Ultrasonic imaging method based on frequency domain generalized sidelobe cancellation and Wiener post-filtering

Abstract

The invention relates to an ultrasonic imaging method based on frequency domain generalized sidelobe cancellation and Wiener post-filtering, and belongs to the technical field of ultrasonic imaging. The method comprises: delaying a sampling signal to obtain an ultrasonic echo signal, converting the signal into ultrasonic echo data of a sub-frequency band, performing space-time smoothing and diagonal loading processing to obtain a frequency domain sample covariance matrix under the sub-frequency band; calculating beamforming weights under each sub-frequency band according to the generalized sidelobe cancellation principle; designing a corresponding Wiener post-filter according to the beamforming weights to obtain Wiener filter coefficients; multiplying the beamforming weights with the post-filter coefficients to obtain a final composite weight vector, and then weighted summing the echo data under the frequency domain to obtain a frequency domain output of a beamformer; converting the frequency domain output value of the beamformer into the time domain, and performing subsequent imaging processing. The invention can significantly improve the imaging resolution and contrast of ultrasonic images.

Description

Technical Field

[0001] The present invention belongs to the technical field of ultrasonic imaging, and relates to an ultrasonic imaging method based on frequency-domain generalized sidelobe cancellation and Wiener post-filtering. Background Art

[0002] The most widely used and simplest beamforming technique in ultrasonic imaging is the delay-and-sum (DAS) algorithm. It calculates the delay amount of the received echo signals according to the geometric position relationship of the array element channels, and then aligns and superimposes the delayed data. The traditional DAS algorithm has low complexity and fast imaging speed. However, due to its use of a fixed window function for weighting, the main lobe width increases and the resolution is relatively low.

[0003] In recent years, in order to improve the contrast and resolution of beamforming algorithms, adaptive algorithms have been applied to the field of ultrasonic imaging. Among them, the minimum variance (MV) beamforming algorithm is the most widely used adaptive algorithm at present. This method minimizes the array output energy while keeping the gain in the desired direction unchanged. By dynamically calculating the weighting vector of the focused delay signals and weighting and summing the signals, the resolution of the image is improved. However, the disadvantage of this algorithm is that its robustness is much lower than that of the traditional delay-and-sum algorithm, and it is easy to cancel useful signals, which has a great impact on the image quality in the case of low signal-to-noise ratio and there are obvious deficiencies in contrast. The generalized sidelobe canceller (GSC), as a robust structure of the minimum variance algorithm, has better background imaging quality than the minimum variance algorithm. At present, some scholars have proposed introducing the generalized sidelobe canceller into the frequency domain (Generalized Sidelobe Canceler Based on Frequency Domain Segmentation, FDSGSC) to improve the applicable conditions of the algorithm. However, the resolution and background quality of the algorithm still need to be further improved.

[0004] Therefore, there is an urgent need to invent an algorithm that can further improve the resolution, contrast and robustness of the generalized sidelobe cancellation beamforming algorithm for ultrasonic imaging, and promote the further improvement of ultrasonic imaging quality. Summary of the Invention

[0005] In view of this, the purpose of the present invention is to provide an ultrasonic imaging method based on frequency-domain generalized sidelobe cancellation and Wiener post-filtering. By designing a frequency-domain Wiener post-filter to weight the output signal on the basis of the output of the frequency-domain generalized sidelobe cancellation algorithm, the imaging resolution and contrast of ultrasonic images are significantly improved, the problem of low contrast of the generalized sidelobe cancellation algorithm can be effectively overcome, and the overall imaging quality of the image is improved.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] An ultrasonic imaging method based on frequency-domain generalized sidelobe cancellation and Wiener post-filtering, specifically including the following steps:

[0008] S1: Perform delay processing on the sampled signals received by the array elements to obtain ultrasonic echo signals;

[0009] S2: Select a suitable frequency-domain segmentation length for the ultrasonic echo signals, and at the same time use the short-time Fourier transform to convert the ultrasonic echo signals into narrow-band sub-signals in the frequency domain, that is, ultrasonic echo data of sub-bands;

[0010] S3: Perform subspace spatial smoothing, diagonal loading, etc. on the ultrasonic echo data of sub-bands in sequence to obtain the frequency-domain sample covariance matrix under sub-bands;

[0011] S4: Calculate the beamforming weights under each sub-band according to the principle of generalized sidelobe cancellation;

[0012] S5: Design a corresponding Wiener post-filter according to the beamforming weights of the frequency-domain generalized sidelobe canceller under sub-bands, and obtain the Wiener filter coefficients;

[0013] S6: Multiply the beamforming weights of the frequency-domain generalized sidelobe canceller under sub-bands by its post-filter coefficients to obtain the final composite weight vector;

[0014] S7: Use the composite weight vector to perform weighted summation on the echo data in the frequency domain to obtain the frequency-domain output of the beamformer;

[0015] S8: Use the inverse fast Fourier transform to convert the frequency-domain output value of the beamformer into a time-domain output value, and obtain the final time-domain output for imaging.

[0016] Furthermore, in step S2, to obtain the ultrasonic echo data of sub-bands, specifically including the following steps:

[0017] S21: Perform frequency-domain segmentation on the ultrasonic echo signal x(k) using the short-time Fourier transform, and the process is shown in the following formula:

[0018]

[0019] Among them, z(k) represents a hanning window with a window length and the number of points of the short-time Fourier transform both being 64 and no signal overlap, x(k) represents the ultrasonic echo time-domain signal, S(m, ω) represents the corresponding value of the signal x(k) in the time-frequency domain, m represents the segmentation serial number, ω represents the frequency-domain serial number under the segmentation serial number m, k represents the sampling moment, K represents the total number of sampling moments, i represents the imaginary unit, and e represents the exponent;

[0020] S22: Based on the previous step, the ultrasonic echo signals of each sensor element are sequentially converted into a number of independent equally spaced narrowband sub-signals. The sub-signal s n (m, ω) of the m-th segment of the n-th aperture is expressed as:

[0021] S n (m, ω) = [S n (m, 1),..., S n (m, W - 1), S n (m, W)]

[0022] where m = 1, 2,..., M, M represents the total number of segments into which the n-th aperture signal is divided; ω = 1, 2,..., W, W represents the length of each narrow sub-band signal; M×W represents the total length of the aperture signal in the time domain, and S n (m, ω) represents the signal amplitude at the point (m, ω) in the time-frequency domain of the n-th aperture.

[0023] Furthermore, in step S3, obtaining the sample covariance matrix under the sub-frequency band specifically includes the following steps:

[0024] S31: Extract the array information of each time-frequency point. The array signal X(m, ω) at the time-frequency point (m, ω) is expressed as:

[0025] X(m, ω) = [S 1 (m, ω), S 2 (m, ω)... S N (m, ω)]

[0026] S32: Perform sub-array partitioning on the array signals of each time-frequency point, and construct the frequency-domain sample covariance matrix R(m, ω) based on the array signals after sub-array partitioning:

[0027]

[0028] where X l (m, ω) = [S l (m, ω), S l+1 (m, ω),..., S l+L-1 (m, ω)] represents the frequency-domain forward smoothing vector of the l-th sub-array, and l = (1, 2,..., N - L + 1), N represents the total length of the ultrasonic array, and L represents the length of the sub-array; X l (m, ω) H is the conjugate transpose of X l (m, ω);

[0029] S33: Perform diagonal loading on the frequency-domain sample covariance matrix R(m, ω) to enhance the robustness of the algorithm, and obtain the covariance matrix after diagonal loading. Expression:

[0030]

[0031] where β = trace(R(m, ω))·Δ, trace(R(m, ω)) is the equivalent power of the signal, trace(·) is the function to calculate the trace of a matrix, and Δ is the ratio of spatial noise to signal power. R 0 is the identity matrix.

[0032] Furthermore, step S4 specifically includes: calculating the beamforming weights at each time-frequency point (m, ω) according to the principle of generalized sidelobe cancellation. The expression is:

[0033] w FDS-GSC (m, ω) = w sq (m, ω) - Bw sa (m, ω)

[0034] where w FDS-GSC (m, ω) represents the beamforming weight of the sub-band generalized sidelobe canceller at the time-frequency point (m, ω), w sq (m, ω) is the non-adaptive weight vector of the main branch of the sub-band generalized sidelobe canceller at the time-frequency point (m, ω), w sa (m, ω) is the adaptive weight vector of the auxiliary branch of the sub-band generalized sidelobe canceller at the time-frequency point (m, ω); where w sq and w sa are related as:

[0035]

[0036] where B H is the conjugate transpose of the blocking matrix B, is the frequency-domain estimated sample covariance matrix after diagonal loading, and (·) -1 represents matrix inversion operation.

[0037] Furthermore, step S5 specifically includes: taking the output of the sub-band generalized sidelobe canceller as the estimated value of the desired signal at the time-frequency point (m, ω), designing a corresponding Wiener post-filter, and obtaining the frequency-domain Wiener post-filter coefficient H FDSGSC-Wiener , and the expression is:

[0038]

[0039] where w FDS-GSCis the adaptive beamforming weight of the sub-band generalized sidelobe canceller. is w FDS-GSC 's conjugate transpose, and S n (m, ω) is the ultrasonic frequency-domain signal of the nth aperture at the time-frequency point (m, ω). is the conjugate transpose of S n (m, ω).

[0040] Furthermore, step S6 specifically includes: multiplying the beamforming weight of the sub-band generalized sidelobe canceller by its post-filter weight to obtain the final composite weight w FDSGSC-Wiener , and the expression is:

[0041] w FDSGSC-Wiener = H FDSGSC-Wiener w FDS-GSC

[0042] where w FDSGSC-Wiener represents the composite weight of the frequency-domain generalized sidelobe canceller combined with the Wiener post-filter; H FDSGSC-Wiener represents the frequency-domain Wiener post-filter coefficient of the generalized sidelobe canceller based on the frequency domain, and w FDS-GSC represents the adaptive beamforming weight of the sub-band generalized sidelobe canceller.

[0043] Furthermore, in step S7, the expression for obtaining the frequency-domain output of the beamformer is:

[0044]

[0045] where y FDSGSC-Wiener (m, ω) represents the frequency-domain output value of the sub-band ultrasonic generalized sidelobe cancelling beamformer with a fused Wiener post-filter at the time-frequency point (m, ω), and w FDSGSC-Wiener represents the composite weight of the frequency-domain generalized sidelobe canceller combined with the Wiener post-filter, represents the conjugate transpose of w FDSGSC-Wiener ; X l (m, ω) represents the frequency-domain data of the lth subarray at the time-frequency point (m, ω); N represents the total length of the ultrasonic array, and L represents the length of the subarray.

[0046] Furthermore, in step S8, the expression for obtaining the final time-domain output is:

[0047]

[0048] where y FDSGSC-Wiene r(k) represents the time-domain output value of the sub-band ultrasonic generalized sidelobe cancelling beamformer with a fused Wiener post-filter at the sampling time k; ISTFT(·) represents the inverse short-time Fourier transform, and y FDSGSC-Wiener(m, ω) represents the frequency-domain output value of the sub-band ultrasonic generalized sidelobe cancellation beamformer with a fused Wiener post-filter at the time-frequency domain point (m, ω).

[0049] The beneficial effects of the present invention are as follows: First, the present invention uses the short-time Fourier transform to divide the time-domain ultrasonic signal into short narrow-band signals and introduce them into the time-frequency domain, and performs the generalized sidelobe cancellation beamforming algorithm in the time-frequency domain. Then, the Wiener post-filter coefficients in the frequency domain are calculated based on the output value of the beamformer in the frequency domain, and the frequency-domain output is further weighted and optimized to further improve the algorithm resolution and sidelobe suppression ability. Finally, the inverse short-time Fourier transform method is used to convert the time-frequency domain output into the time domain for imaging. Therefore, the present invention can further improve the resolution and contrast performance of the traditional ultrasonic generalized sidelobe cancellation algorithm and has a higher effect on sidelobe suppression.

[0050] Other advantages, objectives, and features of the present invention will be described to some extent in the subsequent specification, and to some extent, will be obvious to those skilled in the art based on the study of the following text, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be achieved and obtained through the following specification. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be described in detail preferably with reference to the accompanying drawings, where:

[0052] Figure 1 is a flowchart of the ultrasonic imaging method based on frequency-domain generalized sidelobe cancellation and Wiener post-filtering of the present invention;

[0053] Figure 2 is the imaging effect diagram of point targets of 4 algorithms;

[0054] Figure 3 is the lateral resolution curve graph of point targets at a depth of 50 mm for 4 algorithms;

[0055] Figure 4 is the imaging effect diagram of multi-spot for 4 algorithms;

[0056] Figure 5 is the lateral resolution curve graph of multi-spot imaging at a depth of 50 mm for 4 algorithms. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0057] The following describes the implementation manners of the present invention through specific examples. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific implementation manners. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the diagrams provided in the following embodiments only illustrate the basic concept of the present invention in a schematic manner. Without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0058] Please refer to Figures 1 to 5 , Figure 1 a Wiener post-filter ultrasonic imaging algorithm based on a frequency-domain generalized sidelobe canceller provided by the present invention, which includes the following steps:

[0059] Step S1: Amplify, perform AD conversion, and delay focusing processing on the echo signal received by the ultrasonic array elements to obtain ultrasonic echo data; obtain the signal x(k) after delay focusing processing, and x(k) is expressed as x(k) = [x 1 (k), x 2 (k),..., x N (k)], where N represents the number of array elements of the ultrasonic array, and k represents the sampling moment corresponding to the sampling depth;

[0060] Step S2: Select a suitable frequency-domain segmentation length for the ultrasonic echo signal, and at the same time use the short-time Fourier transform to convert the ultrasonic echo signal into narrowband sub-signals in the frequency domain. Specifically, it includes the following steps:

[0061] S21: Perform frequency-domain segmentation on the ultrasonic echo signal x(k) using the short-time Fourier transform, and the process is shown in the following formula:

[0062]

[0063] where z(k) represents a hanning window with a window length and the number of points of the short-time Fourier transform both being 64 and no signal overlap, x(k) represents the ultrasonic echo time-domain signal; S(m, ω) represents the corresponding value of the x(k) signal in the time-frequency domain; m represents the segmentation serial number, and ω represents the frequency-domain serial number under the segmentation serial number m; k represents the sampling moment, K represents the total number of sampling moments; i represents the imaginary unit, and e represents the exponent.

[0064] S22: Based on the previous step, sequentially convert the ultrasonic echo signals of each sensor array element into a number of independent equally spaced narrowband sub-signals. The expression of the sub-signal S n (m, ω) of the m-th segment on the n-th aperture is:

[0065] S n (m, ω) = [Sn (m, 1),..., S n (m, W - 1), S n (m, W)]

[0066] where m = 1, 2,..., M, M represents the total number of segments into which the aperture signal is divided; ω represents the frequency-domain sequence number at the segmentation sequence number m; ω = 1, 2,..., W, W is the length of each narrow sub-band signal; M×W represents the total length of the aperture signal in the time domain; S n (m, ω) represents the signal amplitude at the point (m, ω) in the time-frequency domain of the nth aperture.

[0067] Step S3: Extract the array information of each time-frequency point, and perform processing such as sub-array spatial smoothing and diagonal loading on the ultrasonic echo data of the sub-bands in sequence to obtain the sample covariance matrix under the sub-bands. Specifically, it includes the following steps:

[0068] S31: Extract the array information of each time-frequency point. The array signal X(m, ω) at the time-frequency point (m, ω) can be expressed as:

[0069] X(m, ω) = [S 1 (m, ω), S 2 (m, ω)...,, S N (m, ω)]

[0070] S32: Perform sub-array partitioning on the array information of each time-frequency point, and construct the frequency-domain sample covariance matrix R(m, ω) based on the array information after sub-array partitioning:

[0071]

[0072] where X l (m, ω) = [S l (m, ω), S l+1 (m, ω),..., S l+L-1 (m, ω)] represents the frequency-domain forward smoothing vector of the lth sub-array, and l = 1, 2,..., N - L + 1, N represents the total length of the ultrasonic array, L represents the length of the sub-array, X l (m, ω) H is the conjugate transpose of X l (m, ω).

[0073] S33: Perform diagonal loading processing on the frequency-domain sample covariance matrix R(m, ω) to enhance the robustness of the algorithm and obtain the covariance matrix after diagonal loading as follows:

[0074]

[0075] Among them, β = trace(R(m, ω))·Δ, where trace(R(m, ω)) is the equivalent power of the signal, trace(·) is the function for calculating the matrix trace, and Δ is the ratio of the spatial noise to the signal power. R 0 is the identity matrix.

[0076] Step S4: Calculate the output weight of the beamformer at each time-frequency point (m, ω) according to the principle of the generalized sidelobe canceller, as shown in the following formula:

[0077] w FDS-GSC (m, ω) = w sq (m, ω) - Bw sa (m, ω)

[0078] In the formula, w FDS-GSC (m, ω) represents the weighted vector of the sub-band generalized sidelobe canceller at the time-frequency point (m, ω), w sq (m, ω) is the non-adaptive weight vector of the main branch of the sub-band generalized sidelobe canceller at the time-frequency point (m, ω), w sa (m, ω) is the adaptive weight vector of the auxiliary branch of the sub-band generalized sidelobe canceller at the time-frequency point (m, ω). Among them, w sq and w sa have the following relationship:

[0079]

[0080] Among them, B H is the conjugate transpose of the blocking matrix B, is the frequency-domain estimated sample covariance matrix after diagonal loading, and (·) -1 represents the matrix inversion operation.

[0081] Step S5: Take the output of the sub-band generalized sidelobe canceller as the estimated value of the desired signal at the time-frequency point (m, ω), design the corresponding Wiener post-filter, and obtain the frequency-domain Wiener post-filter coefficient H FDSGSC-Wiener as follows:

[0082]

[0083] Among them, w FDS-GSC is the adaptive beamforming weight of the sub-band generalized sidelobe canceller, is the conjugate transpose of w FDS-GSC , S n (m, ω) is the ultrasonic frequency-domain signal of the nth aperture at the time-frequency point (m, ω), is the conjugate transpose of S n (m, ω).

[0084] Step S6: Multiply the weights of the sub-band generalized sidelobe canceller by the weights of its post-filter to obtain the final composite weight vector w FDSGSC-Wiener :

[0085] w FDSGSC-Wiener = H FDSGSC-Wiener w FDS-GSC

[0086] where w FDSGSC-Wiener represents the final composite weight vector obtained by the sub-band ultrasonic generalized sidelobe canceller algorithm of the fused Wiener post-filter. H FDSGSC-Wiener represents the frequency-domain Wiener post-filter coefficient of the generalized sidelobe canceller algorithm based on the frequency domain, and w FDS-GSC represents the weighting coefficient of the original frequency-domain generalized sidelobe cancelling beamformer.

[0087] Step S7: Use the composite weight vector to perform weighted summation on the echo data in the frequency domain to obtain the frequency-domain output of the beamformer, as shown in the following formula:

[0088]

[0089] where y FDSGSC-Wiener (m, ω) represents the frequency-domain output value of the sub-band ultrasonic generalized sidelobe cancelling beamformer with a fused Wiener post-filter at the time-frequency point (m, ω), and w FDSGSC-Wiener represents the composite weight of the frequency-domain generalized sidelobe canceller combined with the Wiener post-filter, and represents the conjugate transpose of w FDSGSC-Wiener . X l (m, ω) represents the frequency-domain data of the l-th subarray at the time-frequency point (m, ω); N represents the total length of the ultrasonic array, and L represents the length of the subarray.

[0090] Step S8: Use the inverse fast Fourier transform to convert the frequency-domain output value of the beamformer into a time-domain output value, and obtain the final time-domain output for imaging:

[0091]

[0092] where y FDSGSC-Wiener (k) represents the time-domain output value of the sub-band ultrasonic generalized sidelobe cancelling beamformer with a fused Wiener post-filter at the sampling time k, ISTFT(·) represents the inverse short-time Fourier transform, and y FDSGSC-Wiener (m, ω) represents the frequency-domain output value of the sub-band ultrasonic generalized sidelobe cancelling beamformer with a fused Wiener post-filter at the time-frequency point (m, ω).

[0093] Verification experiment:

[0094] Field II is an ultrasonic experimental simulation platform developed by the Technical University of Denmark based on acoustic principles, which has been widely recognized and used in theoretical research. To verify the effectiveness of the proposed algorithm, Field II is used to image the commonly used point scatter targets and anechoic cyst targets in ultrasonic imaging, and imaging comparison experiments are carried out using actual experimental data. In the point target simulation experiment, 12 point targets with a single-column longitudinal interval of 2.5 mm are set, and the depth is distributed between 40 mm and 70 mm. The emission fixed focusing and reception dynamic focusing methods are adopted, the emission focus is fixed at 50 mm, and the imaging dynamic range of the image is set to 60 dB.

[0095] In the multi-speckle imaging simulation experiment, 3 scatter point targets are set with a transverse position in the center at 0 mm and longitudinal positions at depths of 32.5 mm, 50 mm, and 67.5 mm. In addition, 2 scatter point targets are set at longitudinal 50 mm and transverse ±5 mm to observe the transverse resolution of each algorithm. The synthetic aperture focusing method is adopted, and the imaging dynamic range of the image is set to 60 dB. At the same time, in the speckle medium, 2 anechoic cysts with a radius of 3 mm are set, and the centers of the circles are located at (-5 mm, 40 mm) and (5 mm, 50 mm) respectively, and 2 strong speckles with a radius of 3 mm are set, and the centers of the circles are located at (5 mm, 40 mm) and (-5 mm, 50 mm). The amplitude ratio between the scatter points and the background in the massive cyst is 10 times, and the amplitude ratio between the anechoic cyst and the background is 40 times. In addition, the center frequency of the array element used in the experiment is 7 MHz, the number of array elements is 64, the spacing is 0.24 mm, the sampling frequency is 100 MHz, the sound speed is 1540 m / s, and the imaging dynamic range is set to 60 dB.

[0096] The delay and sum algorithm (DAS), the generalized sidelobe cancellation algorithm (GSC), the fractional sub-band generalized sidelobe cancellation algorithm (FDSGSC), and the fractional sub-band generalized sidelobe cancellation algorithm integrated with a Wiener post-filter (FDSGSC-Wiener) are used to conduct comparative imaging experiments on the above two experimental targets.

[0097] Figure 2 The imaging results of the point targets of the 4 algorithms are given. From Figure 2 it can be seen that the imaging quality of the DAS algorithm is the worst, with serious sidelobe artifacts. Compared with other algorithms, it has higher transverse artifacts and a wider main lobe width, making it difficult to distinguish the target points. The GSC algorithm significantly narrows the main lobe width and significantly improves the transverse resolution, but it still performs poorly in suppressing sidelobe artifacts.

[0098] The resolution of the FDSGSC algorithm is further improved throughout the entire depth range, but artifacts are still visible. The FDSGSC-Wiener algorithm further improves the imaging quality, with the main lobe width and sidelobe level being the best among the four algorithms, and its far-field resolution is significantly improved.

[0099] Figure 3 The lateral resolution curves of the four algorithms at a 50-mm focal point are given. Table 1 presents the main lobe width values at -6 dB for the four imaging algorithms at depths of 40 mm, 50 mm, and 60 mm in the case of a point target. It can be seen that the main lobe width of FDSGSC-Wiener is the best, indicating its best point-target resolution. FDSGSC is second, and the effect of DAS is poor. Compared with the DAS algorithm, the improvement in the main lobe width of FDSGSC-Wiener in resolution is the largest, increasing by 80.7%; compared with the GSC algorithm, the improvement in the main lobe width of FDSGSC-Wiener in resolution is the largest, increasing by 51.6%, and the resolution improvement is obvious.

[0100] Table 1 Comparison of the -6 dB FWHM of the four algorithms at different depths in the point-target experiment

[0101]

[0102] Figure 4 The multi-spot imaging effect diagrams of the four algorithms are given. Table 2 presents the main lobe width data at -6 dB of the lateral resolution curves of the four algorithms for point targets at different depths in the multi-spot imaging experiment. Combining Table 2 and Figure 4 It can be seen that the DAS algorithm has the worst resolution, the widest point imaging, low resolution, and obvious artifacts inside the dark spots. The resolution of the GSC algorithm is improved compared with the DAS algorithm, and the FDSGSC-Wiener further improves the resolution performance compared with the GSC and FDSGSC algorithms. Compared with the DAS and GSC algorithms, the resolution of FDSGSC-Wiener is increased by 74.2% and 50%, respectively. It can be seen that the improvement in the far-field resolution of the FDSGSC-Wiener algorithm is very obvious. To more intuitively observe the resolution situation, Figure 5 The lateral resolution curves of the four algorithms at a 50-mm depth in multi-spot imaging are given. It can be seen that the FDSGSC-Wiener algorithm has the optimal main lobe width and the lowest sidelobe level, indicating its highest point-target imaging performance.

[0103] Table 2 Comparison of the -6 dB FWHM of the four algorithms at different depths in the multi-spot imaging experiment

[0104]

[0105] Table 3 presents a comparison of the imaging performance metrics of different imaging algorithms for multi-spot imaging. It can be seen that, compared with the DAS algorithm, the GSC algorithm incurs a loss in contrast ratio (CR), and the signal-to-noise ratio drops significantly. The improvement of the FDSGSC algorithm in terms of CR is only on par with that of DAS, while the CR value of FDSGSC-Wiener reaches 33.18, which is an increase of 5.75 compared with the GSC algorithm. This indicates that FDSGSC-Wiener significantly improves the quality of spot imaging, and its CNR and SD values are significantly improved compared with the GSC algorithm.

[0106] Table 3 Comparison of the imaging performance metrics of different imaging algorithms for multi-spot imaging

[0107]

[0108] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the present technical solution, and they should all be covered within the scope of the claims of the present invention.

Claims

1. An ultrasonic imaging method based on frequency domain generalized sidelobe cancellation and Wiener post-filtering, characterized in that: The method specifically comprises the following steps: S1: Delay processing is performed on the sampling signal received by the array element to obtain an ultrasonic echo signal; S2: Selecting a frequency domain segment length suitable for the ultrasonic echo signal, and using short-time Fourier transform to convert the ultrasonic echo signal into a narrowband sub-signal in the frequency domain, that is, ultrasonic echo data of the sub-band, specifically comprising the following steps: S21: Use short-time Fourier transform to perform frequency domain segmentation on the ultrasonic echo signal x(k), and the process is shown in the following formula: Wherein, z(k) represents a Hanning window with a window length and a short-time Fourier transform point number of 64 and no signal overlap, x(k) represents the ultrasonic echo time domain signal, S(m,ω) represents the corresponding value of the signal x(k) in the time-frequency domain, m represents the segment number, ω represents the frequency domain number under the segment number m, k represents the sampling time, K represents the total number of sampling times, i represents the imaginary unit, and e represents the exponent; S22: Convert the ultrasonic echo signal of each sensor array element into a number of independent equally spaced narrow-band sub-signals in sequence. The mth narrow-band sub-signal S on the nth aperture n The expression of (m,ω) is: S n (m,ω)=[S n (m,1),...,S n (m,W-1),S n (m,W)] Where m = 1, 2, ..., M, M represents the number of segments into which the nth aperture signal is divided; ω = 1, 2, ..., W, W represents the length of each narrow subband signal; M × W represents the total length of the aperture signal in the time domain, S n (m,ω) represents the ultrasonic frequency domain signal at the time-frequency point (m,ω) at the nth aperture; S3: Processing the ultrasonic echo data of the sub-bands in sequence to obtain the frequency domain sample covariance matrix of the sub-bands; S4: Calculate the beamforming weights at each sub-band according to the generalized sidelobe cancellation principle, specifically including: Calculate the beamforming weights at each time-frequency point (m, ω) according to the generalized sidelobe cancellation principle, the expression is: w FDS-GSC (m,ω)=w sq (m,ω)-Bw sa (m,w) Among them, w FDS-GSC (m,ω) represents the beamforming weight of the sub-band generalized sidelobe canceller at the time-frequency point (m,ω), and w sq (m,ω) is the non-adaptive weight vector of the main branch of the sub-band generalized sidelobe canceller at the time-frequency point (m,ω), w sa (m,ω) is the adaptive weight vector of the auxiliary branch of the sub-band generalized sidelobe canceller at the time-frequency point (m,ω); where w sq With w sa The relationship is: Among them, B H is the conjugate transpose of the blocking matrix B, is the frequency domain estimated sample covariance matrix after diagonal loading, (·) 1 Represents matrix inversion operation; S5: Design a corresponding Wiener post-filter according to the beamforming weights of the frequency domain generalized sidelobe canceller in the sub-band and obtain the Wiener filter coefficients, specifically including: taking the output of the generalized sidelobe canceller in the sub-band as the estimated value of the desired signal at the time-frequency point (m, ω), designing a corresponding Wiener post-filter, and obtaining the frequency domain Wiener post-filter coefficients H based on the sub-band generalized sidelobe canceller FDSGSC-Wiener , the expression is: Among them, w FDS-GSC is the beamforming weight of the sub-band generalized sidelobe canceller, w FDS-GSC The conjugate transpose of S n (m,ω) is the ultrasonic frequency domain signal of the nth aperture at the time-frequency point (m,ω), For S n The conjugate transpose of (m,ω); S6: multiplying the beamforming weight of the frequency domain generalized sidelobe canceller under the sub-band by its post-filter coefficient to obtain the final composite weight vector; S7: performing weighted summation on the echo data in the frequency domain using the composite weight vector to obtain the frequency domain output of the beamformer; S8: The frequency domain output value of the beamformer is converted into a time domain output value by using an inverse fast Fourier transform, and the final time domain beamforming output is obtained and imaged.

2. The ultrasonic imaging method based on frequency domain generalized sidelobe cancellation and Wiener post-filtering according to claim 1, characterized in that: In step S3, the sample covariance matrix under the sub-band is obtained, which specifically includes the following steps: S31: Extract array information of each time-frequency point. The array signal X(m,ω) of the time-frequency point (m,ω) is expressed as: X(m,ω)=[S1(m,ω),S2(m,ω)...,S N (m,ω)] S32: Divide the array signals at each time-frequency point into sub-arrays, and construct a frequency domain sample covariance matrix R(m,ω) based on the array signals after the sub-array division: Among them, X l (m,ω)=[S l (m,ω),S l+1 (m,ω),...,S l+L-1 (m,ω)] represents the frequency domain data of the lth subarray at the time-frequency domain point (m,ω), and l = (1,2,...,N-L+1), N represents the total length of the ultrasound array, and L represents the subarray length; X l (m,ω) H For X l The conjugate transpose of (m,ω); S33: Perform diagonal loading on the frequency domain sample covariance matrix R(m,ω) to obtain the covariance matrix after diagonal loading expression: Where β = trace(R(m,ω))·Δ, trace(R(m,ω)) is the equivalent power of the signal, trace(·) is the function of finding the matrix trace, Δ is the ratio of spatial noise to signal power, R0 is the identity matrix.

3. The ultrasonic imaging method based on frequency domain generalized sidelobe cancellation and Wiener post-filtering according to claim 1, characterized in that: Step S6 specifically includes: multiplying the beamforming weight of the sub-band generalized sidelobe canceller by its post-filter weight to obtain the final composite weight w FDSGSC-Wiener , the expression is: w FDSGSC-Wiener =H FDSGSC-Wiener w FDS-GSC Among them, w FDSGSC-Wiener H represents the composite weight of the frequency domain generalized sidelobe canceller combined with the Wiener postfilter; FDSGSC-Wiener represents the frequency domain Wiener post-filter coefficients of the frequency domain based generalized sidelobe canceller, w FDS-GSC represents the beamforming weights of the sub-band generalized sidelobe canceller.

4. The ultrasonic imaging method based on frequency domain generalized sidelobe cancellation and Wiener post-filtering according to claim 3, characterized in that: In step S7, the frequency domain output expression of the beamformer is obtained as follows: Among them, y FDSGSC-Wiener (m, ω) represents the frequency domain output value of the sub-band ultrasonic generalized sidelobe cancellation beamformer fused with the Wiener post-filter at the time-frequency domain point (m, ω), w FDSGSC-Wiener represents the composite weights of the frequency domain generalized sidelobe canceller combined with the Wiener postfilter, Indicates w FDSGSC-Wiener The conjugate transpose of X l (m, ω) represents the frequency domain data of the lth subarray at the time-frequency domain point (m, ω); N represents the total length of the ultrasound array, and L represents the subarray length.

5. The ultrasonic imaging method based on frequency domain generalized sidelobe cancellation and Wiener post-filtering according to claim 4, characterized in that: In step S8, the expression of the final time domain beamforming output is obtained as: Among them, y FDSGSC-Wiener (k) represents the time domain output value of the sub-band ultrasonic generalized sidelobe cancellation beamformer fused with the Wiener post-filter at the sampling time k; ISTFT(·) represents the inverse short-time Fourier transform, y FDSGSC-Wiener (m, ω) represents the frequency domain output value of the sub-band ultrasonic generalized sidelobe cancellation beamformer fused with the Wiener post-filter at the time-frequency domain point (m, ω).

Images