A method and system for spatial-frequency joint deblurring of inter-element phase difference of distortion towed line spectrum array

Abstract

The application discloses a distortion towed array line spectrum inter-element phase difference space-time frequency joint deblurring method and system, and the method comprises the following steps: firstly, based on the grid theory phase difference deblurring, a three-dimensional phase difference residual tensor composed of time frames, spatial elements and discrete time delay differences is constructed; secondly, a system optimization target is constructed, and time and space smoothing weight parameters are dynamically obtained through the global mean of the phase residual tensor; then, recursive solving of the three-dimensional residual tensor is realized based on the dynamic programming principle, and the cumulative minimum cost, the space-time optimal predecessor state and the path index matrix are calculated point by point; finally, the optimal discrete time delay difference of each time frame and each element is obtained by reverse recursion, and the high-frequency line spectrum phase difference deblurring is completed. The application fully utilizes the space-time two-dimensional correlation of the line spectrum inter-element phase difference, and the global optimal solution is obtained through local optimal path recursion, so that the line spectrum inter-element phase difference deblurring success rate of the ship radiated noise under the condition of low signal-to-noise ratio is significantly improved.

Description

Technical Field

[0001] This invention belongs to the field of underwater acoustic signal processing technology, specifically relating to a method and system for de-ambiguing the phase difference between line spectrum elements in a distorted towed array. It is particularly suitable for de-ambiguing the phase difference of high-frequency line spectrum signals in ship radiated noise to estimate the array time delay vector, and can be used in sonar signal processing systems for underwater target detection, positioning, and tracking. Background Technology

[0002] Line spectrum components generated by the unavoidable vibrations of mechanical equipment such as diesel generators and air conditioning systems are a very important and useful component of ship radiated noise. Typically, the power of a line spectrum component is at least 5–10 dB higher than the power of its neighboring continuous spectrum, making it easy to detect and identify. The phase of these relatively strong line spectrum components contains time delay information from the source to the hydrophone; therefore, these phases can be used to estimate the time delay difference of the radiated noise signal received by different hydrophones. It is worth noting that the accuracy of time delay difference estimation using the phase difference of line spectrum components is sensitive to noise and decreases sharply as the signal-to-noise ratio (SNR) of the line spectrum component decreases. Furthermore, for line spectrum components with the same SNR, the accuracy of time delay difference estimation is proportional to the frequency of the line spectrum component; that is, the estimation accuracy of high-frequency line spectrum components is better than that of low-frequency line spectrum components. However, for phase difference measurements of high-frequency line spectrum components with a half-wavelength smaller than the element spacing, modulo 2π ambiguity often exists. If phase difference ambiguity is ignored and the ambiguous phase difference measurement is still used to estimate the time delay difference, the obtained time delay difference estimate will deviate significantly from the true value.

[0003] McKilliam et al. proposed a least-squares phase difference defuzzification estimator to transform the phase difference defuzzification problem into a problem of nearest points in the lattice, and solved the problem using lattice theory in the least-squares sense. However, this high-frequency line spectrum defuzzification method only utilizes the low-frequency line spectrum information of the current array elements, and the defuzzification success rate is not high when the signal-to-noise ratio is low or the number of line spectra is small. Summary of the Invention

[0004] Purpose of the invention: To overcome the shortcomings of the prior art, this invention provides a spatiotemporal joint defuzzification method and system for phase difference between distorted dragged array elements. By constructing a three-dimensional state space of time frame-spatial element-time delay difference, introducing spatial and temporal smoothing constraints, and realizing global optimal path search based on dynamic programming, the success rate of defuzzification of phase difference between high-frequency array elements under low signal-to-noise ratio conditions is significantly improved.

[0005] Technical Solution: To achieve the above objectives, the present invention provides a method for joint defuzzification of phase difference between elements of a distorted dragged array, comprising the following steps:

[0006] S1: Construct a three-dimensional phase difference residual tensor based on lattice theory phase difference defuzzification, consisting of time frames, spatial array elements, and discrete time delay differences;

[0007] S2: Construct a system optimization objective, find the optimal path in the three-dimensional state space to minimize the joint loss function, which includes a total phase residual term, a time smoothing constraint term, and a spatial smoothing constraint term; wherein, the time and spatial smoothing weight parameters are dynamically obtained through the global mean of the phase residual tensor.

[0008] S3: Based on the principle of dynamic programming, the three-dimensional residual tensor is recursively solved, and the cumulative minimum cost, spatiotemporally optimal predecessor state and path index matrix are calculated point by point.

[0009] S4: Obtain the optimal discrete time delay difference for each time frame and each array element through reverse recursion, and then complete the high-frequency line spectrum phase difference deblurring.

[0010] Preferably, step S1 specifically includes:

[0011] The upper bound of the candidate time delay difference is determined based on the propagation speed of sound waves in water and the nominal element spacing of the towed linear array.

[0012] A discrete set of candidate delay differences is generated based on the sampling rate and the number of candidate delay differences;

[0013] For each time frame, each array element, and each candidate time delay difference, the time delay difference estimate and phase difference residual are iteratively calculated using the phase difference measurement value and signal-to-noise ratio estimate of multiple line spectra through the weighted least squares method.

[0014] By traversing all time frames, array elements, and candidate delay differences, a three-dimensional phase difference residual tensor is constructed. Each element of the tensor corresponds to the phase residual value under a specific time frame, a specific array element, and a specific candidate delay difference.

[0015] Preferably, the joint loss function in step S2 Represented as:

[0016]

[0017] in The total number of time frames observed by the system. The number of phase differences. , , The two numbers to be solved are the first and second numbers. The first time frame, the first The spatial array element, the first The first time frame, the first The spatial array element, the first The first time frame, the first The optimal state corresponding to each spatial array element. Indicates the first The first time frame, the first Each spatial array element and the phase residual value corresponding to the optimal state. and For time and space smoothing weight parameters; ;in The global average residual. This is the preset regularization coefficient.

[0018] Preferably, step S3 specifically includes:

[0019] Initialize the recursion starting point:

[0020]

[0021] in Indicates the first The first time frame, the first The spatial array element, the first The phase residual value corresponding to each time delay difference state Indicates the distance from the starting point to the position. transition to state The cumulative cost; Defined as a time predecessor matrix, used to record the time predecessor state index; Defined as a spatial predecessor matrix, used to record spatial predecessor state indices; Defined as the transition direction matrix; The total number of time frames observed by the system. The number of phase differences. The number of candidate delay differences;

[0022] Set time boundaries:

[0023]

[0024]

[0025]

[0026]

[0027] in The previous frame's state traversal index is used. When the frame is located at a time boundary and there is no preceding spatial position, the transition direction is set to time transition, and the optimal predecessor at the time boundary is recorded. For time smoothing weight parameters;

[0028] Set spatial boundaries:

[0029]

[0030]

[0031]

[0032]

[0033] When located at a spatial boundary and without a preceding time position, set the transfer direction to spatial transfer and record the optimal predecessor at the spatial boundary. These are the time and space regularization smoothing weight parameters;

[0034] For non-boundary internal states, first calculate the time transition cost and the space transition cost separately:

[0035]

[0036]

[0037] like Then execute:

[0038]

[0039]

[0040]

[0041] Otherwise, execute:

[0042]

[0043]

[0044]

[0045] Preferably, step S4 specifically includes:

[0046] After the recursion reaches the endpoint, determine the optimal candidate delay difference for the final position:

[0047]

[0048] in The total number of time frames observed by the system. The number of phase differences. This is the cumulative cost matrix;

[0049] From the finish line Working backwards from the starting point, for and The optimal state sequence to be solved The recursive formula is:

[0050]

[0051] in For the time precursor matrix, For spatial precursor matrix, Here is the transition direction matrix;

[0052] Based on the optimal state sequence, obtain the accurate estimation matrix of the time delay difference between array elements, and calculate the phase difference after deblurring accordingly.

[0053] As a preferred option, the first The first time frame, the first Precise estimate of the time difference of each spatial array element , The first phase difference defuzzification calculation based on lattice theory is the first... The first time frame, the first The first spatial array element, the The estimated time delay difference corresponding to the minimum phase residual of the candidate time delay difference; the final defuzzified time delay difference of the i-th... The first time frame, the first The spatial array element, the first The spectrum at frequency Phase difference at .

[0054] This invention also provides a spatiotemporal joint defuzzification system for the phase difference between spectral elements of a distorted dragged array, used in the method described above. The system includes:

[0055] The 3D phase difference residual tensor construction module is used to construct a 3D phase difference residual tensor composed of time frames, spatial array elements, and discrete time delay differences based on the phase difference defuzzification of lattice theory.

[0056] The optimization model building module is used to construct the system optimization objective and find the optimal path in the three-dimensional state space to minimize the joint loss function. The joint loss function includes a phase total residual term, a time smoothing constraint term, and a spatial smoothing constraint term. The time and spatial smoothing weight parameters are dynamically obtained through the global mean of the phase residual tensor.

[0057] The forward recursion module is used to recursively solve the three-dimensional residual tensor based on the principle of dynamic programming, and calculate the cumulative minimum cost, spatiotemporally optimal predecessor state and path index matrix point by point;

[0058] The reverse recursion module is used to obtain the optimal discrete time delay difference for each time frame and each array element through reverse recursion, thereby completing the high-frequency line spectrum phase difference deblurring.

[0059] The present invention also provides a computer system, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, it implements the steps of the aforementioned method for joint defuzzification of phase difference between distorted dragged array elements in a time-space-frequency manner.

[0060] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the aforementioned method for joint defuzzification of phase difference between spectral elements of a distorted dragged array.

[0061] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the aforementioned method for joint defuzzification of phase difference between distorted dragged array elements in a spatiotemporal frequency domain.

[0062] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages:

[0063] 1. Significantly improve the success rate of defuzzification: By constructing a three-dimensional joint optimization model of time frame-spatial array element-discrete time delay difference state, the continuity and correlation of phase difference in the spatiotemporal dimension are comprehensively utilized, overcoming the limitations of traditional methods that only utilize single frame and single array element information. Especially under the conditions of low signal-to-noise ratio and few line spectra, the success rate of defuzzification is significantly improved.

[0064] 2. By employing dynamic programming for recursive solution and path backtracking, the final obtained time delay difference state sequence is the globally optimal solution, avoiding local optima problems, thus obtaining a more accurate and stable phase difference estimate. Furthermore, this invention introduces an adaptive regularization parameter adjustment mechanism based on residual statistical characteristics, enabling the smoothing weight to change dynamically with the input data, improving the algorithm's adaptability to different marine environments and signal conditions, and enhancing the overall robustness of the system. Attached Figure Description

[0065] Figure 1 This is a schematic flowchart of a method for joint defuzzification of phase difference between elements of a distorted dragged array line spectrum, provided in an embodiment of the present invention.

[0066] Figure 2 This is a schematic diagram showing the nominal and actual positions of the distorted drag array in an embodiment of the present invention.

[0067] Figure 3 The graph shows the success rate of defuzzification between array elements as the signal-to-noise ratio changes, compared to existing methods using the method of this invention. Detailed Implementation

[0068] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0069] like Figure 1 As shown in the figure, the present invention provides a method for joint defuzzification of phase difference between elements of a distorted dragged array, comprising the following steps:

[0070] S1: Construct a three-dimensional phase difference residual tensor based on lattice theory phase difference defuzzification, consisting of time frames, spatial array elements, and discrete time delay differences;

[0071] S2: Construct a system optimization objective, find the optimal path in the three-dimensional state space to minimize the joint loss function, which includes a total phase residual term, a time smoothing constraint term, and a spatial smoothing constraint term; wherein, the time and spatial smoothing weight parameters are dynamically obtained through the global mean of the phase residual tensor.

[0072] S3: Based on the principle of dynamic programming, the three-dimensional residual tensor is recursively solved, and the cumulative minimum cost, spatiotemporally optimal predecessor state and path index matrix are calculated point by point.

[0073] S4: Obtain the optimal discrete time delay difference for each time frame and each array element by reverse recursion, and then complete the high-frequency line spectrum phase difference deblurring;

[0074] Specifically, step S1 may include the following steps:

[0075] Step S101: Determine the upper bound of the candidate delay difference. and the lower world :

[0076]

[0077] Where c is the speed of sound in water, and d is the nominal element spacing of the towed linear array.

[0078] Step S102: Determine candidate time delay differences based on the upper and lower bounds of the time delay difference. :

[0079]

[0080] in The sampling rate during the digitization process of data received by the towed linear array. The number of candidate delay differences.

[0081] Step S103: Initialize the time frame index Array element position index Candidate Delay Difference Index , , The total number of time frames observed by the system. This represents the number of phase differences.

[0082] Step S104: Calculate the phase difference for phase compensation based on the candidate time delay difference. :

[0083]

[0084] in Indicates the input variable After rounding, L represents the number of line spectra detected in the power spectrum. For array element Data Acquisition and Array Elements Data collected at time frame t, frequency The estimated value of the phase difference at that point.

[0085] Step S105, based on phase difference Calculate the weighted least squares delay difference estimate for the current candidate delay difference. :

[0086]

[0087] in The first element in the received signal of a single array element The estimated signal-to-noise ratio of the root line spectrum can then be used to calculate the phase difference residual of the current candidate time delay difference. :

[0088]

[0089] Step S106: Calculate based on time delay difference Phase difference for phase compensation :

[0090]

[0091] in The phase compensation function can be defined as follows:

[0092]

[0093] Step S107, based on phase difference Calculate the weighted least squares delay difference estimate for the current candidate delay difference. :

[0094]

[0095] This allows us to calculate the phase difference residual of the current candidate delay difference. :

[0096]

[0097] Step S108: Compare the least squares residuals and ,if ,So:

[0098]

[0099] Then proceed to step S109. Otherwise:

[0100]

[0101] And return to step S106. Wherein Given time frame t, array element m, and the minimum phase difference residual of the k-th candidate delay difference, It is the estimated time delay difference corresponding to the minimum phase residual.

[0102] Step S109: Proceed with the following logical progression and execute:

[0103] 1. If Then let And return to step S104.

[0104] 2. If and Then let And return to step S104;

[0105] 3. If and Then let And return to step S104;

[0106] 4. If , and Then proceed to step S110.

[0107] Step S110: Construct the three-dimensional phase difference residual tensor :

[0108]

[0109] in Indicates the first The first time frame, the first The spatial array element, the first The phase residual value corresponding to each time delay difference state.

[0110] Specifically, step S2 may include the following steps:

[0111] Step S201: Construct the system optimization objective and find the optimal path in the three-dimensional state space such that the joint loss function... Minimum:

[0112]

[0113] The first term is the total phase residual term, the second term is the time smoothing constraint term, and the third term is the spatial smoothing constraint term. Let be the optimal state to be solved. and These are the time and space smoothing parameters.

[0114] Step S202: Adaptively determine time smoothing parameters based on the global statistical properties of the residual tensor. With spatial smoothing parameters :

[0115]

[0116] in The global average residual. In this embodiment, the preset regularization coefficient is set to... .

[0117] Specifically, step S3 may include the following steps:

[0118] Step S301: Initialize the recursion starting point ( , ):

[0119]

[0120] in Defined as a cumulative cost matrix, Indicates the distance from the starting point to the position. transition to state The cumulative cost; Defined as a time predecessor matrix, used to record the time predecessor state index; Defined as a spatial predecessor matrix, used to record spatial predecessor state indices; Defined as a transfer direction matrix (0: starting point, 1: time transfer, 2: spatial transfer).

[0121] Step S302: Set time boundaries ( , ):

[0122]

[0123]

[0124]

[0125]

[0126] in Use the previous frame's state traversal index. When at a time boundary and without a preceding spatial location, set the transition direction to time transition and record the optimal predecessor at the time boundary.

[0127] Step S303: Set spatial boundaries ( , ):

[0128]

[0129]

[0130]

[0131]

[0132] When located at the spatial boundary, and without a preceding time position, set the transfer direction to spatial transfer and record the optimal predecessor at the spatial boundary.

[0133] Step S304, under normal circumstances ( , ):

[0134] First, calculate the time transfer cost and the space transfer cost separately:

[0135]

[0136]

[0137] like Then execute:

[0138]

[0139]

[0140]

[0141] Otherwise, execute:

[0142]

[0143]

[0144]

[0145] Specifically, step S4 may include the following steps:

[0146] Step S401: After determining that the recursion has reached the endpoint, determine the optimal candidate delay difference for the final position:

[0147]

[0148] Step S402, from the endpoint Working backwards from the starting point, for and The optimal state sequence to be solved The recursive formula can be written as:

[0149]

[0150] The recursive process uses time-decreasing ( ) and spatial decrease ( The double-loop sequence ensures that each state... During computation, its dependent subsequent states and The optimal strategy has been pre-calculated.

[0151] Step S403: Based on the optimal state sequence Obtain the accurate estimation matrix of time difference between array elements. Accurately estimate the matrix based on the time difference between array elements. To obtain the phase difference after final deblurring .

[0152] The following simulation examples demonstrate the effectiveness and advantages of this invention.

[0153] Simulation Example 1: The simulation parameters for ship radiated noise are set as follows: power spectral line frequencies of 200 Hz, 400 Hz, 600 Hz, 1500 Hz, and 3000 Hz, and the number of array elements is 32. The element spacing is d=0.8m, and the distorted drag array is as follows: Figure 2 As shown, the speed of sound underwater is c = 1500 m / s, the narrowband signal-to-noise ratio of the line spectrum is set to -10dB to 10dB, and the sampling frequency is [missing information]. =20kHz, the array data frame number T is 32 frames, the duration of each frame is 1s, and the angle at which the ship's radiated noise is incident on the towed linear array is 60°.

[0154] Based on step S1, determine the upper and lower bounds of the time delay difference. Number of candidate delay differences =215, by iterating through the time state index Spatial state index from 1 to T Status index from 1 to M From 1 to Construct a three-dimensional phase difference residual tensor .

[0155] Based on step S2, set the regularization coefficient. Then, a joint loss function is constructed. .

[0156] Based on step S3, at the starting point of the recursion, set:

[0157]

[0158] At the time boundary ( , ),set up:

[0159]

[0160]

[0161]

[0162]

[0163] At the spatial boundary ( , ),set up:

[0164]

[0165]

[0166]

[0167]

[0168] Under normal circumstances ( , ):

[0169] First, calculate the time transfer cost and the space transfer cost separately:

[0170]

[0171]

[0172] like Then execute:

[0173]

[0174]

[0175]

[0176] Otherwise, execute:

[0177]

[0178]

[0179]

[0180] Based on step S4, after the recursive calculation is completed, the optimal state of the last point is determined:

[0181]

[0182] And work backwards from the finish line to the starting point:

[0183]

[0184] When recursively returning to the starting point, based on the optimal state sequence Accurate estimation of time delay difference between array elements This allows us to obtain the phase difference after final deblurring. The success rate of high-frequency line spectrum deblurring varies with the line spectrum signal-to-noise ratio as follows: Figure 3 As shown.

[0185] Simulation results show that the proposed high-frequency line spectrum phase difference deblurring method based on spatiotemporal-frequency three-dimensional joint deblurring significantly outperforms traditional frequency domain deblurring methods under distorted dragged array conditions, especially exhibiting stronger robustness in low signal-to-noise ratio (SNR) scenarios. Under extremely low SNR conditions of -10 dB in narrowband, the success rate of the three-dimensional spatiotemporal-frequency joint deblurring reaches 0.52, while that of the frequency domain deblurring is only 0.35, representing an improvement of approximately 48.5%. When SNR ≥ 5 dB, the success rate of the three-dimensional method rapidly approaches 1, while the frequency domain deblurring method only approaches saturation at levels above 10 dB. The threshold SNR for successful complete line spectrum deblurring is approximately 4-5 dB lower than that of the frequency domain deblurring method.

[0186] Based on the same inventive concept, this invention also provides a spatiotemporal-frequency joint defuzzification system for phase difference between distorted dragged array elements, used to implement any of the aforementioned spatiotemporal-frequency joint defuzzification methods for phase difference between distorted dragged array elements. The system includes: a three-dimensional phase difference residual tensor construction module, used to construct a three-dimensional phase difference residual tensor composed of time frames, spatial elements, and discrete time delay differences based on lattice theory phase difference defuzzification; an optimization model construction module, used to construct the system optimization objective, find the optimal path in the three-dimensional state space to minimize the joint loss function, the joint loss function including a total phase residual term, a time smoothing constraint term, and a spatial smoothing constraint term; dynamically obtain time and spatial smoothing weight parameters through the global mean of the phase residual tensor; a forward recursion module, used to recursively solve the three-dimensional residual tensor based on the dynamic programming principle, calculating the cumulative minimum cost, spatiotemporally optimal predecessor state, and path index matrix point by point; and a backward recursion module, used to recursively obtain the optimal discrete time delay difference for each time frame and each array element, thereby completing the high-frequency line spectrum phase difference defuzzification.

[0187] The present invention also provides a computer system, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, it implements the steps of the aforementioned method for joint defuzzification of phase difference between distorted dragged array elements in a time-space-frequency manner.

[0188] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the aforementioned method for joint defuzzification of phase difference between spectral elements of any of the distorted dragged arrays.

[0189] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the aforementioned method for joint defuzzification of phase difference between spectral elements of any of the distorted dragged arrays.

[0190] The program code used to implement the method of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the steps of the method of the present invention to be performed. The program code can be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a standalone software package, or entirely on a remote machine or server. All aspects not detailed in this invention are well-known to those skilled in the art.

Claims

1. A method for joint spatial-frequency deambiguation of phase difference between elements of a distorted dragged array linear spectrum, characterized in that, Includes the following steps: S1: Construct a three-dimensional phase difference residual tensor based on lattice theory phase difference defuzzification, consisting of time frames, spatial array elements, and discrete time delay differences; S2: Construct a system optimization objective, find the optimal path in the three-dimensional state space to minimize the joint loss function, which includes a total phase residual term, a time smoothing constraint term, and a spatial smoothing constraint term; wherein, the time and spatial smoothing weight parameters are dynamically obtained through the global mean of the phase residual tensor. S3: Based on the principle of dynamic programming, the three-dimensional residual tensor is recursively solved, and the cumulative minimum cost, spatiotemporally optimal predecessor state and path index matrix are calculated point by point. S4: Obtain the optimal discrete time delay difference for each time frame and each array element through reverse recursion, and then complete the high-frequency line spectrum phase difference deblurring.

2. The method for joint defuzzification of phase difference between elements of a distorted dragged array as described in claim 1, characterized in that, Step S1 specifically includes: The upper bound of the candidate time delay difference is determined based on the propagation speed of sound waves in water and the nominal element spacing of the towed linear array. A discrete set of candidate delay differences is generated based on the sampling rate and the number of candidate delay differences; For each time frame, each array element, and each candidate time delay difference, the time delay difference estimate and phase difference residual are iteratively calculated using the phase difference measurement value and signal-to-noise ratio estimate of multiple line spectra through the weighted least squares method. By traversing all time frames, array elements, and candidate delay differences, a three-dimensional phase difference residual tensor is constructed. Each element of the tensor corresponds to the phase residual value under a specific time frame, a specific array element, and a specific candidate delay difference.

3. The method for joint defuzzification of phase difference between elements of a distorted dragged array as described in claim 1, characterized in that, The joint loss function in step S2 Represented as: ； in The total number of time frames observed by the system. The number of phase differences. , , The two numbers to be solved are the first and second numbers. The first time frame, the first The spatial array element, the first The first time frame, the first The spatial array element, the first The first time frame, the first The optimal state corresponding to each spatial array element. Indicates the first The first time frame, the first Each spatial array element and the phase residual value corresponding to the optimal state. and For time and space smoothing weight parameters; ;in The global average residual. This is the preset regularization coefficient.

4. The method for joint defuzzification of phase difference between elements of a distorted dragged array as described in claim 1, characterized in that, Step S3 specifically includes: Initialize the recursion starting point: ； in Indicates the first The first time frame, the first The spatial array element, the first The phase residual value corresponding to each time delay difference state Indicates the distance from the starting point to the position. transition to state The cumulative cost; Defined as a time predecessor matrix, used to record the time predecessor state index; Defined as a spatial predecessor matrix, used to record spatial predecessor state indices; Defined as the transition direction matrix; The total number of time frames observed by the system. The number of phase differences. The number of candidate delay differences; Set time boundaries: ； ； ； ； in The previous frame's state traversal index is used. When the frame is located at a time boundary and there is no preceding spatial position, the transition direction is set to time transition, and the optimal predecessor at the time boundary is recorded. For time smoothing weight parameters; Set spatial boundaries: ； ； ； ； When located at a spatial boundary and without a preceding time position, set the transfer direction to spatial transfer and record the optimal predecessor at the spatial boundary. These are the time and space regularization smoothing weight parameters; For non-boundary internal states, first calculate the time transition cost and the space transition cost separately: ； ； like Then execute: ； ； ； Otherwise, execute: ； ； 。 5. The method for joint defuzzification of phase difference between elements of a distorted dragged array as described in claim 1, characterized in that, Step S4 specifically includes: After the recursion reaches the endpoint, determine the optimal candidate delay difference for the final position: ； in The total number of time frames observed by the system. The number of phase differences. This is the cumulative cost matrix; From the finish line Working backwards from the starting point, for and The optimal state sequence to be solved The recursive formula is: ； in For the time precursor matrix, For spatial precursor matrix, Here is the transition direction matrix; Based on the optimal state sequence, obtain the accurate estimation matrix of the time delay difference between array elements, and calculate the phase difference after deblurring accordingly.

6. The method for joint defuzzification of phase difference between elements of a distorted dragged array as described in claim 5, characterized in that, No. The first time frame, the first Precise estimate of the time difference of each spatial array element , The first phase difference defuzzification calculation based on lattice theory is the first... The first time frame, the first The first spatial array element, the The estimated time delay difference corresponding to the minimum phase residual of the candidate time delay difference; the final defuzzified time delay difference of the i-th... The first time frame, the first The spatial array element, the first The spectrum at frequency Phase difference at .

7. A spatiotemporal joint defuzzification system for phase difference between elements of a distorted dragged array linear spectrum, used to implement the method according to any one of claims 1-6, characterized in that, include: The 3D phase difference residual tensor construction module is used to construct a 3D phase difference residual tensor composed of time frames, spatial array elements, and discrete time delay differences based on the phase difference defuzzification of lattice theory. The optimization model building module is used to construct the system optimization objective and find the optimal path in the three-dimensional state space to minimize the joint loss function. The joint loss function includes a phase total residual term, a time smoothing constraint term, and a spatial smoothing constraint term. The time and spatial smoothing weight parameters are dynamically obtained through the global mean of the phase residual tensor. The forward recursion module is used to recursively solve the three-dimensional residual tensor based on the principle of dynamic programming, and calculate the cumulative minimum cost, spatiotemporally optimal predecessor state and path index matrix point by point; The reverse recursion module is used to obtain the optimal discrete time delay difference for each time frame and each array element through reverse recursion, thereby completing the high-frequency line spectrum phase difference deblurring.

8. A computer system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the computer program is executed by the processor, it implements the steps of the spatial-temporal joint defuzzification method for phase difference between elements of a distorted dragged array as described in any one of claims 1-6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the spatial-temporal joint defuzzification method for phase difference between elements of a distorted dragged array as described in any one of claims 1-6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the spatial-temporal joint defuzzification method for phase difference between elements of a distorted dragged array as described in any one of claims 1-6.

Images