first similarity result.

[0007] According to the present invention, a method for evaluating the comparative accuracy of multiple sets of two-dimensional directional frequency energy density spectra is provided. The method involves determining the spectral feature elements of each subsequence, determining the similarity of the spectral feature elements between corresponding subsequences, and obtaining a second similarity result, including: The spectrum sequence to be tested and the standard spectrum sequence are respectively divided into a preset number of subsequences according to the observation time order. Each subsequence includes a preset number of two-dimensional directional frequency energy density spectra, and there is a preset number of two-dimensional directional frequency energy density spectra overlapping between adjacent subsequences. Extract the multiple spectral feature elements of each subsequence, determine the similarity of each spectral feature element between each subsequence of the spectral sequence to be tested and the corresponding subsequence in the standard spectral sequence, and use the calculation results of each group as the second similarity result.

[0008] According to the present invention, a method for evaluating the accuracy of comparison of multiple sets of two-dimensional directional frequency energy density spectra is provided, wherein fusing the first similarity result with the second similarity result to construct an initial spectral similarity matrix includes: The first similarity result is used as the first row of data in the matrix, and the calculation results of each group in the second similarity result are used as the remaining rows of data in the matrix. The initial spectral similarity matrix is ​​constructed by arranging them in row order. The number of rows in the initial spectral similarity matrix corresponds to the total number of rows in the overall sequence and each subsequence, and the number of columns corresponds to the number of the multiple spectral feature elements.

[0009] According to the present invention, a method for evaluating the accuracy of comparison of multiple sets of two-dimensional directional frequency energy density spectra is provided, wherein the step of extracting the spectral similarity stable component matrix and the spectral similarity change component matrix from the initial spectral similarity matrix includes: The initial spectral similarity matrix is ​​decomposed in the frequency domain, and the low-frequency components are extracted from the decomposition results as the spectral similarity stable component matrix. The spectral similarity stable component matrix is ​​removed from the initial spectral similarity matrix to obtain the spectral similarity change component matrix.

[0010] According to the present invention, a method for evaluating the comparative accuracy of multiple sets of two-dimensional directional frequency energy density spectra is provided, wherein removing the spectral similarity stable component matrix from the initial spectral similarity matrix to obtain the spectral similarity variation component matrix includes: The initial spectral similarity matrix is ​​subtracted from the spectral similarity stable component matrix by corresponding elements, and the resulting matrix of the same dimension is used as the spectral similarity change component matrix.

[0011] According to the present invention, a method for evaluating the comparative accuracy of multiple sets of two-dimensional directional frequency energy density spectra, before superimposing the stable component matrix of spectral similarity with the variable component matrix of spectral similarity, further includes: The spectral similarity change component matrix is ​​filtered to remove random similarity noise.

[0012] According to the present invention, a method for evaluating the accuracy of comparison of multiple sets of two-dimensional directional frequency energy density spectra, wherein the step of superimposing the stable component matrix of the spectral similarity with the variable component matrix of the spectral similarity to reconstruct the final spectral similarity matrix includes: The stable component matrix of spectral similarity is added to the filtered component matrix of spectral similarity according to their corresponding elements, and the resulting matrix of the same dimension is used as the final spectral similarity matrix.

[0013] According to the present invention, a method for evaluating the accuracy of comparison of multiple sets of two-dimensional directional frequency energy density spectra is provided. The step of determining the spectral similarity evaluation value between the test spectral sequence and the standard spectral sequence based on the final spectral similarity matrix includes: A comprehensive statistical calculation is performed on all elements in the final spectral similarity matrix, and the statistical calculation result is used as the spectral similarity evaluation value between the test spectral sequence and the standard spectral sequence.

[0014] This invention provides a system for evaluating the accuracy of comparison of multiple sets of two-dimensional directional frequency energy density spectra, comprising: The first determining module is used to extract multiple spectral feature elements from the spectrum sequence to be tested and the standard spectrum sequence respectively, determine the similarity between each spectral feature element, and obtain the first similarity result; The second determining module is used to iteratively segment the spectrum sequence to be tested and the standard spectrum sequence into multiple sub-sequences; determine the spectral feature elements of each sub-sequence respectively; determine the similarity of the spectral feature elements between the corresponding sub-sequences; and obtain a second similarity result. The matrix construction module is used to fuse the first similarity result and the second similarity result to construct an initial spectral similarity matrix; extract the spectral similarity stable component matrix and the spectral similarity change component matrix from the initial spectral similarity matrix, and superimpose the spectral similarity stable component matrix and the spectral similarity change component matrix to reconstruct the final spectral similarity matrix; The third determining module is used to determine the spectral similarity evaluation value between the test spectral sequence and the standard spectral sequence based on the final spectral similarity matrix.

[0015] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the method for evaluating the comparative accuracy of multiple sets of two-dimensional directional frequency energy density spectra as described above.

[0016] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for evaluating the comparative accuracy of multiple sets of two-dimensional directional frequency energy density spectra as described above.

[0017] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements the method for evaluating the comparative accuracy of multiple sets of two-dimensional directional frequency energy density spectra as described above.

[0018] The present invention provides a method and system for evaluating the accuracy of comparison of multiple sets of two-dimensional directional frequency energy density spectra. This method extracts multiple spectral feature elements from the test spectral sequence and the standard spectral sequence to determine the first similarity result. It then iteratively segments the sequence into multiple sub-sequences and determines the similarity of each spectral feature element between the corresponding sub-sequences to obtain a second similarity result. These results are then fused to construct an initial spectral similarity matrix. Finally, the method extracts the stable and variable component matrices of spectral similarity from the initial matrix and reconstructs them to obtain the final spectral similarity matrix. This effectively separates the stable component reflecting the inherent bias of the system from the variable component influenced by random environmental factors, significantly suppressing the interference of random marine environmental factors on the evaluation results. Ultimately, the spectral similarity evaluation value determined based on the final spectral similarity matrix can objectively, stably, and reliably reflect the consistency and similarity between the two spectral sequences. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the method for evaluating the accuracy of comparing multiple sets of two-dimensional directional frequency energy density spectra provided by this invention. Figure 2 This is a schematic diagram of the structure of the multi-set two-dimensional directional frequency energy density spectrum comparison accuracy evaluation system provided by the present invention; Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0022] Figure 1 This is a flowchart illustrating the method for evaluating the accuracy of comparing multiple sets of two-dimensional directional frequency energy density spectra provided by this invention.

[0023] like Figure 1 As shown in the figure, this embodiment provides a method for evaluating the comparative accuracy of multiple sets of two-dimensional directional frequency energy density spectra, including: 101. Extract multiple spectral feature elements from the spectrum sequence to be tested and the standard spectrum sequence respectively, determine the similarity between each spectral feature element, and obtain the first similarity result.

[0024] Specifically, two synchronously observed two-dimensional directional frequency energy density spectrum sequences are selected as comparison and evaluation sequences to assess their similarity. Each two-dimensional directional frequency energy density spectrum sequence contains... Two-dimensional directional frequency energy density spectra observed simultaneously. The spectrum sequence to be tested, i.e., the two-dimensional directional frequency energy density spectrum sequence observed by the instrument under test, is denoted as... Data was collected synchronously at fixed time intervals by wave observation equipment whose performance was to be evaluated, including... A two-dimensional directional frequency energy density spectrum arranged in chronological order of observation. The standard spectral sequence, i.e., the two-dimensional directional frequency energy density spectrum sequence observed by a standard instrument, is denoted as... The data, collected synchronously by a standard wave observation device with known measurement accuracy and calibration, and the device under test, serves as the benchmark data for similarity comparison. It contains data with the same number of spectral sequences as the device under test and a one-to-one time correspondence. Two-dimensional directional frequency energy density spectrum.

[0025] Multiple spectral feature elements are extracted from both the test spectral sequence and the standard spectral sequence. The similarity between corresponding spectral feature elements of each class in the two sequences is calculated to obtain the first similarity result. This covers multiple dimensions of wave physical characteristics at the overall time series level, fully reflecting the consistency of the two spectral sequences at the overall statistical level. The spectral feature elements represent quantized parameters extracted from a single or a set of two-dimensional directional frequency-energy density spectra, corresponding to different physical properties of ocean waves such as energy magnitude, frequency distribution, directional distribution, and spectral details. These elements serve as specific comparison dimensions for measuring the similarity between the two sets of spectral data.

[0026] 102. Iteratively divide the spectrum sequence to be tested and the standard spectrum sequence into multiple subsequences.

[0027] Specifically, the spectral sequence to be tested and standard spectral sequence Iteratively segmented into M (where) according to the chronological order of observation time. There are N two-dimensional directional frequency energy density spectrum subsequences, each subsequence containing N consecutive (where N is the number of subsequences). Two-dimensional frequency-energy density spectra are used, with d overlapping spectral data points between adjacent subsequences, to analyze similarity variation patterns from a local timescale. .

[0028] Two-dimensional directional frequency energy density spectrum sequence Divided into The two-dimensional frequency energy density spectral sequence is as follows: ,in =1,2,3… .

[0029] Two-dimensional directional frequency energy density spectrum sequence Divided into The two-dimensional frequency energy density spectral sequence is as follows: ,in =1,2,3… .

[0030] 103. Determine the spectral feature elements of each subsequence, determine the similarity of the spectral feature elements between the corresponding subsequences, and obtain the second similarity result.

[0031] Specifically, the two-dimensional directional frequency energy density spectral sequence is solved separately. and Spectral feature elements, where =1,2,3… The similarity of various spectral feature elements between each subsequence of the test spectral sequence and the corresponding subsequence in the standard spectral sequence is calculated to obtain the second similarity result. Multi-dimensional similarity data is supplemented at the local time segment level to capture the distribution and changes of similarity under different time windows.

[0032] 104. Combine the first similarity result with the second similarity result to construct the initial spectral similarity matrix.

[0033] Specifically, the first similarity result and the second similarity result are fused to construct an initial spectral similarity matrix SM, which organizes the scattered similarity data of multiple scales and dimensions into a standardized two-dimensional data structure, making it easier to process the components separately.

[0034] 105. Extract the stable component matrix and the variable component matrix of spectral similarity from the initial spectral similarity matrix, superimpose the stable component matrix and the variable component matrix of spectral similarity, and reconstruct the final spectral similarity matrix.

[0035] Specifically, the stable and variable component matrices of spectral similarity are extracted from the initial spectral similarity matrix SM. The variable component matrix is ​​then filtered to remove random similarity noise. The stable and variable component matrices are then superimposed at corresponding positions to reconstruct the final spectral similarity matrix. This component separation method separates the inherent, persistent systematic consistency deviation between the two spectral sequences from the fluctuation deviation caused by random environmental factors, reducing the interference of irregular random fluctuations on the overall evaluation results.

[0036] 106. Based on the final spectral similarity matrix, determine the spectral similarity evaluation value between the test spectral sequence and the standard spectral sequence.

[0037] Specifically, the spectral sequence to be tested is calculated based on the final spectral similarity matrix. With standard spectral sequence Spectral similarity evaluation value between By combining multi-dimensional features and comparison results across multiple time scales, the overall similarity between the two wave spectrum sequences can be objectively and stably reflected.

[0038] Furthermore, based on the above embodiments, this embodiment extracts multiple spectral feature elements from the test spectral sequence and the standard spectral sequence respectively, determines the similarity between each spectral feature element, and obtains a first similarity result, including: extracting multiple spectral feature elements from the test spectral sequence and the standard spectral sequence respectively, wherein the spectral feature elements include at least one of the features used to characterize wave energy magnitude, frequency distribution, direction distribution, and spectral shape details; calculating the similarity between each spectral feature element of the test spectral sequence and the corresponding spectral feature element in the standard spectral sequence respectively, and using the calculation result as the first similarity result.

[0039] Specifically, extract the spectral sequence to be tested. and standard spectral sequence Each wave has multiple spectral feature elements, which cover four categories of attributes: wave energy magnitude, frequency distribution characteristics, directional distribution characteristics, and spectral shape details. Specifically, these include the significant wave height sequence, average period sequence, main wave direction sequence, total wave energy sequence, spectral peak direction sequence, spectral peak energy value sequence of the one-dimensional directional energy density spectrum, spectral width sequence of the one-dimensional directional energy density spectrum, wave direction spectral centroid sequence, 0.5 times spectral peak wave direction total energy sequence, 0.5 times spectral peak energy wave direction bandwidth sequence, and the top five coefficients. The top five coefficients The top five coefficients The top five coefficients The top five coefficients Wave-direction frequency distribution, wave-direction extension distribution, wave-direction frequency distribution kurtosis, and one-dimensional directional energy density spectrum.

[0040] Calculate the spectral sequence to be tested respectively Each spectral feature element and the standard spectral sequence The similarity between corresponding spectral feature elements is calculated, and all results are integrated as the first similarity result, totaling 41 similarity indicators. The symbols for each indicator are defined as follows: The one-dimensional correlation coefficient of the effective wave height sequence is The one-dimensional cosine coefficient of the included angle of the effective wave height sequence is: ; The one-dimensional correlation coefficient of the average periodic series is The one-dimensional cosine coefficient of the included angle of the average periodic sequence is ; The one-dimensional correlation coefficient of the main wave sequence is The one-dimensional cosine coefficient of the included angle of the main wave sequence is ; The one-dimensional correlation coefficient of the total wave energy sequence is The one-dimensional cosine coefficient of the included angle in the total wave energy sequence is: ; The one-dimensional correlation coefficient of the spectral peak wave sequence is The one-dimensional cosine coefficient of the included angle of the spectral peak wave direction sequence is ; The one-dimensional correlation coefficient of the peak energy value sequence of the one-dimensional directional energy density spectrum is: The one-dimensional cosine coefficient of the peak energy value sequence of the one-dimensional directional energy density spectrum is: ; The one-dimensional correlation coefficient of the spectral width sequence of the one-dimensional directional energy density spectrum is: The one-dimensional cosine coefficient of the included angle of the spectral width sequence of the one-dimensional directional energy density spectrum is: ; The one-dimensional correlation coefficient of the centroid sequence of the wave spectrum is The one-dimensional cosine coefficient of the included angle of the centroid sequence of the wave spectrum is ; The one-dimensional correlation coefficient of the total energy sequence in the 0.5-fold spectral peak direction is: The one-dimensional cosine coefficient of the included angle of the 0.5 octave peak wave-direction total energy sequence is: ; The one-dimensional correlation coefficient of the 0.5-fold peak energy wave-bandwidth sequence is: The one-dimensional cosine coefficient of the included angle of the 0.5 times peak energy wave-bandwidth sequence is: ; Top five coefficients of The average of the one-dimensional correlation coefficients is The top five coefficients of The average value of the one-dimensional cosine coefficients of the included angle is ; Top five coefficients of The average of the one-dimensional correlation coefficients is The top five coefficients of The average value of the one-dimensional cosine coefficients of the included angle is ; Top five coefficients of The average of the one-dimensional correlation coefficients is The top five coefficients of The average value of the one-dimensional cosine coefficients of the included angle is ; Top five coefficients of The average of the one-dimensional correlation coefficients is The top five coefficients of The average value of the one-dimensional cosine coefficients of the included angle is ; Top five coefficients of The average of the one-dimensional correlation coefficients is The top five coefficients of The average value of the one-dimensional cosine coefficients of the included angle is ; Wave frequency distribution The average of the one-dimensional correlation coefficients is Wave direction frequency distribution The average value of the one-dimensional cosine coefficients of the included angle is ; Wave-direction spread distribution The average of the one-dimensional correlation coefficients is , wave direction expansion distribution The average value of the one-dimensional cosine coefficients of the included angle is ; Kurtosis of wave frequency distribution The average of the one-dimensional correlation coefficients is kurtosis of wave frequency distribution The average value of the one-dimensional cosine coefficients of the included angle is ; One-dimensional directional energy density spectrum The average of the one-dimensional correlation coefficients is One-dimensional directional energy density spectrum The average value of the one-dimensional cosine coefficients of the included angle is ; Two-dimensional directional frequency energy density spectrum The average of the two-dimensional correlation coefficients is Two-dimensional directional frequency energy density spectrum The average of the correlation coefficients of the two-dimensional normalized products is ; The average of the minimum-to-maximum ratio of the total energy of each wave is: .

[0041] For time-series parameters such as significant wave height and mean period, where each observation moment corresponds to a single value, the one-dimensional correlation coefficient and one-dimensional cosine coefficient of the included angle are directly calculated for the entire sequence. For parameters such as the first five coefficients and wave direction frequency distribution, where each single spectrum corresponds to a set of values, the similarity of the parameters corresponding to each single spectrum at each observation moment is first calculated, and then the average of the similarity results for all L observation moments is taken. For the complete two-dimensional directional frequency energy density spectrum, the two-dimensional correlation coefficient and two-dimensional normalized product correlation coefficient of each single spectrum are calculated and then averaged. At the same time, the minimum-to-maximum ratio of the total wave energy of each single spectrum is calculated and then averaged. This multi-dimensional, multi-index similarity calculation comprehensively covers all physical characteristics of the wave spectrum, avoiding the one-sidedness caused by comparing single parameters and improving the comprehensiveness of the overall evaluation.

[0042] Taking 306 spectral sequences from simultaneous observations by two sets of wave buoys in a certain sea area as an example, i.e., L=306, the following calculations are obtained: =0.997894, =0.999166, =0.941904, =0.988059, =0.906970, =0.926420, =0.883589, and the similarity calculation process for other spectral feature elements is similar.

[0043] Furthermore, based on the above embodiments, this embodiment determines the spectral feature elements of each subsequence, determines the similarity of the spectral feature elements between corresponding subsequences, and obtains a second similarity result, including: dividing the spectrum sequence to be tested and the standard spectrum sequence into a preset number of subsequences according to the observation time order, each subsequence including a preset number of two-dimensional directional frequency energy density spectra, and adjacent subsequences having a preset number of two-dimensional directional frequency energy density spectra with overlap; extracting multiple spectral feature elements of each subsequence, determining the similarity of the spectral feature elements between each subsequence of the spectrum sequence to be tested and the corresponding subsequence in the standard spectrum sequence, and using each set of calculation results as the second similarity result.

[0044] Specifically, the spectral sequence to be tested and standard spectral sequence The data is divided into M subsequences according to the observation time sequence. Each subsequence contains N consecutive two-dimensional directional frequency energy density spectra. d overlapping two-dimensional directional frequency energy density spectra are maintained between any two adjacent subsequences, with the overlap number satisfying the following condition: .

[0045] Standard spectral sequence The resulting M two-dimensional directional frequency energy density spectral sequences are denoted as follows: ,in =1,2,3… Spectral sequence to be tested Divided into The two-dimensional frequency energy density spectrum sequence is as follows: ,in =1,2,3… The sliding overlap segmentation method ensures data continuity between adjacent subsequences, avoids data abrupt changes caused by segmentation boundaries, and provides sufficient data support for subsequent frequency domain component separation processing due to the sufficient number of subsequences and the length of a single subsequence.

[0046] Extract each subsequence separately and Based on all the aforementioned spectral feature elements, and following calculation rules completely consistent with the overall sequence, the similarity of each spectral feature element between corresponding subsequences is determined. The similarity index of the i-th item in the j-th subsequence is denoted as... Where i = 1, 2, 3, ..., 41. The calculation results of each group of subsequences are integrated in chronological order as the second similarity result. The multi-dimensional similarity calculation at the subsequence level reflects the degree of consistency between the two spectral sequences within different time windows, and fully presents the change law of similarity over time.

[0047] Taking a complete sequence of 306 spectra as an example, it can be divided into 13 subsequences, i.e., M=13. Each subsequence contains 170 two-dimensional directional frequency energy density spectra, i.e., N=170. There are 160 overlapping spectral data between two adjacent subsequences, i.e., d=160, which meets the above parameter value requirements.

[0048] The first subsequence was calculated as follows: =0.997629、 =0.999120; =0.916357、 =0.984425; =0.892319、 =0.916070; =0.870912, and the calculation method for the remaining subsequences is the same.

[0049] Based on the above embodiments, this embodiment fuses the first similarity result and the second similarity result to construct an initial spectral similarity matrix, including: using the first similarity result as the first row of data in the matrix, and using the calculation results of each group in the second similarity result as the remaining rows of data in the matrix, and arranging them in row order to construct the initial spectral similarity matrix; the number of rows in the initial spectral similarity matrix corresponds to the total number of the overall sequence and each subsequence, and the number of columns corresponds to the number of multiple spectral feature elements.

[0050] Specifically, the first similarity result is used as the first row of the matrix, and the calculation results of each subsequence in the second similarity result are used as the second row and subsequent rows of the matrix. These subsequences are arranged row by row according to their chronological order to construct the initial spectral similarity matrix SM. The total number of rows in the initial spectral similarity matrix SM corresponds to the total number of rows in the overall sequence and all subsequences, i.e., M+1 rows; the total number of columns corresponds to the number of all spectral feature similarity indices involved in the calculation, totaling 41 columns.

[0051] The element at each position in the matrix is ​​denoted as ,in =1,2,3…14; =1,2,3…41. When When =1, it corresponds to the first similarity result of the overall sequence. When p=2 to p=M+1, the similarity results for the subsequences of the j=p-1th group. The well-organized matrix structure unifies scattered multi-scale similarity data into standardized two-dimensional data, facilitating frequency domain processing and component separation.

[0052] Furthermore, based on the above embodiments, this embodiment extracts the spectral similarity stable component matrix and the spectral similarity change component matrix from the initial spectral similarity matrix, including: performing frequency domain decomposition on the initial spectral similarity matrix, extracting low-frequency components from the decomposition results as the spectral similarity stable component matrix; and removing the spectral similarity stable component matrix from the initial spectral similarity matrix to obtain the spectral similarity change component matrix.

[0053] Specifically, a two-dimensional Fourier transform is performed on the initial spectral similarity matrix SM to transform the matrix data from the original row and column space to the frequency domain, completing frequency domain decomposition. Low-frequency components are extracted from the frequency domain decomposition results, and then an inverse two-dimensional Fourier transform is used to convert the low-frequency components back to a spatial domain matrix with the same dimensions as the original matrix, yielding the stable spectral similarity component matrix. The low-frequency components correspond to the gently changing and persistent similarity characteristics in the matrix, reflecting the inherent and systematic consistency between the two spectral sequences, unaffected by instantaneous random factors. Removing the stable spectral similarity component matrix from the initial spectral similarity matrix SM yields the variable spectral similarity component matrix. The variable component matrix corresponds to the similarity characteristics that fluctuate over time, reflecting the similarity changes caused by dynamic factors such as random fluctuations in the marine environment and observational noise.

[0054] Furthermore, based on the above embodiments, this embodiment removes the spectral similarity stable component matrix from the initial spectral similarity matrix to obtain the spectral similarity change component matrix, including: subtracting the initial spectral similarity matrix from the spectral similarity stable component matrix by corresponding elements, and using the matrix of the same dimension obtained after the subtraction as the spectral similarity change component matrix.

[0055] Specifically, the initial spectral similarity matrix SM and the stable spectral similarity component matrix are matched one-to-one according to their corresponding row and column numbers. The values ​​of the two elements at each corresponding position are subtracted one by one. All the results are then arranged in their original row and column positions to obtain a spectral similarity variation component matrix with the exact same dimensions as the initial spectral similarity matrix SM. This element-by-element subtraction method accurately isolates the stable components from the matrix while fully preserving all fluctuation information, ensuring the accuracy of component separation.

[0056] Furthermore, based on the above embodiments, this embodiment further includes, before superimposing the spectral similarity stable component matrix and the spectral similarity change component matrix, performing filtering processing on the spectral similarity change component matrix to filter out random similarity noise in the spectral similarity change component matrix.

[0057] Specifically, a Gaussian filter is used to perform two-dimensional smoothing filtering on the spectral similarity variation component matrix. The filtering process is based on a neighborhood-weighted average of matrix elements, with weights following a Gaussian distribution. This removes high-frequency random spikes while preserving the continuous variation trend within the neighborhood, filtering out random similarity noise caused by random observation noise and instantaneous environmental disturbances in the spectral similarity variation component matrix. The filtering process retains the true and regular similarity fluctuations in the variation components, removing irregular random interference, further improving the stability and reliability of the final evaluation results.

[0058] Furthermore, based on the above embodiments, this embodiment superimposes the spectral similarity stable component matrix and the spectral similarity change component matrix to reconstruct the final spectral similarity matrix, including: adding the spectral similarity stable component matrix and the filtered spectral similarity change component matrix element by element, and using the matrix of the same dimension obtained after addition as the final spectral similarity matrix.

[0059] Specifically, the stable component matrix of spectral similarity and the filtered component matrix of spectral similarity variation are matched one-to-one according to their corresponding row and column numbers. The values ​​of the two elements at corresponding positions are added one by one. All the results are then arranged in their original row and column positions to obtain the final spectral similarity matrix with the same dimensions as the initial spectral similarity matrix (SM). The reconstructed matrix retains both the inherent system consistency characteristics between the two spectral sequences and the similarity fluctuation characteristics with actual physical meaning, while eliminating the interference of random noise, thus balancing the accuracy and stability of the evaluation results.

[0060] Furthermore, based on the above embodiments, this embodiment determines the spectral similarity evaluation value between the test spectral sequence and the standard spectral sequence according to the final spectral similarity matrix, including: performing comprehensive statistical calculations on all elements in the final spectral similarity matrix, and using the statistical calculation results as the spectral similarity evaluation value between the test spectral sequence and the standard spectral sequence.

[0061] Specifically, a comprehensive statistical calculation is performed on all elements in the final spectral similarity matrix. This involves summing the values ​​of all elements in the matrix, dividing by the total number of elements, and obtaining the arithmetic mean. This mean is then used as the spectral sequence to be tested. With standard spectral sequence Spectral similarity evaluation value between The calculation method of averaging all matrix elements integrates similarity information across all dimensions and time scales, resulting in a unique and directly comparable comprehensive evaluation result, which facilitates intuitive judgment of the overall consistency between two spectral sequences.

[0062] Taking the comparative data of the above 306 spectra as an example, the final calculated spectral similarity evaluation value =0.916659.

[0063] The following describes the comparison accuracy evaluation system for multiple sets of two-dimensional directional frequency energy density spectra provided by the present invention. The comparison accuracy evaluation system for multiple sets of two-dimensional directional frequency energy density spectra described below can be referred to in correspondence with the comparison accuracy evaluation method for multiple sets of two-dimensional directional frequency energy density spectra described above.

[0064] Figure 2 This is a schematic diagram of the structure of a multi-set two-dimensional directional frequency energy density spectrum comparison accuracy evaluation system provided by the present invention.

[0065] like Figure 2 As shown in the figure, this embodiment provides a comparative accuracy evaluation system for multiple sets of two-dimensional directional frequency energy density spectra, including: The first determining module 201 is used to extract multiple spectral feature elements of the spectrum sequence to be tested and the standard spectrum sequence respectively, determine the similarity between each spectral feature element, and obtain the first similarity result; The second determining module 202 is used to iteratively divide the spectrum sequence to be tested and the standard spectrum sequence into multiple sub-sequences; determine the spectral feature elements of each sub-sequence; determine the similarity of the spectral feature elements between the corresponding sub-sequences; and obtain the second similarity result. The matrix construction module 203 is used to fuse the first similarity result and the second similarity result to construct an initial spectral similarity matrix; extract the spectral similarity stable component matrix and the spectral similarity change component matrix from the initial spectral similarity matrix, superimpose the spectral similarity stable component matrix and the spectral similarity change component matrix, and reconstruct the final spectral similarity matrix; The third determining module 204 is used to determine the spectral similarity evaluation value between the test spectral sequence and the standard spectral sequence based on the final spectral similarity matrix.

[0066] Figure 3 A schematic diagram of the physical structure of an electronic device is provided.

[0067] like Figure 3 As shown, the electronic device may include: a processor 310, a communications interface 320, a memory 330, and a communications bus 340, wherein the processor 310, the communications interface 320, and the memory 330 communicate with each other through the communications bus 340. The processor 310 can call logic instructions in the memory 330 to execute a method for evaluating the accuracy of multiple sets of two-dimensional directional frequency energy density spectra. This method includes: extracting multiple spectral feature elements from the test spectrum sequence and the standard spectrum sequence respectively, determining the similarity between each spectral feature element, and obtaining a first similarity result; iteratively dividing the test spectrum sequence and the standard spectrum sequence into multiple sub-sequences; determining the spectral feature elements of each sub-sequence respectively, determining the similarity between the spectral feature elements of the corresponding sub-sequences, and obtaining a second similarity result; fusing the first similarity result and the second similarity result to construct an initial spectral similarity matrix; extracting the stable spectral similarity component matrix and the variable spectral similarity component matrix from the initial spectral similarity matrix, superimposing the stable spectral similarity component matrix and the variable spectral similarity component matrix to reconstruct a final spectral similarity matrix; and determining the spectral similarity evaluation value between the test spectrum sequence and the standard spectrum sequence based on the final spectral similarity matrix.

[0068] Furthermore, the logical instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0069] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the method for evaluating the accuracy of comparison of multiple sets of two-dimensional directional frequency energy density spectra provided by the above methods. The method includes: extracting multiple spectral feature elements of the spectrum sequence to be tested and the standard spectrum sequence respectively, determining the similarity between each spectral feature element, and obtaining a first similarity result; iteratively dividing the spectrum sequence to be tested and the standard spectrum sequence into multiple sub-sequences; determining the spectral feature elements of each sub-sequence respectively, determining the similarity between each spectral feature element of the corresponding sub-sequences, and obtaining a second similarity result; fusing the first similarity result and the second similarity result to construct an initial spectral similarity matrix; extracting the spectral similarity stable component matrix and the spectral similarity change component matrix from the initial spectral similarity matrix, superimposing the spectral similarity stable component matrix and the spectral similarity change component matrix to reconstruct a final spectral similarity matrix; and determining the spectral similarity evaluation value between the spectrum sequence to be tested and the standard spectrum sequence based on the final spectral similarity matrix.

[0070] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements a method for evaluating the accuracy of comparison of multiple sets of two-dimensional directional frequency energy density spectra provided by the methods described above. This method includes: extracting multiple spectral feature elements from a test spectral sequence and a standard spectral sequence, determining the similarity between each spectral feature element, and obtaining a first similarity result; iteratively segmenting the test spectral sequence and the standard spectral sequence into multiple sub-sequences; determining the spectral feature elements of each sub-sequence, determining the similarity between the spectral feature elements of corresponding sub-sequences, and obtaining a second similarity result; fusing the first similarity result and the second similarity result to construct an initial spectral similarity matrix; extracting the spectral similarity stable component matrix and the spectral similarity change component matrix from the initial spectral similarity matrix, superimposing the spectral similarity stable component matrix and the spectral similarity change component matrix, and reconstructing a final spectral similarity matrix; and determining the spectral similarity evaluation value between the test spectral sequence and the standard spectral sequence based on the final spectral similarity matrix.

[0071] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0072] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0073] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for evaluating the comparative accuracy of multiple sets of two-dimensional directional frequency energy density spectra, characterized in that, include: Multiple spectral feature elements are extracted from the spectrum sequence to be tested and the standard spectrum sequence, and the similarity between each spectral feature element is determined to obtain the first similarity result; The test spectral sequence and the standard spectral sequence are iteratively divided into multiple sub-sequences; The spectral feature elements of each subsequence are determined, and the similarity of the spectral feature elements between the corresponding subsequences is determined to obtain the second similarity result; The first similarity result and the second similarity result are fused to construct an initial spectral similarity matrix; Extract the stable component matrix and the variable component matrix of spectral similarity from the initial spectral similarity matrix, and superimpose the stable component matrix of spectral similarity and the variable component matrix of spectral similarity to reconstruct the final spectral similarity matrix; Based on the final spectral similarity matrix, the spectral similarity evaluation value between the test spectral sequence and the standard spectral sequence is determined.

2. The method for evaluating the comparative accuracy of multiple sets of two-dimensional directional frequency energy density spectra according to claim 1, characterized in that, The process involves extracting multiple spectral feature elements from the test spectral sequence and the standard spectral sequence, determining the similarity between each spectral feature element, and obtaining a first similarity result, including: Multiple spectral feature elements are extracted from both the spectrum sequence to be tested and the standard spectrum sequence. The spectral feature elements include at least one of the following: wave energy magnitude, frequency distribution, direction distribution, and spectral shape detail features. The similarity between each spectral feature element of the spectrum sequence to be tested and the corresponding spectral feature element in the standard spectrum sequence is calculated, and the calculation result is used as the first similarity result.

3. The method for evaluating the comparative accuracy of multiple sets of two-dimensional directional frequency energy density spectra according to claim 1, characterized in that, The step of determining the spectral feature elements of each subsequence, determining the similarity of the spectral feature elements between corresponding subsequences, and obtaining a second similarity result includes: The spectrum sequence to be tested and the standard spectrum sequence are respectively divided into a preset number of subsequences according to the observation time order. Each subsequence includes a preset number of two-dimensional directional frequency energy density spectra, and there is a preset number of two-dimensional directional frequency energy density spectra overlapping between adjacent subsequences. Extract the multiple spectral feature elements of each subsequence, determine the similarity of each spectral feature element between each subsequence of the spectral sequence to be tested and the corresponding subsequence in the standard spectral sequence, and use the calculation results of each group as the second similarity result.

4. The method for evaluating the comparative accuracy of multiple sets of two-dimensional directional frequency energy density spectra according to claim 1, characterized in that, The step of fusing the first similarity result and the second similarity result to construct an initial spectral similarity matrix includes: The first similarity result is used as the first row of data in the matrix, and the calculation results of each group in the second similarity result are used as the remaining rows of data in the matrix. The initial spectral similarity matrix is ​​constructed by arranging them in row order. The number of rows in the initial spectral similarity matrix corresponds to the total number of rows in the overall sequence and each subsequence, and the number of columns corresponds to the number of the multiple spectral feature elements.

5. The method for evaluating the comparative accuracy of multiple sets of two-dimensional directional frequency energy density spectra according to claim 1, characterized in that, The extraction of the stable spectral similarity component matrix and the variable spectral similarity component matrix from the initial spectral similarity matrix includes: The initial spectral similarity matrix is ​​decomposed in the frequency domain, and the low-frequency components are extracted from the decomposition results as the spectral similarity stable component matrix. The spectral similarity stable component matrix is ​​removed from the initial spectral similarity matrix to obtain the spectral similarity change component matrix.

6. The method for evaluating the comparative accuracy of multiple sets of two-dimensional directional frequency energy density spectra according to claim 5, characterized in that, The step of removing the stable spectral similarity component matrix from the initial spectral similarity matrix to obtain the spectral similarity variation component matrix includes: The initial spectral similarity matrix is ​​subtracted from the spectral similarity stable component matrix by corresponding elements, and the resulting matrix of the same dimension is used as the spectral similarity change component matrix.

7. The method for evaluating the comparative accuracy of multiple sets of two-dimensional directional frequency energy density spectra according to claim 5, characterized in that, Before superimposing the stable component matrix of spectral similarity with the variable component matrix of spectral similarity, the method further includes: The spectral similarity change component matrix is ​​filtered to remove random similarity noise.

8. The method for evaluating the comparative accuracy of multiple sets of two-dimensional directional frequency energy density spectra according to claim 7, characterized in that, The step of superimposing the stable component matrix of spectral similarity with the variable component matrix of spectral similarity to reconstruct the final spectral similarity matrix includes: The stable component matrix of spectral similarity is added to the filtered component matrix of spectral similarity according to their corresponding elements, and the resulting matrix of the same dimension is used as the final spectral similarity matrix.

9. The method for evaluating the comparative accuracy of multiple sets of two-dimensional directional frequency energy density spectra according to any one of claims 1-8, characterized in that, The step of determining the spectral similarity evaluation value between the test spectral sequence and the standard spectral sequence based on the final spectral similarity matrix includes: A comprehensive statistical calculation is performed on all elements in the final spectral similarity matrix, and the statistical calculation result is used as the spectral similarity evaluation value between the test spectral sequence and the standard spectral sequence.

10. A comparative accuracy evaluation system for multiple sets of two-dimensional directional frequency energy density spectra, characterized in that, include: The first determining module is used to extract multiple spectral feature elements from the spectrum sequence to be tested and the standard spectrum sequence respectively, determine the similarity between each spectral feature element, and obtain the first similarity result; The second determining module is used to iteratively divide the spectrum sequence to be tested and the standard spectrum sequence into multiple sub-sequences; The spectral feature elements of each subsequence are determined, and the similarity of the spectral feature elements between the corresponding subsequences is determined to obtain the second similarity result; A matrix construction module is used to fuse the first similarity result and the second similarity result to construct an initial spectral similarity matrix; Extract the stable component matrix and the variable component matrix of spectral similarity from the initial spectral similarity matrix, and superimpose the stable component matrix of spectral similarity and the variable component matrix of spectral similarity to reconstruct the final spectral similarity matrix; The third determining module is used to determine the spectral similarity evaluation value between the test spectral sequence and the standard spectral sequence based on the final spectral similarity matrix.

Images