Doppler blood flow velocity estimation method based on high-frequency pulse differential frequency

Abstract

The invention discloses a Doppler blood flow velocity estimation method based on high-frequency pulse differential frequency, and solves the problems of velocity estimation resolution reduction and artifacts caused by alternate transmission of a B mode and Doppler pulses in an ultrasonic dual-mode system. The method comprises the following steps: transmitting a high-frequency broadband ultrasonic signal compatible with B-mode imaging; the method comprises the following steps: receiving a superficial blood vessel echo signal, performing time domain sampling, converting into a discrete signal, extracting a narrow-band frequency spectrum through frequency sampling, and generating a low-frequency equivalent signal by using a multi-frequency-point differential frequency technology; calculating a covariance matrix of each signal, performing average weighting to remove artifacts, and generating low-frequency equivalent covariance data; and estimating the blood flow velocity through eigenvalue decomposition and spectral peak search. According to the method, two pulses do not need to be transmitted alternately, resolution reduction and artifacts are avoided, Doppler velocity estimation precision is remarkably improved, integrated waveform design is supported, and hardware implementation is simplified.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of ultrasonic detection, and particularly relates to a Doppler blood flow velocity estimation method based on high-frequency pulse differential frequency. BACKGROUND

[0002] Ultrasound imaging technology uses the echo signal of the interaction between high-frequency ultrasound waves and biological tissues to realize non-invasive medical diagnosis and is widely used in the fields of cardiovascular and fetal monitoring. It is favored due to its non-radiation, real-time and portability. Ultrasound imaging mainly includes B mode and Doppler mode. The B mode generates a two-dimensional gray-scale image through echo amplitude, adopts a high-frequency broadband signal to obtain a high-resolution image, and is suitable for blood vessel and tissue anatomical structure imaging. The Doppler mode estimates blood flow velocity based on the frequency shift of the echo and is suitable for blood flow dynamics evaluation.

[0003] In the prior art, a dual-mode imaging system combines B mode and Doppler mode to simultaneously provide anatomical images and blood flow velocity information by alternately transmitting two kinds of pulses. However, the alternately transmitting leads to time resource segmentation, reduces the pulse repetition frequency (PRF), limits the resolution of Doppler velocity estimation and introduces estimation artifacts, especially when the blood flow is fast. In addition, when a high-frequency signal is directly used for Doppler estimation, the frequency resolution is insufficient due to the broadband characteristic, it is difficult to accurately detect a small frequency shift, and the signal is severely attenuated in deep tissues, which limits the application range. Although the existing improved methods such as coded pulse and adaptive filtering can partially improve the signal-to-noise ratio, they still need to alternately transmit, and cannot fundamentally solve the problems of artifacts and resolution.

[0004] Therefore, there is an urgent need for a method that does not need to alternately transmit pulses, takes into account B mode imaging and Doppler velocity estimation, eliminates artifacts, improves velocity estimation accuracy and simplifies system design. SUMMARY

[0005] The main purpose of the application is to solve the problems of artifacts and velocity estimation resolution reduction caused by alternately transmitting B mode and Doppler pulses in an ultrasonic dual-mode system, and the defect of insufficient frequency resolution when a high-frequency broadband signal is directly used for Doppler estimation, and to provide a Doppler blood flow velocity estimation method based on high-frequency pulse differential frequency.

[0006] In order to achieve the above purpose, the application adopts the following technical scheme: A Doppler blood flow velocity estimation method based on high-frequency pulse differential frequency, the Doppler blood flow velocity estimation method comprising the following steps: S1, transmitting a high-frequency broadband ultrasonic signal compatible with B mode imaging to a superficial blood vessel, wherein the B mode generates a two-dimensional gray-scale image through echo amplitude, and a high-frequency broadband signal is adopted to obtain a high-resolution ultrasonic imaging image; S2, receiving echo signals of superficial blood vessels, converting into discrete signals by time domain sampling; S3, extracting narrowband spectrum of echo signals by frequency sampling, obtaining multiple frequency components, and applying Hilbert transform to convert into complex signals; S4, performing differential operation on a pair of frequency components in the narrowband spectrum of the echo signals that satisfy a fixed differential frequency, calculating the covariance matrix of the differential signals, repeating the operation on all signals that satisfy the differential frequency, calculating the mean of the covariance matrix, and generating an equivalent low-frequency covariance matrix; S5, dividing the equivalent low-frequency covariance matrix into subarrays and calculating the spatially smoothed covariance matrix, then performing eigenvalue decomposition, dividing the subspace, performing spectral peak search, and applying spectral estimation algorithm to estimate blood flow velocity.

[0007] Further, the high-frequency wideband ultrasonic signal is a Gaussian modulated pulse sequence , and the expression is as follows:

[0008] wherein, represents a continuous time variable, is a slow time sampling index, recording the pulse sequence and capturing the phase change caused by blood flow velocity, represents the total number of pulse sequences, represents the pulse repetition period, is a Gaussian envelope function, is a cosine function.

[0009] In ultrasonic blood flow detection, low-frequency signals are used for blood flow velocity detection, and high-frequency signals are used for B-mode imaging. The present application transmits high-frequency signals and reduces the working frequency by differential frequency to balance B-mode imaging and blood flow detection. Wideband signals are transmitted to facilitate subsequent frequency sampling and eliminate the influence of differential frequency.

[0010] Further, the echo signals received in step S2 are delayed and attenuated versions of the transmitted signals scattered by blood cells, which are sampled by an analog-to-digital converter at a sampling rate to obtain discrete signals The expression is as follows:

[0011] wherein, is a fast time sampling index, capturing the time domain shape of the signal and reflecting the information in the depth direction, is a slow time sampling index, recording the pulse sequence and capturing the phase change caused by blood flow velocity, is the reflection coefficient, is a Gaussian envelope function, is the carrier center frequency, denotes the pulse repetition period, denotes the distance of the high-frequency broadband ultrasound signal transmitter from the center of the blood vessel, is the propagation speed of ultrasound in the blood flow, denotes the target radial movement speed, in m / s, is the cosine function.

[0012] Further, step S3 performs frequency sampling on the discrete signal by a set of band-pass filters. The multiple narrow-band components obtained by frequency sampling facilitate subsequent elimination of the negative effects of the differential frequency method.

[0013] extracts narrow-band frequency components, the center frequency of the th narrow-band component is :

[0014] wherein, denotes the frequency interval, denotes the starting frequency, denotes the narrow-band frequency component index, denotes the number of narrow-band frequency components, and the linear increase of the designed center frequencies at intervals of can ensure more frequency data pairs, facilitating subsequent differential operations.

[0015] applies the Hilbert transform to the th narrow-band component, converts the signal into a complex exponential signal that is easier to process, and obtains the complex signal :

[0016] wherein, is the reflection coefficient for the th narrow-band component, denotes the imaginary symbol, and after demodulation of the signal and removal of irrelevant terms, the matrix expression of can be written as:

[0017] wherein, denotes the slow-time vector, is the serial number of the blood flow scatterer, is the total number of blood flow scatterers, and the direction vector = , denotes the reflection coefficient of the th scatterer for the th narrow-band component, denotes the reflection coefficient of the a Gaussian white noise with a narrowband component added, is the first radial moving speed of the blood flow scatterer. By derivation, the expression has partially met the conditions of estimation, but since its working frequency is , the frequency is too high to cause problems such as low estimation frequency resolution, under-sampling, etc., so the working frequency needs to be reduced.

[0018] Further, the difference operation in step S4 reduces the working frequency, specifically including: selecting any frequency in the narrowband component that meets the condition, where the is a low frequency suitable for blood flow velocity estimation, and the received data corresponding to the frequency is written as a data pair , , , , The calculated difference frequency data is obtained by using Hadamard product:

[0019] wherein represents multiplication of corresponding elements of a matrix, represents taking complex conjugate of data ; since , are complex exponential signals, the complex conjugate multiplication can obtain the difference frequency data, and the expansion is:

[0020] wherein is the difference frequency term, and since the signals operated by each pair of difference frequency meet , this term constructs an equivalent received signal term of , i.e.

[0021] The second cross-term of the expansion will cause artifacts in the estimation result, and the artifact position is determined by the target velocity to be estimated and the transmission frequency.

[0022] The covariance matrix is calculated for each difference signal: wherein represents the total number of pulse sequences, represents the difference frequency data ​​Take the conjugate transpose. right Decompose:

[0023] in This represents the covariance matrix of the low-frequency equivalent signal, while It is the pseudo-image covariance matrix. This is the noise covariance matrix. To suppress artifact terms and reduce the impact of noise, for all... The average of the covariance matrices of the difference frequency signals is taken: Because the signal covariance is averaged, the low-frequency... and parameters to be estimated Consistent, enhanced after averaging. However, artifact terms vary across different... The values ​​differ slightly, therefore, after averaging, they gradually approach 0 when the frequency sampling number... big enough to satisfy Therefore, the signal term is preserved while artifacts are suppressed. That is, equivalent low frequency Receive signals.

[0024] Furthermore, the spectral estimation algorithm in step S5 employs a spatial smoothing MUSIC algorithm. Since blood flow particles are homogeneous scattering particles with strong coherence, decoherence can improve estimation accuracy. The pulses set in this invention have a continuous arrangement, allowing for the application of a low-complexity spatial smoothing decoherence method, specifically including: The equivalent low-frequency covariance matrix R is divided into... The overlapping subarrays are calculated, and the spatially smoothed covariance matrix is ​​computed. ,in Indicates the first The covariance matrix of each subarray; Again Perform eigenvalue decomposition to divide the signal subspace and noise subspace; construct direction vectors: Using the peak search function :

[0025] Estimate blood flow velocity v̂, where The direction vector representing the differential frequency. Represents the direction vector Take the conjugate transpose, and Represents the noise subspace.

[0026] Furthermore, the bandpass filter is an ideal filter, and its bandwidth is set to... ,in Bandwidth preference factor to ensure sufficient frequency spacing between narrowband components to allow for smooth filtering of a sufficient number of narrowband frequency components.

[0027] Compared with the prior art, the present application has the following advantages and beneficial effects: (1) The Doppler blood flow velocity estimation method proposed by the present application does not need to alternately transmit B mode and Doppler pulses, thereby avoiding the problem of sparse transmission caused by time resource segmentation in the traditional dual-mode system. In the traditional method, B mode short pulses and high-resolution broadband signals are used to generate two-dimensional gray-scale images, while Doppler mode longer pulses are used for narrowband frequency resolution, resulting in a decrease in pulse repetition frequency (PRF) and signal discontinuity, which in turn introduces artifacts and a decrease in velocity resolution. The present application realizes integrated waveform design by transmitting a single Gaussian-modulated broadband signal, combined with frequency sampling to extract narrowband components and difference frequency operation to construct an equivalent low-frequency signal. This not only eliminates artifacts caused by sparse transmission, but also maintains a stable PRF, improves the robustness of Doppler frequency shift estimation, is suitable for dynamic blood flow scenarios, and improves the reliability and efficiency of clinical diagnosis.

[0028] (2) The Doppler blood flow velocity estimation method proposed by the present application overcomes the problem of insufficient frequency resolution of high-frequency broadband signals through frequency sampling and difference frequency technology, improves the accuracy of Doppler velocity estimation, and is particularly suitable for superficial blood vessels such as the carotid artery. In high-frequency broadband signals, the traditional method causes the frequency spectrum to be wide (3-5 MHz) due to the short pulse length (0.1-0.2 µs), resulting in low frequency resolution and an inability to accurately capture small Doppler frequency shifts. The present application first generates a discrete signal by time-domain sampling , then applies a bandpass filter to extract narrowband components . Subsequently, the difference frequency technology is used to construct a low-frequency equivalent signal by Hadamard product of the data . This technology suppresses artifacts and preserves Doppler information by averaging the covariance matrix to improve velocity resolution, is suitable for superficial blood vessels, has low high-frequency attenuation, and has high signal-to-noise ratio, and can realize early atherosclerosis detection in clinical practice.

[0029] (3) The present application supports integrated waveform design, simplifies the hardware implementation of the ultrasound system, and improves real-time performance and practicality. In the traditional system, B mode and Doppler mode need independent hardware modules to support different pulse designs, resulting in system complexity and poor real-time performance. The present application uses a single Gaussian-modulated broadband signal transmission, compatible with B-mode high-resolution imaging and Doppler estimation, narrowband components are extracted by time-domain sampling and frequency sampling, low-frequency equivalent matrix is generated by difference frequency operation and covariance average, and finally the velocity is estimated by MUSIC algorithm. This design reduces the number of transmitting modules, reduces the hardware complexity (such as no need to switch pulse types), improves the real-time performance, is suitable for portable ultrasonic equipment, and improves the practicability in clinical applications, such as real-time carotid blood flow monitoring and cardiovascular diagnosis. BRIEF DESCRIPTION OF DRAWINGS

[0030] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0031] Figure 1 is a flow chart of the Doppler blood flow velocity estimation method based on high-frequency pulse difference frequency disclosed by the present application; Figure 2 is a blood flow velocity spectrum diagram for directly using high-frequency pulses to estimate Doppler blood flow velocity in the embodiments of the present application; Figure 3 is a blood flow velocity estimation spectrum diagram for using high-frequency pulse difference frequency to estimate Doppler blood flow velocity in the embodiments of the present application; Figure 4 is a blood flow velocity estimation spectrum diagram for using low-frequency pulses to estimate Doppler blood flow velocity in the embodiments of the present application; Figure 5 is a pseudo-spectrum diagram for using only a single frequency pair to estimate fixed velocity in the embodiments of the present application; Figure 6 is a pseudo-spectrum diagram for using difference frequency method to estimate fixed velocity in the embodiments of the present application. DETAILED DESCRIPTION

[0032] In order to make the person skilled in the art better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0033] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application can be combined with other embodiments.

[0034] Example 1 This embodiment discloses a blood flow velocity estimation method based on high-frequency pulse differential frequency, the flowchart of which is shown below. Figure 1 As shown. This embodiment is implemented using the MATLAB simulation platform combined with the Field II ultrasound simulation toolkit. Compared to traditional dual-mode system ultrasound blood flow estimation methods, this method avoids the problem of alternating transmission of B-mode and Doppler pulses. Compared to methods that directly use high-frequency or low-frequency pulses, this invention maintains the resolution advantage of high-frequency signals while obtaining the velocity detection sensitivity of low-frequency signals through differential processing, effectively solving the artifact problem and improving the estimation accuracy. The specific implementation steps are as follows: S1. Transmitting high-frequency broadband ultrasonic signals. In this embodiment, an 8-10MHz broadband high-frequency ultrasonic pulse signal is transmitted through an ultrasonic probe. The parameter is set to the pulse center frequency. Pulse repetition period The transmitted pulse is a uniform pulse with a length of 128, that is... .

[0035] S2. Receive echo signals from superficial blood vessels. This embodiment uses the Field II ultrasound simulation toolkit to simulate blood flow particle velocity, employing a 100MHz sampling frequency (…). The echo signal is sampled to obtain a discrete signal. .

[0036] S3, through 40 bandpass filters ( Narrowband frequency components were extracted, with each filter having a bandwidth of 0.1MHz and a center frequency of: in , Then, the Hilbert transform is used to convert each narrowband component into a complex signal:

[0037] in, Indicates for the first The first narrowband component The reflection coefficients of each scatterer were obtained through simulation using the Field II ultrasonic simulation toolkit. is the serial number of blood scatterers, M is the total number of blood scatterers (estimated according to the eigenvalue distribution in the blood flow coherence hypothesis), = , is the propagation speed of ultrasound in blood flow, and its value is 1540 , is the pulse repetition interval, and its value is .

[0038] At this time, the direct use of high-frequency pulse (e.g. ) to estimate the blood flow velocity of Field II simulation is shown in Figure 2 , but due to the limitations of high-frequency signals in low-speed blood flow detection, including insufficient frequency resolution, serious blood flow estimation artifacts caused by under-sampling, etc., the spectrum shown in Figure 2 is seriously blurred and has artifacts, and the blood flow velocity distribution cannot be seen, so the blood flow velocity estimation fails. Therefore, this paper proposes a differential frequency method, which can effectively overcome the limitations of direct estimation of high-frequency signals, and improve the reliability of blood flow velocity estimation accuracy.

[0039] S4, from the 40 narrow-band spectrum components selected in step S3 that meet the condition constitute data pairs ; and calculate the covariance matrix of the differential signal , where the total number of pulses P=128. S5, divide the equivalent low-frequency covariance matrix into overlapping sub-arrays ( ) and calculate the spatially smoothed covariance matrix ; Then perform eigenvalue decomposition to divide the signal subspace and the noise subspace; construct the steering vector , , , =1540m / s.

[0040] Spectrum peak search is performed by music algorithm, and spectrum estimation algorithm is used to estimate blood flow velocity:

[0041] As shown in Figure 3 , the Field II simulated blood flow velocity graph estimated by the differential frequency pulse signal accurately shows the blood flow velocity distribution, and the distribution curve has good continuity.

[0042] In contrast, traditional methods directly transmit low frequencies. The results of pulse-based blood flow velocity estimation are as follows Figure 4 As shown, although blood flow signals can be detected, the continuity of velocity estimation is significantly worse than that of low-frequency signals due to limitations in resolution and the number of snapshots. Figure 3 The results can also be found through comparison. Figure 4 The existence of more pronounced velocity estimation ambiguity also proves that the Doppler resolution (velocity dimension resolution) is inferior. Figure 3 The differential frequency method used.

[0043] Example 2 This embodiment refers to the steps in Embodiment 1 to achieve fixed blood flow velocity estimation. This embodiment will demonstrate the advantages of the average covariance matrix method proposed in this invention in effectively suppressing artifacts and accurately estimating blood flow velocity.

[0044] S1. Transmit a high-frequency broadband ultrasonic signal, referring to step S1 in Example 1.

[0045] S2, Receive echo signals from superficial blood vessels. Here, the blood flow particle velocity is set to a fixed velocity ( A sampling frequency of 100MHz was used. The echo signal is sampled to obtain a discrete signal. .in , These are the fast-time sampling index and the slow-time sampling index, respectively.

[0046] Steps S3 and S4 refer to steps S3 and S4 in Example 1, where frequency sampling and differential operation are performed on the received echo signal to obtain... An equivalent low-frequency covariance matrix .

[0047] S5. Select one of the steps in S4. ( Eigenvalues ​​are calculated and the signal and noise subspaces are divided. Blood flow velocity is estimated using the orthogonality of the signal and noise subspaces. The estimation results are as follows: Figure 5 As shown, there are obvious spurious peaks in the estimation, and even the correct peak value will be estimated to be off.

[0048] By using the covariance matrix averaging method proposed in this paper, the covariance matrices of all difference signals that meet the conditions are averaged. Generate an equivalent low-frequency covariance matrix, where Subsequently, the averaged covariance matrix was analyzed. The estimation results obtained by performing eigenvalue decomposition, subspace partitioning, and peak search are as follows: Figure 6shown, Figure 6 The clear and accurate velocity peak is shown, which verifies that the average covariance matrix calculated by the method can better suppress the artifacts.

[0049] It should be noted that, for the foregoing method embodiments, in order to facilitate the description, they are all expressed as a series of action combinations, but those skilled in the art should know that the present application is not limited by the action sequence described, because according to the present application, certain steps can be performed in other sequences or simultaneously.

[0050] The technical features of the above embodiments can be combined in any manner. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described, but as long as the combinations of the technical features do not exist contradictions, they should be considered as the scope of the present application.

[0051] The above embodiments are the preferred embodiments of the present application, but the embodiments of the present application are not limited by the above embodiments, and any changes, modifications, substitutions, combinations, simplifications made without departing from the spirit and principles of the present application are equivalent replacement methods and are included in the protection scope of the present application.

Claims

1. A method of Doppler blood flow velocity estimation based on high frequency pulse differential frequency, characterized by, The Doppler blood flow velocity estimation method comprises the following steps: S1, emitting high-frequency broadband ultrasonic signals compatible with B-mode imaging to superficial blood vessels, wherein the B-mode generates a two-dimensional gray-scale image by echo amplitude, and a high-frequency broadband signal is used to obtain a high-resolution ultrasonic imaging image; S2, receiving echo signals of the superficial blood vessels, and converting the echo signals into discrete signals through time domain sampling; S3, extracting a narrowband spectrum of the echo signals through frequency sampling, obtaining a plurality of frequency components, and applying a Hilbert transform to convert the frequency components into complex signals; S4, performing a difference operation on a pair of frequency components in the narrowband spectrum of the echo signals that satisfy a fixed difference frequency, calculating a covariance matrix of the difference signals, repeating the operation on all signals that satisfy the difference frequency, calculating a mean value of the covariance matrices, generating an equivalent low-frequency covariance matrix, and performing a difference operation on a pair of frequency components in the narrowband spectrum of the echo signals that satisfy a fixed difference frequency, calculating a covariance matrix of the difference signals, repeating the operation on all signals that satisfy the difference frequency, calculating a mean value of the covariance matrices, generating an equivalent low-frequency covariance matrix; S5, performing subarray division on the equivalent low-frequency covariance matrix and calculating a spatially smoothed covariance matrix, performing eigenvalue decomposition on the spatially smoothed covariance matrix, dividing a subspace, performing spectral peak searching, and applying a spectral estimation algorithm to estimate blood flow velocity.

2. The method of claim 1, wherein, The high-frequency broadband ultrasound signal is a Gaussian-modulated pulse sequence The expression is as follows: wherein, denotes a continuous time variable, is a slow time sampling index, recording the pulse sequence, capturing the phase change caused by blood flow velocity, denotes the total number of pulse sequences, denotes the pulse repetition period, is a Gaussian envelope function, is a cosine function.

3. The method of claim 1, wherein, The echo signal received in step S2 is a delayed and attenuated version of the transmitted signal after scattering by the blood cells, and is converted by an analog-to-digital converter at a sampling rate The time domain sampling is performed to obtain a discrete signal The expression is: wherein, is a fast-time sampling index, capturing the time-domain shape of the signal, reflecting information in the depth direction, is a slow-time sampling index, recording the pulse sequence, capturing the phase change caused by blood flow velocity, is a reflection coefficient, is a Gaussian envelope function, is a carrier center frequency, denotes the pulse repetition period, denotes the distance of the high-frequency broadband ultrasound signal transmitter from the center of the blood vessel, is the propagation speed of ultrasound in the blood flow, denotes the target radial moving speed, in m / s, is a cosine function.

4. The method of claim 3, wherein the method is based on high frequency pulse differential frequency Doppler blood flow velocity estimation. Step S3 performs frequency sampling on the discrete signal extracting a number of narrow-band frequency components, the center frequency of the first narrow-band component being : wherein denotes a frequency interval, denotes a start frequency, denotes a narrowband frequency component index, denotes the number of narrowband frequency components, and a Hilbert transform is applied to the first narrowband component to obtain a complex signal ​ wherein is the reflection coefficient for the first narrowband component, denotes the imaginary unit, and after demodulating the signal and removing the irrelevant terms the matrix expression for can be written as in, Represents the slow-time vector, The serial number of the blood flow scattering body. The total number of blood flow scatterers, and their direction vectors. = , Indicates for the first The first narrowband component The reflection coefficient of each scatterer Then it means that for the first Gaussian white noise added to a narrowband component For the first The radial velocity of each blood flow scatterer.

5. The method of claim 4, wherein the method is based on high frequency pulse differential frequency Doppler blood flow velocity estimation. The difference operation in step S4 specifically comprises: selecting any frequency in the narrow band component that satisfies conditions , the received data corresponding to the frequency , is written as a data pair , , and the calculated difference frequency data is obtained using the Hadamard product : wherein denotes multiplication of corresponding elements of matrices, denotes taking the complex conjugate of data denotes taking the complex conjugate; The covariance matrix is calculated for each differential signal : where, denotes the total number of pulse sequences, denotes the differential frequency data The conjugate transpose is taken and then the average of all differential signal covariance matrices that meet the criteria is taken to obtain the equivalent low frequency covariance matrix , where denotes the number of selected frequency pairs that meet the criteria.

6. The method of claim 5, wherein the method is based on high frequency pulse differential frequency Doppler blood flow velocity estimation. The spectral estimation algorithm in step S5 adopts a spatially smoothed MUSIC algorithm, and specifically comprises: The equivalent low-frequency covariance matrix R is divided into sub-arrays, and the spatially smoothed covariance matrix is calculated, where denotes the covariance matrix of the th sub-array. Re-against Eigenvalue decomposition is performed to divide signal subspace and noise subspace; direction vector is constructed: ; the spectral peak search function : Estimate the blood flow velocity v̂, where denotes the direction vector of the differential frequency, denotes the direction vector of the differential frequency, take the conjugate transpose, and denotes the noise subspace.

7. The method of claim 4, wherein the method is a Doppler blood flow velocity estimation method based on high frequency pulse differential frequency. The bandpass filter is an ideal filter, with a bandwidth set to wherein is a bandwidth preference factor, ensuring a sufficient frequency separation between the individual narrowband components.

8. The method of claim 5, wherein the method is a Doppler blood flow velocity estimation method based on high frequency pulse differential frequency. The covariance matrix averaging process can effectively suppress the cross-term artifact generated by different scatterers in the differential frequency calculation process when the number of frequency samples is large enough, the artifact term approaches zero .

Images