Method and device for separating ultrasonic aliasing signals and electronic equipment

Abstract

The invention provides an ultrasonic aliasing signal separation method and device and electronic equipment, and relates to the field of ultrasonic signal processing, and the method comprises the steps: obtaining a to-be-processed ultrasonic aliasing signal; the ultrasonic aliasing signal is a signal formed by aliasing a plurality of source ultrasonic signals; an encoder in the trained separation model is adopted to encode the ultrasonic aliasing signal, and a submerged space variable is obtained; and decoding the submerged space variable by adopting a decoder in the trained separation model to obtain a reconstructed ultrasonic aliasing signal. According to the method, the ultrasonic aliasing signals are processed through the variational auto-encoder unsupervised model, rapid recognition and dimension reduction processing of the ultrasonic aliasing signals are achieved, the ultrasonic aliasing signal separation problem is converted into the function optimization problem through the loss function formed by the minimum expected logarithm likelihood function and the KLD regularization calculation function, and the ultrasonic aliasing signal separation efficiency is improved. Separation of ultrasonic aliasing signals is achieved, and the quality and efficiency of ultrasonic signal recovery are further improved.

Description

Technical Field

[0001] The present application relates to the field of ultrasonic signal processing, and in particular to a method, device and electronic equipment for separating ultrasonic aliased signals. Background Art

[0002] Ultrasonic nondestructive testing has been widely used in aerospace, new energy, medical and other fields due to its non-destructive and safe characteristics. However, a practical challenge of ultrasonic nondestructive testing is that the ultrasonic transducer may receive multiple reflected signals from different interfaces or multiple internal defects at the same time during the receiving process. The aliasing of these signals makes data analysis and defect diagnosis more complicated. Therefore, it is necessary to effectively identify and separate ultrasonic aliased signals. In ultrasonic signal processing, one of the commonly used methods is empirical mode decomposition, which first decomposes the aliased signal into a series of intrinsic mode functions. Secondly, by calculating the correlation coefficient, areas with different signal strengths can be identified and distinguished. Finally, wavelet transform combined with threshold processing technology is usually used to separate these signals. Although this conventional method shows good results in distinguishing areas with large differences in signal strength, its separation effect still needs to be further improved when processing complex aliased signals with similar frequencies. Summary of the invention

[0003] The purpose of the embodiments of the present application is to provide a method, device and electronic equipment for separating ultrasonic aliasing signals, so as to solve the above-mentioned problems existing in the prior art and to fully separate ultrasonic aliasing signals.

[0004] In a first aspect, a method for separating an ultrasonic aliasing signal is provided, and the method may include:

[0005] Acquire an ultrasonic aliasing signal to be processed; the ultrasonic aliasing signal is a signal aliased by multiple source ultrasonic signals;

[0006] encoding the ultrasonic aliasing signal using an encoder in the trained separation model to obtain a latent space variable;

[0007] The latent space variables are decoded using a decoder in the trained separation model to obtain a reconstructed ultrasonic aliasing signal.

[0008] In a possible implementation, the training process of the separation model includes:

[0009] The preset generation model is trained through the configured training set to obtain the initial separation model;

[0010] The parameters of the initial separation model are optimized by using the determined loss function to determine the optimized parameters, and the separation model is determined based on the optimized parameters.

[0011] In a possible implementation, the process of determining the loss function includes:

[0012] After the encoder of the initial separation model outputs the initial latent space variables, determining a first Gaussian distribution satisfied by the initial latent space variables;

[0013] and determining a second Gaussian distribution satisfied by the initial latent space variables and a posterior probability distribution satisfied by the reconstructed signal output by the decoder;

[0014] The loss function is determined according to the first Gaussian distribution, the second Gaussian distribution and the posterior probability distribution.

[0015] In a possible implementation, the loss function includes a reconstruction function and a KL divergence function.

[0016] In a possible implementation, the configuration process of the training set includes:

[0017] The toneburst signal is processed with different time shift changes and amplitude changes to obtain multiple change signals;

[0018] The multiple change signals are randomly combined to determine the training set including multiple training samples and the test set including multiple test samples.

[0019] In a second aspect, a device for separating ultrasonic aliasing signals is provided, and the device may include:

[0020] An acquisition unit, used for acquiring an ultrasonic aliasing signal to be processed; the ultrasonic aliasing signal is a signal aliased by multiple source ultrasonic signals;

[0021] An encoding unit, configured to encode the ultrasonic aliasing signal using an encoder in a trained separation model to obtain a latent space variable;

[0022] The decoding unit is used to decode the latent space variables using a decoder in the trained separation model to obtain a reconstructed ultrasonic aliasing signal.

[0023] In a possible implementation, the device further includes a training unit;

[0024] The training unit is further used to train the configured generation model through the configured training set to obtain an initial separation model;

[0025] The parameters of the initial separation model are optimized by using the determined loss function, the optimized parameters are determined, and the separation model is determined based on the optimized parameters.

[0026] In a third aspect, an electronic device is provided, the electronic device comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus;

[0027] Memory, used to store computer programs;

[0028] The processor is used to implement any method step described in the first aspect when executing the program stored in the memory.

[0029] The present application provides a method for separating ultrasonic aliasing signals, the method comprising obtaining an ultrasonic aliasing signal to be processed; the ultrasonic aliasing signal is a signal aliased by multiple source ultrasonic signals; the ultrasonic aliasing signal is encoded by an encoder in a trained separation model to obtain a latent space variable; the latent space variable is decoded by a decoder in the trained separation model to obtain a reconstructed ultrasonic aliasing signal. The present application processes ultrasonic aliasing signals through an unsupervised model of a variational autoencoder to achieve rapid recognition and dimensionality reduction of ultrasonic aliasing signals, and transforms the ultrasonic aliasing signal separation problem into a function optimization problem by minimizing the loss function composed of an expected log-likelihood function and a KLD regularization calculation function, thereby achieving separation of ultrasonic aliasing signals and further improving the quality and efficiency of ultrasonic signal recovery. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments of the present application will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0031] Figure 1 A system architecture diagram of a method for separating ultrasonic aliasing signals provided in an embodiment of the present application;

[0032] Figure 2 A schematic flow chart of a method for separating ultrasonic aliasing signals provided in an embodiment of the present application;

[0033] Figure 3 A schematic diagram of a separation process of an ultrasonic aliasing signal by a separation model provided in an embodiment of the present application;

[0034] Figure 4 A schematic diagram for comparing the reconstructed source ultrasonic signal 1 and the source ultrasonic signal 1 provided in an embodiment of the present application;

[0035] Figure 5 A schematic diagram for comparing the reconstructed source ultrasonic signal 2 and the source ultrasonic signal 2 provided in an embodiment of the present application;

[0036] Figure 6 A schematic diagram for comparing the reconstructed ultrasonic aliasing signal and the original ultrasonic aliasing signal provided in an embodiment of the present application;

[0037] Figure 7 A schematic diagram of the structure of a device for separating ultrasonic aliasing signals provided in an embodiment of the present application;

[0038] Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0039] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0040] The ultrasonic aliasing signal separation method provided in the embodiment of the present application can be applied in Figure 1 In the system architecture shown in Figure 1 As shown, the system may include: a server and a terminal. The server may be a physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDN), and basic cloud computing services such as big data and artificial intelligence platforms. The terminal may be a user equipment (UE) such as a mobile phone, smart phone, laptop, digital broadcast receiver, personal digital assistant (PDA), tablet computer (PAD), handheld device, vehicle-mounted device, wearable device, computing device or other processing device connected to a wireless modem, mobile station (MS), mobile terminal, etc. The terminal and the server may be directly or indirectly connected via wired or wireless communication, which is not limited in this application.

[0041] The terminal is used to obtain the ultrasonic aliasing signal to be processed and send the ultrasonic aliasing signal to be processed to the server.

[0042] The server is used to obtain an ultrasonic aliasing signal to be processed to execute a method for separating ultrasonic aliasing signals provided in the present application.

[0043] Ultrasonic nondestructive testing has been widely used in aerospace, new energy, medical and other fields due to its non-destructive and safe characteristics. However, a practical challenge of ultrasonic nondestructive testing is that the ultrasonic transducer may receive multiple reflected signals from different interfaces or multiple internal defects at the same time during the receiving process. The aliasing of these signals makes data analysis and defect diagnosis more complicated. Therefore, it is necessary to effectively identify and separate ultrasonic aliased signals. In ultrasonic signal processing, one of the commonly used methods is empirical mode decomposition, which first decomposes the aliased signal into a series of intrinsic mode functions. Secondly, by calculating the correlation coefficient, areas with different signal strengths can be identified and distinguished. Finally, wavelet transform combined with threshold processing technology is usually used to separate these signals. Although this conventional method shows good results in distinguishing areas with large differences in signal strength, its separation effect still needs to be further improved when processing complex aliased signals with similar frequencies.

[0044] Therefore, the present application provides a method for separating ultrasonic aliasing signals, which is used to solve the above-mentioned problems existing in the prior art and can fully separate ultrasonic aliasing signals.

[0045] The preferred embodiments of the present application are described below in conjunction with the drawings in the specification. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present application, and are not used to limit the present application. In addition, the embodiments and features in the embodiments of the present application may be combined with each other if there is no conflict.

[0046] Figure 2 The flowchart of a method for separating ultrasonic aliasing signals provided in an embodiment of the present application is shown in FIG. Figure 2 As shown, the method may include:

[0047] Step S210: construct a training set and a test set.

[0048] The test set includes ultrasonic aliasing signals and change signals to be processed.

[0049] Specifically, the toneburst signal is processed with different time shifts and amplitude changes to obtain multiple change signals. In short, the change signal is a signal in which the toneburst signal (which can be understood as the original ultrasonic signal) is translated along the time axis for a certain time interval without changing its amplitude and frequency characteristics and / or a signal aliasing in which the amplitude of the toneburst signal changes with time or space.

[0050] The multiple change signals are randomly combined multiple times, and 80% of the training samples in the change signals are taken to form a training set, and the rest are taken as test samples and form a test set together with the ultrasonic signal to be processed.

[0051] The ultrasonic signal to be processed is a composite signal formed by aliasing of multiple ultrasonic signals during the transducer receiving process;

[0052] The aliasing of ultrasonic signals refers to the fact that when ultrasonic waves encounter different interfaces or multiple defect areas in the object being tested, they will cause multiple wave behaviors such as reflection, refraction and scattering. These waves cause ultrasonic waves to propagate along different paths. If the ultrasonic waves on these different paths return to the transducer almost at the same time or in a close time, signal superposition may occur, thus forming ultrasonic aliasing signals. The existence of such aliasing signals increases the complexity of signal analysis.

[0053] Step S220: input the training set into the variational autoencoder to obtain a trained separation model.

[0054] The preset generation model is trained through the configured training set to obtain the initial separation model;

[0055] The parameters of the initial separation model are optimized through the determined loss function, the optimized parameters are determined, and based on the optimized parameters, a separation model that can be actually applied is determined.

[0056] Among them, the preset generation model includes an encoder and a decoder, each of which is composed of five fully connected feedforward neural network layers, each of which contains a linear layer, a ReLU activation function and a batch normalization processing function.

[0057] Table 1 shows the types and dimensions of the dataset in an experimental example (D x ) and the dimensions of the fully connected layers in each layer of the encoder's feed-forward network, and the corresponding dimensions of the fully connected layers of the decoder increase layer by layer.

[0058] Table 1

[0059]

[0060] The process of determining the loss function can specifically include:

[0061] Assume that any training sample x in the training set consists of M source signals s; (noise-free), through the encoder (q φ , parameter φ) maps the input training samples to the low-dimensional latent space, thereby determining the probability distribution of the initial latent space variable Z; where K is the number of source signals corresponding to the training samples. It should be noted that M is the number of source signals that actually constitute the aliasing signal (M may be unknown), and K is the number of source signals artificially set in the model to infer the composition of the aliasing signal (satisfying K≥M), and K is used in the calculation process.

[0062] Typically, the encoder consists of two branches, one for outputting the mean μ and the other for outputting the variance σ 2 The branch outputs here are all obtained based on the latent space variable Z, which is used for comparison calculation with the first Gaussian distribution at the loss function.

[0063] Afterwards, through the decoder (g θ , parameter θ) decodes the initial latent space variables into the reconstructed signal Specifically, it can be expressed as: the corresponding training sample is composed of K source signals s;

[0064] Under the condition of a given input x, the approximate posterior distribution of the initial latent space variable Z satisfies the independent first Gaussian distribution, which can be specifically expressed as:

[0065]

[0066] Among them, μ φ (x) and is the matrix of the mean and variance of the encoder output, and N represents a Gaussian distribution.

[0067] And, assuming that the initial latent space variable Z satisfies an independent prior Gaussian distribution (second Gaussian distribution), it can be specifically expressed as:

[0068]

[0069] The prior Gaussian distribution in this method assumes that each element of each initial latent space variable changes independently, which helps to separate the changing factors in the data. It should be noted that the assumptions about the variable distribution are all prerequisites. The purpose is to compare the distribution obtained by the model with the assumed distribution when calculating the loss function, so that the output of the model is close to the assumed distribution.

[0070] The decoder decodes the initial latent space variables into the reconstructed signal Afterwards, given the latent space variable Z, the data X satisfies the posterior probability distribution, which is the Laplace distribution with variance b (variance b is the set distribution parameter, which is different from the previous latent space output branch):

[0071]

[0072] According to the first Gaussian distribution, the second Gaussian distribution and the posterior probability distribution, the loss function is determined:

[0073]

[0074] Furthermore,

[0075] in, Indicates that the training sample X contains N independent and identically distributed samples x (n) .

[0077] The above loss function includes two parts:

[0078] Part I Indicates that in x (n) In the initial latent space variable Z and the approximate posterior distribution q φ Calculate p under the condition (z|x) θ (x (n) |Z), in short, the first part is the reconstruction function, which is used to measure the difference between the initial latent space variables output by the decoder and the input training samples.

[0079] Part II.D KL (q φ (Z|x (n) )||p(Z)) represents the KL divergence regularization calculation function between the posterior probability distribution and the prior Gaussian distribution, where the KL divergence is Kullback-Leibler Divergence, abbreviated as KLD. In short, the second part is the KL divergence function, which is used to measure the difference between the posterior probability distribution output by the encoder and the prior Gaussian distribution to ensure that the initial latent space variable distribution is close to the prior Gaussian distribution.

[0080] Then, by minimizing the loss function Optimize the parameters of the encoder and decoder, determine the optimized parameters of the encoder and decoder, and determine a separation model that can be actually applied. It can be understood that the separation model is a variational autoencoder.

[0081] Step S230: input the test set into the trained separation model to separate the ultrasonic aliasing signal to be processed.

[0082] Specifically, the separation model after optimizing the parameters is tested through the test set to determine the generalization performance of the separation model after optimizing the parameters, that is, the test set is processed through the separation model after optimizing the parameters to determine whether it can separate the reconstructed signal that is infinitely close to the test set, so as to determine whether the model can be used in practice; if it can, it is determined to be the trained separation model. If not, the training set and test set are reconstructed or the model hyperparameters are adjusted to continue training the current model until it can be determined through the test set that the model can be used in practice.

[0083] Combination Figure 3 As shown, the encoder in the trained separation model is used to encode the ultrasonic aliasing signal to be processed to obtain multiple latent space variables of the latent space;

[0084] The decoder in the trained separation model is used to decode and reconstruct each latent space variable to obtain a reconstructed ultrasonic aliasing signal, that is, the separation result of the separation model on the ultrasonic signal. Figure 3 It means that multiple aliased input signals are processed and separated signals corresponding to each aliased signal are output.

[0085] The output of the separation model is a reconstructed ultrasound aliased signal that attempts to approximate the original signal but is not identical to the original signal. This is because the separation model also takes into account the distribution characteristics of the latent space, in this way it can be used not only to reconstruct known input signals but also to generate new, similar signals.

[0086] In some embodiments, in combination Figures 4 to 6 As shown, the comparison result of the reconstructed source ultrasonic signal (reconstructed source signal) 1 and the source ultrasonic signal (source signal) 1. The comparison result of the reconstructed source ultrasonic signal 2 and the source ultrasonic signal 2. The comparison result of the reconstructed ultrasonic aliasing signal and the original ultrasonic aliasing signal.

[0087] in, Figure 6 The solid line is a randomly selected aliased signal in the test set, which consists of two source signals, denoted as source signal 1 ( Figure 4 solid line) and source signal 2 ( Figure 5 Solid line). Figure 6 After the aliased signal in is input into the trained model (the setting value of K in the model is 2), the reconstructed source signal 1 ( Figure 4 dotted line), reconstructed source signal 2 ( Figure 5 dotted line) and the composite aliased reconstructed signal ( Figure 6 dotted line).

[0088] The present application provides a method for separating ultrasonic aliasing signals, the method comprising obtaining an ultrasonic aliasing signal to be processed; the ultrasonic aliasing signal is a signal aliased by multiple source ultrasonic signals; the ultrasonic aliasing signal is encoded by an encoder in a trained separation model to obtain a latent space variable; the latent space variable is decoded by a decoder in the trained separation model to obtain a reconstructed ultrasonic aliasing signal. The present application processes ultrasonic aliasing signals through an unsupervised model of a variational autoencoder to achieve rapid recognition and dimensionality reduction of ultrasonic aliasing signals, and transforms the ultrasonic aliasing signal separation problem into a function optimization problem by minimizing the loss function composed of an expected log-likelihood function and a KLD regularization calculation function, thereby achieving separation of ultrasonic aliasing signals and further improving the quality and efficiency of ultrasonic signal recovery.

[0089] The method proposed in this application is a technique for separating ultrasonic aliasing signals through a variational autoencoder. Variational reasoning is a technique that uses approximate Bayesian inference to solve complex problems that are difficult to calculate in machine learning. It approximates complex distributions by selecting a parameterized distribution family and obtains approximate posterior distributions by solving specific optimization problems, which can effectively handle inference problems of high-dimensional data and complex models. Therefore, the technique for separating ultrasonic aliasing signals based on a variational autoencoder can achieve comprehensive separation of ultrasonic aliasing signals.

[0090] Corresponding to the above method, the embodiment of the present application also provides a separation device for ultrasonic aliasing signals, such as Figure 7 As shown, the device comprises:

[0091] An acquisition unit 710 is used to acquire an ultrasonic aliasing signal to be processed; the ultrasonic aliasing signal is a signal aliased by multiple source ultrasonic signals;

[0092] An encoding unit 720 is used to encode the ultrasonic aliasing signal using an encoder in the trained separation model to obtain a latent space variable;

[0093] The decoding unit 730 is used to decode the latent space variables using a decoder in the trained separation model to obtain a reconstructed ultrasonic aliasing signal.

[0094] The functions of each functional unit of the ultrasonic aliasing signal separation device provided in the above-mentioned embodiment of the present application can be realized through the above-mentioned method steps. Therefore, the specific working process and beneficial effects of each unit in the ultrasonic aliasing signal separation device provided in the embodiment of the present application are not repeated here.

[0095] The present application also provides an electronic device, such as Figure 8 As shown, it includes a processor 810 , a communication interface 820 , a memory 830 and a communication bus 840 , wherein the processor 810 , the communication interface 820 , and the memory 830 communicate with each other via the communication bus 840 .

[0096] Memory 830, used for storing computer programs;

[0097] The processor 810 is used to execute the program stored in the memory 830 to implement the following steps:

[0098] Acquire an ultrasonic aliasing signal to be processed; the ultrasonic aliasing signal is a signal aliased by multiple source ultrasonic signals;

[0099] encoding the ultrasonic aliasing signal using an encoder in the trained separation model to obtain a latent space variable;

[0100] The latent space variables are decoded using a decoder in the trained separation model to obtain a reconstructed ultrasonic aliasing signal.

[0101] The communication bus mentioned above can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0102] The communication interface is used for communication between the above electronic device and other devices.

[0103] The memory may include a random access memory (RAM) or a non-volatile memory (NVM), such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0104] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0105] The implementation methods and beneficial effects of the components of the electronic device in the above embodiments to solve the problems can be seen in Figure 2 The various steps in the illustrated embodiment are implemented, therefore, the specific working process and beneficial effects of the electronic device provided by the embodiment of the present application are not repeated here.

[0106] In another embodiment provided by the present application, a computer program product including instructions is also provided. When the computer program product is executed on a computer, the computer is enabled to execute a method for separating ultrasonic aliasing signals as described in any one of the above embodiments.

[0107] Those skilled in the art should understand that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the embodiments of the present application can be in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware.

[0108] The embodiments of the present application are described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0109] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0110] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0111] Unless otherwise defined, the technical terms or scientific terms used in this application should be understood by people with ordinary skills in the field to which the present invention belongs. "First", "second" and similar words used in this application do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect", "couple" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0112] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they are aware of the basic creative concepts. Therefore, the present application embodiments are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present application embodiments.

[0113] Obviously, those skilled in the art can make various changes and modifications to the embodiments in the present application without departing from the spirit and scope of the embodiments in the present application. Thus, if these modifications and variations of the embodiments in the present application are within the scope of the embodiments in the present application and their equivalents, the embodiments in the present application are also intended to include these modifications and variations.

Claims

1. A method for separating ultrasonic aliasing signals, characterized in that: The method comprises: Acquire an ultrasonic aliasing signal to be processed; the ultrasonic aliasing signal is a signal aliased by multiple source ultrasonic signals; encoding the ultrasonic aliasing signal using an encoder in the trained separation model to obtain a latent space variable; The latent space variables are decoded using a decoder in the trained separation model to obtain a reconstructed ultrasonic aliasing signal.

2. The method according to claim 1, characterized in that The training process of the separation model includes: The preset generation model is trained through the configured training set to obtain the initial separation model; The parameters of the initial separation model are optimized by using the determined loss function to determine the optimized parameters, and the separation model is determined based on the optimized parameters.

3. The method according to claim 2, characterized in that The process of determining the loss function includes: After the encoder of the initial separation model outputs the initial latent space variables, determining a first Gaussian distribution satisfied by the initial latent space variables; and determining a second Gaussian distribution satisfied by the initial latent space variables and a posterior probability distribution satisfied by the reconstructed signal output by the decoder; The loss function is determined according to the first Gaussian distribution, the second Gaussian distribution and the posterior probability distribution.

4. The method according to claim 3, characterized in that The loss function includes a reconstruction function and a KL divergence function.

5. The method according to claim 4, characterized in that The configuration process of the training set includes: The toneburst signal is processed with different time shift changes and amplitude changes to obtain multiple change signals; The toneburst signal and a plurality of variation signals are randomly combined to determine the training set including a plurality of training samples and the test set including a plurality of test samples.

6. A device for separating ultrasonic aliased signals, characterized in that: The device comprises: An acquisition unit, used for acquiring an ultrasonic aliasing signal to be processed; the ultrasonic aliasing signal is a signal aliased by multiple source ultrasonic signals; An encoding unit, configured to encode the ultrasonic aliasing signal using an encoder in a trained separation model to obtain a latent space variable; The decoding unit is used to decode the latent space variables using a decoder in the trained separation model to obtain a reconstructed ultrasonic aliasing signal.

7. The device according to claim 6, characterized in that The device also includes a training unit; The training unit is further used to train the configured generation model through the configured training set to obtain an initial separation model; The parameters of the initial separation model are optimized by using the determined loss function, the optimized parameters are determined, and the separation model is determined based on the optimized parameters.

8. An electronic device, characterized in that: The electronic device comprises a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus; Memory, used to store computer programs; A processor, for implementing the method steps described in any one of claims 1 to 5 when executing a program stored in a memory.

Images