A phase unwrapping method and device applied to ship-bridge collision bridge body vibration monitoring

Abstract

The application provides a phase unwrapping method and device applied to ship-bridge collision bridge body vibration monitoring, and relates to the technical field of distributed optical fiber acoustic wave sensing. The method comprises the following steps: based on a self-attention mechanism, self-adaptive noise suppression is performed on a window phase signal sequence to obtain a denoised phase signal sequence; according to the denoised phase signal sequence, multi-scale time-frequency feature fusion is performed through a convolutional neural network to obtain a space-time matrix and a time-frequency feature; according to the space-time matrix, spatial correlation feature extraction is performed through a two-dimensional convolutional network to obtain a spatial feature; based on a long short-term memory network, weighted fusion is performed on the time-frequency feature and the spatial feature to obtain a space-time fusion feature; and according to the space-time fusion feature, unwrapping phase mapping is performed through a full connection layer to obtain a real phase signal sequence. The application is a phase unwrapping method for distributed optical fiber acoustic wave vibration in a ship collision scene, which has high accuracy and good robustness.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of distributed optical fiber acoustic sensing, in particular to a phase unwrapping method and device applied to ship-bridge collision body vibration monitoring. BACKGROUND

[0002] Ship collision is one of the major safety threats faced by bridge structures, especially in water areas with frequent shipping, bridges are at risk of being hit by out-of-control ships, overloading or extreme weather. As a high-precision, long-range structural health monitoring technology, distributed optical fiber acoustic vibration sensing system can capture real-time vibration signals of bridge structures through sensing optical fibers, providing key data support for ship collision event warning and damage assessment. However, the distributed optical fiber acoustic vibration sensing system has a significant phase wrapping problem in high-frequency vibration scenarios (such as shock waves generated by ship collisions): due to the periodicity limitation of phase measurement (usually 2π), when the vibration frequency or amplitude exceeds the system dynamic range, the phase signal will be truncated modulo 2π, resulting in serious distortion of amplitude, frequency and phase information. This problem directly weakens the ability of the distributed optical fiber acoustic vibration sensing system to identify ship collision events, such as being unable to accurately determine the impact energy, locate the impact position or distinguish between impact and other environmental noise (such as wind vibration, vehicle load), significantly reducing the reliability of the warning system.

[0003] Traditional phase unwrapping methods (such as time domain integration method, Doppler unwrapping, etc.) face multiple challenges in complex engineering scenarios, the instantaneous high-frequency impact signal generated by ship collision exceeds the dynamic range of traditional algorithms, resulting in increased unwrapping error; these methods rely on manually designed feature extraction rules, making it difficult to adapt to the diversity of different bridge structures and ship collision scenarios; the calculation complexity of traditional algorithms is high, which cannot meet the real-time requirements of the warning system (response time often exceeds seconds), missing the best warning opportunity.

[0004] Distributed optical fiber acoustic vibration sensing system can monitor the vibration signals along the long-distance optical fiber in real time through the distributed sensing characteristics of the optical fiber, and is widely used in bridge safety monitoring, geological disaster warning and other fields. However, the distributed optical fiber acoustic vibration sensing system generally faces the problem of phase wrapping in the signal demodulation process, which is due to the periodic truncation of phase information within 2π range caused by the interference measurement principle, so that the phase value of the original vibration signal is limited between -π and π, which cannot directly reflect the true vibration amplitude and frequency information. To solve this problem, existing technologies mainly rely on traditional mathematical algorithms and model-driven methods, but these methods have significant limitations in practical applications.

[0005] The core of traditional phase unwrapping techniques is to extend the truncated phase values to the continuous domain through mathematical operations. For example, the Least Square Method (LSM) estimates the unwrapped phase by fitting polynomials or curves, but this method is sensitive to noise, especially in low signal-to-noise ratio environments, and can introduce errors, and cannot effectively handle rapid phase jumps in high-frequency vibration signals. Although high-order polynomial fitting can improve fitting accuracy, it requires a large amount of prior knowledge to set the polynomial order and initial conditions, has high computational complexity, and is difficult to implement real-time processing, especially in long-distance optical fiber monitoring, the exponential growth of data volume will further reduce the efficiency of the algorithm. In addition, iterative unwrapping methods based on models (such as phase unwrapping algorithms) rely on assumptions about the physical model of the vibration signal, and when the actual vibration signal contains complex noise or nonlinear characteristics, model mismatch can cause cumulative errors, leading to systematic deviations in the unwrapping results. Some technologies also try to combine hardware compensation methods, such as increasing the sampling frequency or introducing auxiliary reference signals to alleviate the wrapping problem, but hardware upgrades increase system costs, and it is still difficult to completely eliminate the wrapping phenomenon in dynamic environments or long-distance deployments.

[0006] The deficiencies of existing phase unwrapping techniques for distributed optical fiber acoustic vibration sensing systems mainly manifest in the following aspects: First, traditional phase unwrapping algorithms lack robustness to noise, especially in complex environments such as bridges, wind vibration, vehicle loads, and environmental temperature changes can introduce multi-source noise, leading to distortion of the unwrapped phase signal. Second, these methods perform poorly when dealing with non-stationary vibration signals, and the vibration frequency and amplitude of bridge structures vary greatly under different working conditions, and traditional algorithms have difficulty in adaptively adjusting model parameters, which can easily produce local optimal solutions or unwrapping errors. Third, the contradiction between computational efficiency and real-time performance limits the engineering application of distributed optical fiber acoustic vibration sensing systems, for example, in long-distance monitoring, traditional algorithms need to process phase data point by point or in segments, which cannot meet the real-time alarm requirements of bridge safety monitoring. In addition, existing phase unwrapping methods are highly dependent on the initial phase estimate, and if the initial value is biased, the subsequent unwrapping process will fail due to error accumulation, especially when the fiber layout path is complex or there is multi-node interference, this problem is more prominent. Finally, traditional phase unwrapping techniques cannot simultaneously consider the unwrapping accuracy of different frequency components, rapid phase changes in high-frequency vibration signals can exceed the processing capacity of the algorithm, and small phase fluctuations in low-frequency signals can be easily masked by noise, leading to limited overall monitoring performance. These defects make it difficult for existing phase unwrapping techniques to meet the needs of high accuracy, high real-time performance, and strong environmental adaptability in actual applications, especially in bridge scenarios that require extremely high safety and reliability, a new unwrapping scheme that can break through the limitations of traditional methods is urgently needed.

[0007] In the prior art, there is a lack of a phase unwrapping method for distributed fiber acoustic wave vibration in a ship collision scenario, which is accurate and robust. SUMMARY

[0008] To solve the technical problems of insufficient vibration signal demodulation accuracy, poor dynamic complex scene adaptability and low real-time processing efficiency caused by phase wrapping phenomenon when the distributed fiber acoustic wave vibration sensing system is applied to ship collision bridge monitoring in the prior art, the embodiments of the present application provide a phase unwrapping method and device applied to ship collision bridge body vibration monitoring. The technical solution is as follows:

[0009] On the one hand, a phase unwrapping method applied to ship collision bridge body vibration monitoring is provided, which is realized by a phase unwrapping device, and the method comprises:

[0010] The signal acquisition module is configured to acquire monitoring information by a distributed fiber vibration sensing system, and obtain a phase signal sequence of bridge vibration; and based on a preset length of a sliding window, the phase signal sequence is segmented to obtain a window phase signal sequence.

[0011] The signal denoising module is configured to perform adaptive noise suppression on the window phase signal sequence based on a self-attention mechanism to obtain a denoised phase signal sequence.

[0012] According to the denoised phase signal sequence, multi-scale time-frequency feature fusion is performed through a convolutional neural network to obtain a space-time matrix and a time-frequency feature.

[0013] According to the space-time matrix, spatial correlation feature extraction is performed through a two-dimensional convolutional network to obtain a spatial feature.

[0014] Based on a long short-term memory network, the time-frequency feature and the spatial feature are weighted and fused to obtain a space-time fusion feature.

[0015] According to the space-time fusion feature, unwrapping phase mapping is performed through a fully connected layer to obtain a real phase signal sequence.

[0016] On the other hand, a phase unwrapping device applied to ship collision bridge body vibration monitoring is provided, which is applied to the phase unwrapping method for ship collision bridge body vibration monitoring, and the device comprises:

[0017] The signal acquisition module is configured to acquire monitoring information by a distributed fiber vibration sensing system, and obtain a phase signal sequence of bridge vibration; and based on a preset length of a sliding window, the phase signal sequence is segmented to obtain a window phase signal sequence.

[0018] The signal denoising module is configured to perform adaptive noise suppression on the window phase signal sequence based on a self-attention mechanism to obtain a denoised phase signal sequence.

[0019] a time-frequency feature extraction module configured to obtain a time-space matrix and time-frequency features by performing multi-scale time-frequency feature fusion on the denoised phase signal sequence through a convolutional neural network;

[0020] a spatial feature extraction module configured to obtain spatial features by performing spatial correlation feature extraction on the time-space matrix through a two-dimensional convolutional network;

[0021] a feature fusion module configured to obtain time-space fusion features by performing weighted fusion on the time-frequency features and the spatial features based on a long short-term memory network;

[0022] a signal unwrapping module configured to obtain a real phase signal sequence by performing unwrapping phase mapping on the time-space fusion features through a fully connected layer.

[0023] In another aspect, a phase unwrapping device is provided, which comprises a processor and a memory having computer readable instructions stored thereon, the computer readable instructions being executed by the processor to implement any one of the above phase unwrapping methods applied to ship-bridge body vibration monitoring.

[0024] In another aspect, a computer readable storage medium is provided, which stores at least one instruction, the at least one instruction being loaded and executed by a processor to implement any one of the above phase unwrapping methods applied to ship-bridge body vibration monitoring.

[0025] The technical solutions provided by the embodiments of the present application have at least the following beneficial effects:

[0026] The application provides a phase unwrapping method applied to ship-bridge body vibration monitoring, the inherent physical law and noise distribution characteristics of a phase signal are learned through large-scale training data of a deep learning model, random noise, environmental temperature fluctuation, and multi-source vibration superposition have stronger robustness, especially in a low signal-to-noise ratio environment, the phase unwrapping method still has high-precision unwrapping ability, and the phase jump and cumulative error problems caused by noise sensitivity of a traditional algorithm are solved; real-time processing of massive phase data along a long-distance optical fiber can be realized, the calculation efficiency bottleneck caused by point-by-point expansion of a traditional iterative algorithm is broken through, and the demand of bridge safety monitoring for real-time alarm and rapid response is met; through an adaptive parameter optimization mechanism, the deep learning model can dynamically adapt to vibration signal characteristics of different frequency components, high-frequency vibration rapid phase changes can be accurately captured, and phase fluctuations of low-frequency micro-vibration can be effectively identified, and the compatibility and overall monitoring performance of the system for multi-band vibration signals are significantly improved; through the feature fusion capability of deep learning, the recognition capability for complex vibration modes can be further enhanced, high-reliability and high-resolution vibration sensing data are provided for bridge structure health monitoring, the dependence of the system on hardware compensation means is significantly reduced, and the engineering deployment cost is reduced.

[0027] The application reconstructs the algorithm framework of phase unwrapping through a deep learning technology, fundamentally solves the limitations of traditional methods in noise robustness, calculation efficiency, dynamic adaptability, and global accuracy, and lays a technical foundation for large-scale application of a distributed optical fiber acoustic vibration sensing system in bridge safety monitoring and other high-reliability scenes. The application is a phase unwrapping method with high accuracy and good robustness for distributed optical fiber acoustic vibration in a ship impact scene. BRIEF DESCRIPTION OF DRAWINGS

[0028] In order to more clearly illustrate the technical solutions in the embodiments of the application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0029] Figure 1 is a phase unwrapping method flow chart applied to ship-bridge body vibration monitoring provided by the embodiment of the application;

[0030] Figure 2 is a phase unwrapping device block diagram applied to ship-bridge body vibration monitoring provided by the embodiment of the application;

[0031] Figure 3 is a structural schematic diagram of a phase unwrapping device provided by the embodiment of the application. DETAILED DESCRIPTION

[0032] The technical solutions in the present application will be described below with reference to the drawings.

[0033] In the embodiments of the present application, the words such as "example", "for example" are used to represent an example, illustration or description. Any embodiment or design scheme described as "example" in the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. Rather, the word "example" is intended to present the concept in a specific manner. In addition, in the embodiments of the present application, the meaning expressed by "and / or" can be both, or can be one of the two.

[0034] In the embodiments of the present application, "image" and "picture" can be used interchangeably at times, and it should be pointed out that the meanings expressed are consistent when the distinction is not emphasized. "Of", "corresponding" and "corresponding" can be used interchangeably at times, and it should be pointed out that the meanings expressed are consistent when the distinction is not emphasized.

[0035] In the embodiments of the present application, sometimes the subscript such as W1 can be written in the form of non-subscript such as W1, and the meanings expressed are consistent when the distinction is not emphasized.

[0036] In order to make the technical problems, technical solutions and advantages of the present application clearer, the following will be described in detail with reference to the drawings and specific embodiments.

[0037] The embodiments of the present application provide a phase unwrapping method applied to ship-bridge body vibration monitoring, which can be realized by a phase unwrapping device, which can be a terminal or a server. As shown in the flow chart of the phase unwrapping method applied to ship-bridge body vibration monitoring, the processing flow of the method can include the following steps: Figure 1 As shown in the flow chart of the phase unwrapping method applied to ship-bridge body vibration monitoring, the processing flow of the method can include the following steps:

[0038] S1, monitoring information collection is performed through a distributed optical fiber vibration sensing system to obtain a phase signal sequence of bridge vibration; based on a preset length sliding window, the phase signal sequence is segmented to obtain a window phase signal sequence.

[0039] In a feasible implementation manner, the received distributed optical fiber vibration sensing system phase signal sequence is processed, is a time index. The phase signal sequence is segmented using a fixed length sliding window to ensure the continuity of event characteristics.

[0040] S2, based on a self-attention mechanism, self-adaptive noise suppression is performed on the window phase signal sequence to obtain a denoised phase signal sequence.

[0041] Optionally, based on the self-attention mechanism, the window phase signal sequence is subjected to adaptive noise suppression to obtain a denoised phase signal sequence, comprising:

[0042] According to the window phase signal sequence, long-distance dependence is captured to obtain a weighted value vector of an intermediate quantity of the window phase signal sequence;

[0043] According to the weighted value vector and a preset output weight matrix, calculation is performed to obtain an attention output of the window phase signal sequence;

[0044] Based on the attention output, dynamic residual analysis is performed on the window phase signal sequence according to the window phase signal sequence to obtain an adaptive setting threshold of a statistical characteristic;

[0045] According to the adaptive setting threshold and the attention output, binary mask denoising processing is performed on the window phase signal sequence to obtain a denoised phase signal sequence.

[0046] In a feasible implementation, the window phase signal sequence is processed using the self-attention mechanism, and a weighted value vector of an i-th position of an intermediate quantity is as follows:

[0047] (1)

[0048] wherein, , , is a learnable matrix for mapping the original feature to an attention space; is an i-th row vector of K, is a j-th column vector of K, is a j-th row vector of K; TN represents the number of elements of the window phase signal sequence. The attention output h is as follows:

[0049]

[0050] (2)

[0051] wherein, is an output weight matrix.

[0052] Through residual analysis and dynamic threshold, noise and signal regions are distinguished, and a mask is generated for noise suppression. The absolute difference between the original input and the attention output can be represented as a residual vector, and a threshold is adaptively set according to the statistical characteristics of the current residual as follows:

[0053] ​​​​​​​ (3);

[0054] in, is a coefficient.

[0055] The set of locations where the residual exceeds a threshold is defined as the noise region. Window phase signal sequence. The output X after noise reduction The value X at each position i As shown in equation (4):

[0056] (4);

[0057] in, It is a binary mask. When the residual is less than or equal to the threshold, then... (Preserve the original signal), when the residual is greater than the threshold, then (Replace with attention output).

[0058] S3. Based on the denoised phase signal sequence, multi-scale time-frequency feature fusion is performed through a convolutional neural network to obtain the spatiotemporal matrix and time-frequency features.

[0059] Optionally, based on the denoised phase signal sequence, multi-scale time-frequency feature fusion is performed using a convolutional neural network to obtain a spatiotemporal matrix and time-frequency features, including:

[0060] Based on preset signal conversion parameters, the denoised phase signal sequence is converted into a spatiotemporal matrix;

[0061] Based on the spatiotemporal matrix, multi-scale feature extraction is performed using a convolutional neural network to obtain multi-scale time-frequency local features.

[0062] Based on preset fusion weights, channel fusion is performed according to multi-scale time-frequency local features to obtain time-frequency features.

[0063] In one feasible implementation, the denoised window phase signal sequence X is converted into a spatiotemporal matrix based on preset signal conversion parameters, including repetition frequency and data acquisition card sampling speed. . Indicates time Spatiotemporal location The vibration phase. The spatial dimension of the spatiotemporal matrix corresponds to the sampling points along the sensing optical cable of the distributed optical fiber vibration sensing system, the time dimension corresponds to the sampling time points during continuous monitoring of the system, and the element values ​​of the matrix correspond to the amplitude.

[0064] The spatiotemporal matrix is ​​processed by N parallel convolution kernels. Extract local features at different time scales. Each convolutional kernel in time The local features extracted at the location are as follows (5):

[0065] (5);

[0066] in, It is the first Each convolutional kernel at time offset The weight parameters at that location, It is the first The size of each convolutional kernel, It is a bias term used to adjust the linear combination of features.

[0067] Multiscale features After splicing, pass 1 1. Convolution is used to perform channel fusion, and the fused time-frequency features are as follows (6):

[0068] (6);

[0069] in, It is the first The fusion weights of each scale feature are used to adjust the contribution of each scale feature to the final fusion result. It is the bias term of the fusion layer.

[0070] Will Feature sequences at different scales are concatenated along the channel dimension to form a high-dimensional feature vector, preserving information from all time scales. The fused feature sequence... It integrates time-frequency information of high-frequency, mid-frequency, and low-frequency vibrations.

[0071] In this system, the convolutional kernels are parallel; the number of kernels is 3; and the kernel size is 3. 3, 5 5 and 7 7.

[0072] In one feasible implementation, different convolutional kernel sizes correspond to different time windows. Small convolutional kernels capture local details of high-frequency vibrations, while large convolutional kernels extract long-term trends of low-frequency vibrations. Through parallel processing, the network can simultaneously focus on features at different time scales.

[0073] S4. Based on the spatiotemporal matrix, spatial correlation features are extracted using a two-dimensional convolutional network to obtain spatial features.

[0074] Among them, the two-dimensional convolutional network captures the spatial correlation of sensor information in the spatiotemporal matrix by establishing a local correlation model in the spatial dimension.

[0075] In one feasible implementation, a two-dimensional convolutional network is used to process phase data along the fiber at adjacent time points, spatial features. As shown in equation (7):

[0076] (7) ;

[0077] The two-dimensional convolution kernel considers the phase values of the current spatial position and its adjacent spatial positions at time t, models the local correlation in the spatial dimension, and captures the spatial correlation between sensors.

[0078] S5, based on a long short-term memory network, the time-frequency features and the spatial features are weighted and fused to obtain spatio-temporal fusion features.

[0079] Optionally, based on a long short-term memory network, the time-frequency features and the spatial features are weighted and fused to obtain spatio-temporal fusion features, including:

[0080] Based on the gating mechanism, the hidden state is obtained by capturing the dynamic long-term dependency relationship through the long short-term memory network according to the time-frequency features;

[0081] Based on the preset attention mechanism weight, the spatial attention weight is calculated according to the spatial features and the hidden state;

[0082] Based on the spatial attention weight, the time-frequency features and the spatial features are weighted and fused to obtain spatio-temporal fusion features.

[0083] In a feasible implementation, the fused feature sequence is sequentially modeled using a long short-term memory (LSTM) network. The LSTM dynamically updates the hidden state through a gating mechanism (forget gate, input gate, and output gate), and can remember long-term time dependencies, i.e., periodic or trend changes in the vibration signal, and capture long-term dependencies. The hidden state of the LSTM at time t is as follows: (8) ;

[0084] (8) ;

[0085] wherein, is the hidden state, which is dynamically updated through a gating mechanism (forget gate, input gate, and output gate).

[0086] The spatial features and the time features are combined and weighted through an attention mechanism, and the process is as follows:

[0087] (9) ;

[0088] wherein, is the spatial attention weight, which is calculated by a spatial attention module, and the process is as follows: ​

[0089] (10);

[0090] wherein, is the weight based on the attention mechanism, for the linear combination of joint time-frequency and spatial features.

[0091] S6, according to the spatio-temporal fusion feature, the phase mapping is unwrapped through the full connection layer to obtain a real phase signal sequence.

[0092] In a feasible implementation manner, the spatio-temporal feature and the nonlinear mapping of the real phase are established through the full connection layer:

[0093] (11);

[0094] wherein, is the weight matrix of the full connection layer, for the linear mapping of the feature to the frequency; is the bias term of the full connection layer.

[0095] The present application unwraps the phase wrapping problem of the distributed optical fiber acoustic vibration sensing system through the deep learning method, significantly improves the monitoring accuracy, real-time performance and environmental adaptability of the system in complex environment. Compared with the traditional mathematical algorithm, the deep learning model can automatically extract the multi-scale vibration features in the phase signal through the end-to-end feature learning mechanism, effectively overcome the influence of noise interference and nonlinear distortion.

[0096] In addition, the present application does not need to rely on artificial setting of polynomial order, initial phase value or physical model parameters, avoids the systematic deviation caused by lack of prior knowledge or model mismatch in traditional method, especially in the bridge scene with complex optical fiber layout path and multiple node interference, can effectively suppress the error accumulation effect, and ensure the global consistency of the unwrapping result.

[0097] The application provides a phase unwrapping method applied to ship-bridge body vibration monitoring, and the inherent physical law and noise distribution characteristics of a phase signal are learned through large-scale training data of a deep learning model, so that the method has stronger robustness to complex interference such as random noise, environmental temperature fluctuation and multi-source vibration superposition, and can still maintain high-precision unwrapping ability in a low signal-to-noise ratio environment, thereby solving the phase jump and cumulative error problems caused by noise sensitivity of a traditional algorithm; the method can realize real-time processing of massive phase data along a long-distance optical fiber, breaks through the calculation efficiency bottleneck caused by point-by-point expansion of a traditional iterative algorithm, meets the demand of bridge safety monitoring on real-time alarm and rapid response, and through an adaptive parameter optimization mechanism, the deep learning model can dynamically adapt to vibration signal characteristics of different frequency components, accurately captures rapid phase changes of high-frequency vibration, effectively identifies phase fluctuations of low-frequency micro-vibration, and significantly improves the compatibility and overall monitoring performance of the system to multi-band vibration signals; through the feature fusion capability of deep learning, the recognition capability to complex vibration modes can be further enhanced, high-reliability and high-resolution vibration sensing data are provided for bridge structure health monitoring, the dependence of the system on hardware compensation means is significantly reduced, and the engineering deployment cost is reduced.

[0098] The application reconstructs the algorithm framework of phase unwrapping through a deep learning technology, fundamentally solves the limitations of traditional methods in noise robustness, calculation efficiency, dynamic adaptability and global accuracy, and lays a technical foundation for large-scale application of a distributed optical fiber acoustic vibration sensing system in bridge safety monitoring and other high-reliability scenes. The application is a phase unwrapping method with high accuracy and good robustness for distributed optical fiber acoustic vibration in a ship impact scene.

[0099] Figure 2 A phase unwrapping device block diagram applied to ship-bridge body vibration monitoring is shown according to an exemplary embodiment, and the device is used for the phase unwrapping method applied to ship-bridge body vibration monitoring. Referring to Figure 2 The device comprises a signal acquisition module 210, a signal denoising module 220, a time-frequency feature extraction module 230, a spatial feature extraction module 240, a feature fusion module 250 and a signal unwrapping module 260. Wherein:

[0100] The signal acquisition module 210 is used for collecting monitoring information through a distributed optical fiber vibration sensing system, obtaining a phase signal sequence of bridge vibration, and segmenting the phase signal sequence based on a preset length sliding window to obtain a window phase signal sequence.

[0101] The signal denoising module 220 is used for adaptively suppressing noise of the window phase signal sequence based on a self-attention mechanism to obtain a denoised phase signal sequence.

[0102] The time-frequency feature extraction module 230 is configured to perform multi-scale time-frequency feature fusion on the denoised phase signal sequence through a convolutional neural network to obtain a space-time matrix and time-frequency features.

[0103] The spatial feature extraction module 240 is configured to perform spatial correlation feature extraction on the space-time matrix through a two-dimensional convolutional network to obtain spatial features.

[0104] The feature fusion module 250 is configured to perform weighted fusion on the time-frequency features and the spatial features based on a long short-term memory network to obtain space-time fusion features.

[0105] The signal unwrapping module 260 is configured to perform unwrapping phase mapping on the space-time fusion features through a fully connected layer to obtain a real phase signal sequence.

[0106] Optionally, the signal denoising module 220 is further configured to:

[0107] perform long-distance dependence capture on the window phase signal sequence to obtain a weighted value vector of the intermediate quantity of the window phase signal sequence;

[0108] perform calculation on the weighted value vector and a preset output weight matrix to obtain an attention output of the window phase signal sequence;

[0109] perform dynamic residual analysis on the window phase signal sequence based on the attention output to obtain an adaptive setting threshold of the statistical characteristics;

[0110] perform binary mask denoising processing on the window phase signal sequence according to the adaptive setting threshold and the attention output to obtain a denoised phase signal sequence.

[0111] Optionally, the time-frequency feature extraction module 230 is further configured to:

[0112] convert the denoised phase signal sequence into a space-time matrix according to a preset signal conversion parameter;

[0113] perform multi-scale feature extraction on the space-time matrix through a convolutional neural network to obtain multi-scale time-frequency local features;

[0114] perform channel fusion on the multi-scale time-frequency local features based on a preset fusion weight to obtain time-frequency features.

[0115] The convolutional neural network has a parallel relationship between convolution kernels, the number of the convolution kernels is 3, and the size of the convolution kernel is 3 3, 5 5 and 7 7.

[0116] The two-dimensional convolution network captures the spatial correlation of the sensor information in the space-time matrix by establishing a local correlation model in the spatial dimension.

[0117] Optionally, the feature fusion module 250 is further used for:

[0118] Based on the gating mechanism, the hidden state is obtained by capturing dynamic long-term dependency through a long short-term memory network according to the time-frequency feature.

[0119] Based on the preset attention mechanism weight, the spatial attention weight is calculated according to the spatial feature and the hidden state.

[0120] Based on the spatial attention weight, the time-space fusion feature is obtained by weighted fusion according to the time-frequency feature and the spatial feature.

[0121] The application proposes a phase unwrapping method applied to ship collision bridge beam vibration monitoring, which learns the internal physical law and noise distribution characteristics of the phase signal through large-scale training data of a deep learning model, has stronger robustness to complex interference such as random noise, environmental temperature fluctuation and multi-source vibration superposition, especially can still maintain high-precision unwrapping ability in a low signal-to-noise ratio environment, solves the phase jump and cumulative error problems caused by noise sensitivity of traditional algorithms, can realize real-time processing of massive phase data along the long-distance optical fiber, breaks through the calculation efficiency bottleneck caused by point-by-point expansion of traditional iterative algorithms, meets the needs of real-time alarm and rapid response of bridge safety monitoring, and through the self-adaptive parameter optimization mechanism of the deep learning model, can dynamically adapt to vibration signal characteristics of different frequency components, can accurately capture rapid phase changes of high-frequency vibration, and can effectively identify phase fluctuations of low-frequency micro-vibration, significantly improve the compatibility and overall monitoring performance of the system to multi-band vibration signals, and through the feature fusion capability of deep learning, the recognition capability to complex vibration modes can be further enhanced, high-reliability and high-resolution vibration perception data are provided for bridge structure health monitoring, the dependence of the system on hardware compensation means is significantly reduced, and the engineering deployment cost is reduced.

[0122] The application reconstructs the algorithm framework of phase unwrapping through deep learning technology, fundamentally solves the limitations of traditional methods in noise robustness, calculation efficiency, dynamic adaptability and global accuracy, and lays a technical foundation for the large-scale application of a distributed optical fiber acoustic vibration sensing system in bridge safety monitoring and other high-reliability scenes. The application is a phase unwrapping method for distributed optical fiber acoustic vibration, which has high accuracy and good robustness in a ship collision scene.

[0123] Figure 3 is a structural schematic view of a phase unwrapping device provided by an embodiment of the application, as Figure 3 indicated, the phase unwrapping device can include the aboveFigure 2 The phase unwrapping device shown is applied to ship collision bridge body vibration monitoring. Optionally, the phase unwrapping device 310 can include a first processor 2001.

[0124] Optionally, the phase unwrapping device 310 can also include a memory 2002 and a transceiver 2003.

[0125] The first processor 2001, the memory 2002, and the transceiver 2003 can be connected through a communication bus.

[0126] The specific components of the phase unwrapping device 310 will be described below. Figure 3 The specific components of the phase unwrapping device 310 will be described below.

[0127] The first processor 2001 is the control center of the phase unwrapping device 310 and can be one processor or a plurality of processing elements. For example, the first processor 2001 can be one or more central processing units (CPUs), application specific integrated circuits (ASICs), or one or more integrated circuits configured to perform the embodiments of the present application, such as one or more digital signal processors (DSPs) or one or more field programmable gate arrays (FPGAs).

[0128] Optionally, the first processor 2001 can execute various functions of the phase unwrapping device 310 by running or executing software programs stored in the memory 2002 and calling data stored in the memory 2002.

[0129] In a specific implementation, as an embodiment, the first processor 2001 can include one or more CPUs, such as the CPU0 and CPU1 shown in FIG. 1. Figure 3

[0130] In a specific implementation, as an embodiment, the phase unwrapping device 310 can also include a plurality of processors, such as the first processor 2001 and the second processor 2004 shown in FIG. 1. Each of these processors can be a single-CPU or a multi-CPU. The processor here can refer to one or more devices, circuits, and / or processing cores for processing data (such as computer program instructions). Figure 3

[0131] ​​The memory 2002 is configured to store a software program for implementing the scheme of the present application, and the first processor 2001 is configured to control the execution of the software program. The specific implementation can refer to the method embodiments described above, and will not be described here.

[0132] Alternatively, the memory 2002 can be a read-only memory (ROM) or other type of static storage device that can store static information and instructions, a random access memory (RAM) or other type of dynamic storage device that can store information and instructions, an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disk storage, a magneto-optical disk storage (including a compact disk, a laser disk, an optical disk, a digital versatile disk, a Blu-ray disk, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited to this. The memory 2002 can be integrated with the first processor 2001, or can exist independently and be coupled to the first processor 2001 through an interface circuit (not shown in the figure) of the phase unwrapping device 310. The embodiments of the present application are not limited in this regard. Figure 3

[0133] The transceiver 2003 is configured to communicate with a network device or a terminal device.

[0134] Alternatively, the transceiver 2003 can include a receiver and a transmitter (not shown separately in the figure). The receiver is configured to implement the receiving function, and the transmitter is configured to implement the transmitting function. Figure 3

[0135] Alternatively, the transceiver 2003 can be integrated with the first processor 2001, or can exist independently and be coupled to the first processor 2001 through an interface circuit (not shown in the figure) of the phase unwrapping device 310. The embodiments of the present application are not limited in this regard. Figure 3 It should be noted that the structure of the phase unwrapping device 310 shown in the figure does not constitute a limitation on the router, and the actual knowledge structure recognition device can include more or fewer components than those shown in the figure, or combine certain components, or different component arrangements.

[0136] Figure 3

[0137] ​​​​In addition, the technical effects of the phase unwrapping device 310 can refer to the technical effects of the phase unwrapping method applied to the ship collision bridge body vibration monitoring described in the above method embodiments, which will not be repeated here.

[0138] It should be understood that the first processor 2001 in the embodiments of the present application can be a central processing unit (CPU), and the processor can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0139] It should also be understood that the memory in the embodiments of the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically EPROM (EEPROM) or a flash memory. The volatile memory can be a random access memory (RAM) used as an external cache. By way of example but not limitation, many forms of random access memory (RAM) are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM) and direct memory bus random access memory (direct rambus RAM, DR RAM).

[0140] The above-described embodiments can be implemented in whole or in part by software, hardware (such as a circuit), firmware, or any combination thereof. When implemented in software, the above-described embodiments can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are wholly or partially generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions can be transferred from one website, computer, server, or data center to another website, computer, server, or data center through a wired (such as infrared, wireless, microwave, etc.) manner. The computer-readable storage medium can be any available medium accessible by a computer or a data storage device such as a server, data center, etc. containing one or more available medium collections. The available medium can be a magnetic medium (such as a floppy disk, a hard disk, a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state disk.

[0141] It should be understood that the term "and / or" herein merely describes an association relationship of associated objects, which means that there can be three relationships, for example, A and / or B can represent three cases of A alone, A and B together, and B alone, where A and B can be singular or plural. In addition, the character " / " herein generally represents an "or" relationship between the front and rear associated objects, but can also represent an "and / or" relationship, which can be understood in the context before and after.

[0142] In the present application, "at least one" means one or more, and "multiple" means two or more. "At least one of the following" or the like means any combination of the items, including any combination of single or multiple items. For example, at least one of a, b, or c can represent a, b, c, a-b, a-c, b-c, or a-b-c, where a, b, and c can be single or multiple.

[0143] It should be understood that in various embodiments of the present application, the size of the sequence number of the above-described processes does not mean the order of execution, and the execution order of the processes should be determined by their functions and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0144] Those skilled in the art can clearly understand that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0145] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working processes of the devices, apparatuses and units described above can refer to the corresponding processes in the foregoing method embodiments, which will not be repeated here.

[0146] In several embodiments provided by the present application, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely schematic, for example, the division of the units is only a logical function division, and actual implementation can have another division manner, for example, multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0147] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0148] In addition, each functional unit in each embodiment of the present application can be integrated into a processing unit, or each unit can exist physically independently, or two or more units can be integrated into one unit.

[0149] If the functions are realized in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the parts of the present application that essentially contribute to the prior art or the parts of the technical solutions can be embodied in the form of software products. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the various embodiments of the present application. The aforementioned storage medium includes a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0150] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A phase unwrapping method applied to ship collision bridge body vibration monitoring, characterized in that, The method includes: Monitoring information is collected through a distributed fiber optic vibration sensing system to obtain the phase signal sequence of bridge vibration; the phase signal sequence is then segmented based on a sliding window of preset length to obtain a windowed phase signal sequence. Based on the self-attention mechanism, adaptive noise suppression is performed on the window phase signal sequence to obtain a denoised phase signal sequence. The step of adaptively suppressing noise in the window phase signal sequence based on a self-attention mechanism to obtain a denoised phase signal sequence includes: Long-distance dependency capture is performed based on the window phase signal sequence to obtain the weighted vector of intermediate quantities in the window phase signal sequence; The attention output of the window phase signal sequence is obtained by calculating based on the weighted value vector and the preset output weight matrix; Based on attention output, dynamic residual analysis is performed on the window phase signal sequence to obtain an adaptive threshold for statistical characteristics; Based on the adaptively set threshold and attention output, the window phase signal sequence is subjected to binary mask denoising to obtain a denoised phase signal sequence. Based on the denoised phase signal sequence, multi-scale time-frequency feature fusion is performed through a convolutional neural network to obtain the spatiotemporal matrix and time-frequency features; The step of fusing multi-scale time-frequency features using a convolutional neural network based on the denoised phase signal sequence to obtain the spatiotemporal matrix and time-frequency features includes: Based on preset signal conversion parameters, the denoised phase signal sequence is converted into a spatiotemporal matrix; Based on the spatiotemporal matrix, multi-scale feature extraction is performed using a convolutional neural network to obtain multi-scale time-frequency local features. Based on preset fusion weights, channel fusion is performed according to multi-scale time-frequency local features to obtain time-frequency features; Based on the spatiotemporal matrix, spatial correlation features are extracted using a two-dimensional convolutional network to obtain spatial features; Based on long short-term memory networks, spatiotemporal fusion features are obtained by weighted fusion according to time-frequency features and spatial features. Based on the spatiotemporal fusion characteristics, the phase mapping is deconvoluted through a fully connected layer to obtain the real phase signal sequence.

2. The phase unwrapping method for ship collision bridge body vibration monitoring according to claim 1, characterized in that, The convolutional neural network has parallel convolutional kernels; the number of convolutional kernels is 3; and the size of the convolutional kernels is 3*3, 5*5, and 7*7.

3. The phase unwrapping method for ship collision bridge body vibration monitoring according to claim 1, characterized in that, The two-dimensional convolutional network captures the spatial correlation of sensor information in the spatiotemporal matrix by establishing a local correlation model in the spatial dimension.

4. The phase unwrapping method for ship collision bridge body vibration monitoring according to claim 1, characterized in that, The method based on a long short-term memory network, which performs weighted fusion based on time-frequency and spatial features to obtain spatiotemporal fusion features, includes: Based on the gating mechanism, and according to the time-frequency characteristics, the hidden state is obtained by dynamically capturing long-term dependencies through a long short-term memory network. Based on the preset attention mechanism weights, spatial attention weights are calculated according to spatial features and hidden states. Based on spatial attention weights, a weighted fusion is performed according to time-frequency features and spatial features to obtain spatiotemporal fusion features.

5. A phase unwrapping device for ship collision bridge body vibration monitoring, the phase unwrapping device for ship collision bridge body vibration monitoring being used to implement the phase unwrapping method for ship collision bridge body vibration monitoring according to any one of claims 1-4, characterized in that, The device includes: The signal acquisition module is used to collect monitoring information through a distributed fiber optic vibration sensing system to obtain the phase signal sequence of bridge vibration; and to segment the phase signal sequence based on a sliding window of preset length to obtain a windowed phase signal sequence. The signal denoising module is configured to perform adaptive noise suppression on the window phase signal sequence based on a self-attention mechanism to obtain a denoised phase signal sequence. The time-frequency feature extraction module is configured to perform multi-scale time-frequency feature fusion on the denoised phase signal sequence through a convolutional neural network to obtain a space-time matrix and time-frequency features. The spatial feature extraction module is configured to perform spatial correlation feature extraction on the space-time matrix through a two-dimensional convolutional network to obtain spatial features. The feature fusion module is configured to perform weighted fusion on the time-frequency features and the spatial features based on a long short-term memory network to obtain space-time fusion features. The signal unwrapping module is configured to perform unwrapping phase mapping on the space-time fusion features through a fully connected layer to obtain a real phase signal sequence.

6. The phase unwrapping device for ship collision bridge body vibration monitoring according to claim 5, characterized in that, The signal denoising module is further configured to: capture long-distance dependencies based on the window phase signal sequence to obtain a weighted value vector of an intermediate quantity of the window phase signal sequence; perform calculation based on the weighted value vector and a preset output weight matrix to obtain an attention output of the window phase signal sequence; perform dynamic residual analysis based on the attention output and the window phase signal sequence to obtain an adaptive setting threshold of a statistical characteristic; perform binary mask denoising processing on the window phase signal sequence based on the adaptive setting threshold and the attention output to obtain the denoised phase signal sequence.

7. A phase unwrapping apparatus characterized by comprising: The phase unwrapping device comprises: a processor; a memory having computer readable instructions stored thereon, wherein the computer readable instructions, when executed by the processor, implement the method of any one of claims 1 to 4.

8. A computer readable storage medium, characterized in that, The computer readable storage medium stores program code, which can be called and executed by the processor to implement the method of any one of claims 1 to 4.

Images