Time domain beamforming acceleration method and device based on sparse matrix

Abstract

The present invention discloses a time domain beamforming acceleration method and device based on sparse matrix, the method comprises the following steps: averagely divide the pixel points in the three-dimensional imaging space into blocks; calculate the delay vector according to the distance vector of each group of pixel points to the two-dimensional plane receiving array and the two-dimensional plane transmitting array respectively; calculate the sampling index according to the sampling rate and delay vector of the two-dimensional plane receiving array, and calculate the interpolation coefficient vector according to the sampling index; calculate the block sparse matrix corresponding to each group of pixel points according to the interpolation coefficient vector and the delay vector; combine all the block sparse matrices into an overall sparse matrix; perform beamforming calculation on the input time domain multi-channel signal and the overall sparse matrix. The present invention can effectively improve the efficiency of time domain beamforming, and the overall sparse matrix can be imaged multiple times after being stored and read, which greatly reduces the single imaging time.

Description

Technical Field

[0001] The invention belongs to the technical field of underwater acoustic array signal processing, and in particular relates to a time domain beamforming acceleration method and device based on a sparse matrix. Background Art

[0002] Due to the serious attenuation of light waves and electromagnetic waves in water, the imaging equipment and technical means commonly used on land and in the air have limited application scope underwater. Sound waves can not only propagate over long distances in water with low attenuation, but also have good directionality and penetration. Therefore, sound waves have become the most commonly used signal carrier for underwater detection. The three-dimensional sonar imaging system can not only overcome the problem of short visual distance of optical imaging systems, but also obtain more intuitive three-dimensional space-target image information compared to traditional two-dimensional imaging sonar, so it has received more and more attention and application.

[0003] According to different imaging methods, three-dimensional imaging sonar can be divided into acoustic lens imaging, beamforming acoustic imaging and holographic acoustic imaging. With the development of digital signal processing technology, beamforming acoustic imaging sonar has been more and more widely used. When working, the beamforming acoustic imaging sonar uses the transmitting transducer to actively transmit sound waves of a specific frequency, receives the scattered signals within a specific field of view through the receiving transducer, and uses the signal processor that implements the beamforming algorithm to process the sonar signal in real time, and finally inverts the spatial distribution of the sound pressure amplitude within the field of view. This process not only requires the algorithm to have a high degree of real-time and accuracy, but also places strict requirements on the computing power of the signal processor.

[0004] In beamforming acoustic imaging sonar, signal processing based on beamforming is a time-consuming key part of the imaging calculation process. Its computational complexity and efficiency directly affect the real-time performance and accuracy of the entire imaging system. At present, common beamforming imaging algorithms include time-domain beamforming, deconvolution imaging, and compressed beamforming. Among them, the time-domain beamforming algorithm appeared the earliest and is the most mature. With its advantages of good robustness and small amount of calculation, it has become a technical solution commonly used in various sonar devices. The algorithm calculates the time delay of sound wave propagation at each spatial pixel position, and performs phase compensation and superposition on the received signal accordingly, thereby realizing the reconstruction of the target sound field.

[0005] However, the time-domain beamforming algorithm still faces many challenges in practical applications. First, the time-domain beamforming algorithm needs to calculate the time delay of beamforming for each spatial pixel position. If a dense matrix storage method is used, the storage volume will increase dramatically, thereby increasing the hardware cost and operation burden of the system. Secondly, in conventional time-domain beamforming calculations, the frequent indexing operations of time delay parameters become the main bottleneck restricting the improvement of algorithm performance. These indexing operations not only increase the computational complexity, but also reduce the real-time performance of the algorithm. As an important technical means in the field of underwater detection, the performance improvement of the three-dimensional sonar imaging system is inseparable from the continuous optimization and innovation of the beamforming algorithm to achieve more efficient three-dimensional underwater imaging. Summary of the invention

[0006] In view of the above, the purpose of the present invention is to provide a time-domain beamforming acceleration method and device based on a sparse matrix, which uses a sparse matrix to store signal interpolation coefficients and corresponding time delays, and uses multiplication operations instead of index operations with time-domain multi-channel signals in time-domain beamforming calculations. This can greatly improve the efficiency of beamforming and reduce the storage space required for time delay parameters. It is suitable for receiving beamforming algorithms based on plane wave imaging using two-dimensional arrays in three-dimensional imaging sonars, and has the advantages of small storage space, high computational efficiency, and easy parallelization.

[0007] In order to achieve the above-mentioned invention object, the technical solution provided by the present invention is as follows:

[0008] An embodiment of the present invention provides a sparse matrix-based time-domain beamforming acceleration method, comprising the following steps:

[0009] Expanding the spatial distribution of three-dimensional pixel points in the set three-dimensional imaging space to one dimension, and then evenly dividing the pixel points into blocks to obtain a plurality of groups of pixel points;

[0010] A delay vector is calculated based on the distance vectors from each group of pixels to the two-dimensional plane receiving array and the two-dimensional plane transmitting array; a sampling index is calculated based on the sampling rate of the two-dimensional plane receiving array and the delay vector, and an interpolation coefficient vector is calculated based on the sampling index; a block sparse matrix corresponding to each group of pixels is calculated based on the interpolation coefficient vector and the delay vector;

[0011] Combine the block sparse matrices corresponding to all groups of pixels into an overall sparse matrix;

[0012] The time-domain multi-channel signal emitted by the two-dimensional planar transmit array is obtained and expanded into a one-dimensional arrangement. A one-dimensional matrix is ​​calculated based on the overall sparse matrix and the one-dimensionally arranged time-domain multi-channel signal. The one-dimensional matrix is ​​transformed in dimension and reconstructed into three-dimensional space to complete the beamforming calculation of the time-domain multi-channel signal.

[0013] Preferably, the step of expanding the spatial distribution of three-dimensional pixel points in the set three-dimensional imaging space into one dimension and then evenly dividing the pixel points into blocks to obtain a plurality of groups of pixel points includes:

[0014] Set the three-dimensional imaging space in , and The resolution and imaging range of the direction are obtained respectively. , and The number of imaging points in the three directions are , and , the total number of pixels is ;

[0015] Flatten the three-dimensional space points to a one-dimensional arrangement along the row direction. Pixels are divided into Group, that is, the average per group pixels, each group of pixels corresponds to axis, Axis and The axis coordinates form the coordinate vector , and .

[0016] Preferably, the delay vector is calculated based on the distance vectors from each group of pixels to the two-dimensional plane receiving array and the two-dimensional plane transmitting array, including:

[0017] Calculate each group of pixels to the two-dimensional plane receiving array separately , and Distance of direction , and ,in , , , , and They are the first The coordinate vectors corresponding to the three coordinate axes of the array elements are based on , and , the distance vector of each group of pixels from the two-dimensional plane receiving array is calculated according to the following formula: :

[0018] ;

[0019] Calculate each group of pixels to the two-dimensional plane emission array separately , and Distance of direction , and ,in , , , , and They are the first The coordinate vectors corresponding to the three coordinate axes of the array elements are based on , and , the distance vector of each group of pixels from the two-dimensional plane emission array is calculated according to the following formula: :

[0020] ;

[0021] according to and Calculate the delay vector required for the signal to be beamformed ,in is the speed of sound in water.

[0022] Preferably, the step of calculating a sampling index according to the sampling rate and the delay vector of the two-dimensional plane receiving array, and calculating an interpolation coefficient vector according to the sampling index comprises:

[0023] According to the sampling rate of the two-dimensional plane receiving array and delay vector Calculate the sampling index corresponding to the delay ;

[0024] right Round to get an array ,calculate and The difference , then the interpolation coefficient vector is , superscript is transposed.

[0025] Preferably, the method further comprises: after calculating the sampling index, determining whether the sampling index size satisfies the condition that it is greater than zero and less than or equal to the number of sampling points of a single channel of the two-dimensional plane receiving array. If the threshold range is not met, the corresponding delay vector is excluded to update the delay vector. .

[0026] Preferably, the calculating of the block sparse matrix corresponding to each group of pixels according to the interpolation coefficient vector and the delay vector includes:

[0027] Construct a block sparse matrix to store the signal interpolation coefficients and the corresponding delays, and calculate the non-zero element values ​​in the block sparse matrix as ,in is the interpolation coefficient vector, is the delay vector, with superscript is transposed, and the number of rows of the block sparse matrix corresponding to each group of pixels is the number of pixels in each group, and the number of columns is ,in is the number of channels of the two-dimensional planar receiving array, is the number of sampling points of a single channel of the two-dimensional planar receiving array.

[0028] Preferably, combining the block sparse matrices corresponding to all groups of pixel points into an overall sparse matrix includes:

[0029] The block sparse matrices corresponding to all groups of pixels are stacked and combined in the row direction to obtain the overall sparse matrix , The dimension of ,in is the total number of pixels, The number of columns for each block sparse matrix.

[0030] Preferably, the step of acquiring the time domain multi-channel signal emitted by the two-dimensional planar transmit array and expanding it into a one-dimensional arrangement, and calculating the one-dimensional matrix according to the overall sparse matrix and the one-dimensionally arranged time domain multi-channel signal, comprises:

[0031] Acquire the time domain multi-channel signal emitted by the two-dimensional planar transmit array, expand the two-dimensional time domain multi-channel signal to one dimension in a row-first manner, and perform beam domain signal calculation ,in is the calculated one-dimensional matrix, is the overall sparse matrix, is the acquired time domain multi-channel signal.

[0032] Preferably, the step of reconstructing the one-dimensional matrix into a three-dimensional space after performing a dimension transformation to complete the beamforming calculation of the time domain multi-channel signal includes:

[0033] The one-dimensional matrix is ​​reconstructed into a three-dimensional matrix through dimension transformation. The dimension of the three-dimensional matrix is ,in , and The three-dimensional imaging space is set as , and The number of imaging points in three directions.

[0034] To achieve the above-mentioned purpose of the invention, an embodiment of the present invention further provides a sparse matrix-based time-domain beamforming acceleration device, comprising: a three-dimensional imaging space block module, a block sparse matrix construction module, an overall sparse matrix construction module and a time-domain beamforming calculation module;

[0035] The three-dimensional imaging space block module is used to expand the three-dimensional pixel point spatial distribution in the set three-dimensional imaging space into one dimension and then averagely block the pixel points to obtain a plurality of groups of pixel points;

[0036] The block sparse matrix construction module is used to calculate the delay vector according to the distance vectors from each group of pixels to the two-dimensional plane receiving array and the two-dimensional plane transmitting array respectively; calculate the sampling index according to the sampling rate and the delay vector of the two-dimensional plane receiving array, and calculate the interpolation coefficient vector according to the sampling index; calculate the block sparse matrix corresponding to each group of pixels according to the interpolation coefficient vector and the delay vector;

[0037] The overall sparse matrix construction module is used to combine the block sparse matrices corresponding to all groups of pixel points into an overall sparse matrix;

[0038] The time domain beamforming calculation module is used to obtain the time domain multi-channel signal transmitted by the two-dimensional planar transmitting array and expand it into a one-dimensional arrangement, calculate the one-dimensional matrix according to the overall sparse matrix and the one-dimensionally arranged time domain multi-channel signal, reconstruct the one-dimensional matrix into a three-dimensional space after dimension transformation, and complete the beamforming calculation of the time domain multi-channel signal.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] The present invention utilizes the advantages of sparse matrices in storage and indexing, and transforms the conventional time-domain beamforming insertion delay and superposition calculation into a point multiplication operation between a sparse matrix and an input signal vector, wherein the sparse matrix includes the spatial pixel point delay corresponding to parameters such as the imaging space size, imaging resolution, and sampling frequency, and the interpolation coefficients based on a predetermined interpolation method, thereby improving the efficiency of beamforming. After being stored and read, the sparse matrix can be used for multiple imaging, greatly reducing the single imaging time. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0042] Figure 1It is a flowchart of a sparse matrix-based time-domain beamforming acceleration method provided by an embodiment of the present invention;

[0043] Figure 2 is a schematic diagram of a sparse matrix construction method provided by an embodiment of the present invention;

[0044] Figure 3 It is a schematic diagram of performing time domain beamforming calculation on a received signal based on a sparse matrix provided by an embodiment of the present invention;

[0045] Figure 4 3 is a schematic diagram comparing imaging time between a conventional time-domain beamforming method and a sparse matrix-based time-domain beamforming method provided by an embodiment of the present invention;

[0046] Figure 5 It is a structural schematic diagram of a sparse matrix-based time-domain beamforming acceleration device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific implementation methods described herein are only used to explain the present invention and do not limit the protection scope of the present invention.

[0048] The inventive concept of the present invention is: in view of the problem that the conventional time-domain beamforming algorithm in the prior art has low computational efficiency and leads to poor real-time imaging speed, the embodiment of the present invention provides a time-domain beamforming acceleration method and device based on a sparse matrix, constructs an overall sparse matrix through three-dimensional imaging space blocking, delay calculation and signal interpolation coefficient calculation, and optimizes the time-domain beamforming technology used in three-dimensional sonar. By utilizing the advantages of the overall sparse matrix in storage and indexing, the overall sparse matrix is ​​multiplied with the input time-domain multi-channel signal vector, thereby improving the efficiency of time-domain beamforming, and the overall sparse matrix can be used for multiple imaging after being stored and read, which greatly reduces the single imaging time.

[0049] Figure 1 FIG. 1 is a flow chart of a sparse matrix-based time domain beamforming acceleration method provided by an embodiment of the present invention. Figure 1 As shown, the embodiment provides a sparse matrix-based time-domain beamforming acceleration method, comprising the following steps:

[0050] Step 1, such as Figure 2 As shown in (a), the three-dimensional imaging space is set in , and The resolution and imaging range of the direction are obtained respectively. , and The number of imaging points in the three directions are , and , the total number of pixels is .

[0051] Step 2, such as Figure 2 As shown in (b) in the figure, the three-dimensional space points are flattened to a one-dimensional arrangement along the row direction. Pixels are divided into Group, that is, the average per group pixels, each group of pixels corresponds to axis, Axis and The axis coordinates form the coordinate vector , and .

[0052] Step 3: Calculate the distance from each pixel group to the two-dimensional plane receiving array (i.e. signal receiving device) , and Distance of direction , and ,in , , , , and They are the first The coordinate vectors corresponding to the three coordinate axes of the array elements are based on , and , the distance vector of each group of pixels from the two-dimensional plane receiving array is calculated according to the following formula: :

[0053] .

[0054] Step 4: Calculate the distance from each group of pixels to the two-dimensional plane transmission array (i.e., signal transmission device) , and Distance of direction , and ,in , , , , and They are the first The coordinate vectors corresponding to the three coordinate axes of the array elements are based on , and , the distance vector of each group of pixels from the two-dimensional plane emission array is calculated according to the following formula: :

[0055] .

[0056] Step 5, according to and Calculate the delay vector required for the signal to be beamformed ,in is the speed of sound in water.

[0057] Step 6: Based on the sampling rate of the two-dimensional plane receiving array Calculate the sampling index corresponding to the delay ,in is the delay vector. After calculating the sampling index, determine if the sampling index size satisfies the requirement that it is greater than zero and less than or equal to the number of sampling points of a single channel of the two-dimensional plane receiving array. If the threshold range is not met, the corresponding delay vector is excluded to update the delay vector. .

[0058] Step 7, calculate the coefficients required for linear interpolation of the signal. Round to get an array ,calculate The next smaller integer The difference , then the required interpolation coefficient vector is , superscript is transposed.

[0059] Step 8: Construct a block sparse matrix according to the coo format to store the signal interpolation coefficients and the corresponding delays, such as Figure 2 As shown in (c), the number of rows of the block sparse matrix corresponding to each group of pixels The number of pixels in each group , the number of columns is ,in is the number of channels of the two-dimensional planar receiving array, is the number of sampling points of a single channel of the two-dimensional plane receiving array. Since the interpolation coefficient vector There are only two non-zero elements in each row, so each row in the block sparse matrix has two non-zero element values. The non-zero element values ​​in the block sparse matrix are ,in is the interpolation coefficient vector, is the delay vector, with superscript is transposed, and the row and column indices corresponding to the non-zero element values ​​are The row and column indices corresponding to the valid values.

[0060] Step 9, repeat steps 3 to 8 until the construction of the block sparse matrices of all groups in step 2 is completed. The block sparse matrices corresponding to all groups of pixels are stacked and combined in the row direction to obtain the following Figure 3 The overall sparse matrix shown in (a) is , The dimension of ,in is the total number of pixels, The number of columns for each block sparse matrix.

[0061] Step 10: For a single imaging operation, directly use the overall sparse matrix obtained based on step 9 to perform the next operation. For multiple imaging operations, the overall sparse matrix generated in step 9 is Storing in CSR format reduces the storage space of time delay and interpolation coefficient information, and can be read when imaging calculation is required.

[0062] Step 11, such as Figure 3 (a) Figure 3 (b) and Figure 3 As shown in (c), the time domain beamforming signal calculation is performed. The time domain multi-channel signal transmitted by the two-dimensional planar transmit array is obtained. , expand the two-dimensional time domain multi-channel signal to one dimension in a row-first manner, and perform beam domain signal calculation ,in is the calculated one-dimensional matrix. Since a sparse matrix is ​​used, its index format can speed up the dot multiplication process.

[0063] Step 12, such as Figure 3 As shown in (d) in the figure, the one-dimensional matrix Through dimensional transformation, it is reconstructed into a three-dimensional matrix. The dimension of the three-dimensional matrix is , so that subsequent image display processing can be carried out.

[0064] In the embodiment, Figure 4 As shown, when the two-dimensional plane receiving array is 16×16 array elements, , and When the axial imaging range is between (-1m, 1m), (-1m, 1m), and (4.5m, 5.5m), the sparse matrix-based time-domain beamforming acceleration method provided in an embodiment of the present invention is used to perform beamforming calculations (computer parameters are: dual-core Intel(R) Xeon(R) Silver4210 CPU processor, main frequency 2.19 GHz, memory size 64G), which has obvious speed advantages over conventional time-domain beamforming calculations.

[0065] In summary, the sparse matrix-based time-domain beamforming acceleration method provided in the embodiment of the present invention utilizes the advantages of sparse matrices in storage and indexing to transform the conventional time-domain beamforming insertion delay and superposition calculation into a point multiplication operation between a sparse matrix and an input signal vector, wherein the sparse matrix contains spatial pixel point delays and interpolation coefficients based on a predetermined interpolation method corresponding to parameters such as the imaging space size, imaging resolution, and sampling frequency, thereby improving the efficiency of beamforming. After storage and reading, multiple imaging can be performed, greatly reducing the single imaging time, thereby providing more powerful technical support for the fields of marine scientific research, underwater resource development, and underwater safety monitoring.

[0066] Based on the same inventive concept, Figure 5 As shown, an embodiment of the present invention further provides a sparse matrix-based time-domain beamforming acceleration device 500, comprising: a three-dimensional imaging space blocking module 510, a block sparse matrix construction module 520, an overall sparse matrix construction module 530 and a time-domain beamforming calculation module 540.

[0067] The three-dimensional imaging space blocking module 510 is used to expand the three-dimensional pixel point spatial distribution in the set three-dimensional imaging space into one dimension and then averagely block the pixel points to obtain a plurality of groups of pixel points.

[0068] The block sparse matrix construction module 520 is used to calculate the delay vector according to the distance vector from each group of pixel points to the two-dimensional plane receiving array and the two-dimensional plane transmitting array respectively; calculate the sampling index according to the sampling rate and delay vector of the two-dimensional plane receiving array, and calculate the interpolation coefficient vector according to the sampling index; calculate the block sparse matrix corresponding to each group of pixel points according to the interpolation coefficient vector and the delay vector.

[0069] The overall sparse matrix construction module 530 is used to combine the block sparse matrices corresponding to all groups of pixel points into an overall sparse matrix.

[0070] The time domain beamforming calculation module 540 is used to obtain the time domain multi-channel signal transmitted by the two-dimensional planar transmitting array and expand it into a one-dimensional arrangement, read the overall sparse matrix and the one-dimensionally arranged time domain multi-channel signal to calculate to obtain a one-dimensional matrix, transform the one-dimensional matrix and reconstruct it into a three-dimensional space to complete the beamforming calculation of the time domain multi-channel signal.

[0071] It should be noted that the sparse matrix-based time-domain beamforming acceleration device provided in the above embodiment belongs to the same inventive concept as the sparse matrix-based time-domain beamforming acceleration method. The specific implementation process is detailed in the embodiment of the sparse matrix-based time-domain beamforming acceleration method, which will not be repeated here.

[0072] The specific implementation methods described above provide a detailed description of the technical solutions and beneficial effects of the present invention. It should be understood that the above is only the most preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, supplements and equivalent substitutions made within the scope of the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A time-domain beamforming acceleration method based on sparse matrix, characterized in that: The following steps are involved: Expanding the spatial distribution of three-dimensional pixel points in the set three-dimensional imaging space to one dimension, and then evenly dividing the pixel points into blocks to obtain a plurality of groups of pixel points; A delay vector is calculated based on the distance vectors from each group of pixels to the two-dimensional plane receiving array and the two-dimensional plane transmitting array; a sampling index is calculated based on the sampling rate of the two-dimensional plane receiving array and the delay vector, and an interpolation coefficient vector is calculated based on the sampling index; a block sparse matrix corresponding to each group of pixels is calculated based on the interpolation coefficient vector and the delay vector; Combine the block sparse matrices corresponding to all groups of pixels into an overall sparse matrix; The time domain multi-channel signal transmitted by the two-dimensional planar transmitting array is obtained and expanded to a one-dimensional arrangement, and a one-dimensional matrix is ​​obtained by calculating the time domain multi-channel signal according to the overall sparse matrix and the one-dimensionally arranged time domain multi-channel signal, including: obtaining the time domain multi-channel signal transmitted by the two-dimensional planar transmitting array, expanding the two-dimensional time domain multi-channel signal to one dimension in a row-first manner, and performing beam domain signal calculation ,in is the calculated one-dimensional matrix, is the overall sparse matrix, To obtain the time domain multi-channel signal; then the one-dimensional matrix is ​​transformed into a three-dimensional space to complete the beamforming calculation of the time domain multi-channel signal.

2. The time-domain beamforming acceleration method based on sparse matrix according to claim 1, characterized in that: The three-dimensional pixel point spatial distribution in the set three-dimensional imaging space is expanded to one dimension and then the pixel points are evenly divided into blocks to obtain a plurality of groups of pixel points, including: Set the three-dimensional imaging space in , and The resolution and imaging range of the direction are obtained respectively. , and The number of imaging points in the three directions are , and , the total number of pixels is ; Flatten the three-dimensional space points to a one-dimensional arrangement along the row direction. Pixels are divided into Group, that is, the average per group pixels, each group of pixels corresponds to axis, Axis and The axis coordinates form the coordinate vector , and .

3. The time-domain beamforming acceleration method based on sparse matrix according to claim 2, characterized in that: The delay vector is obtained by calculating the distance vectors from each group of pixels to the two-dimensional plane receiving array and the two-dimensional plane transmitting array, including: Calculate each group of pixels to the two-dimensional plane receiving array separately , and Distance of direction , and ,in , , , , and They are the first The coordinate vectors corresponding to the three coordinate axes of the array elements are based on , and , the distance vector of each group of pixels from the two-dimensional plane receiving array is calculated according to the following formula: : ； Calculate each group of pixels to the two-dimensional plane emission array separately , and Distance of direction , and ,in , , , , and They are the first The coordinate vectors corresponding to the three coordinate axes of the array elements are based on , and , the distance vector of each group of pixels from the two-dimensional plane emission array is calculated according to the following formula: : ； according to and Calculate the delay vector required for the signal to be beamformed ,in is the speed of sound in water.

4. The sparse matrix-based time-domain beamforming acceleration method according to claim 1, characterized in that: The step of calculating a sampling index based on the sampling rate and the delay vector of the two-dimensional plane receiving array and calculating an interpolation coefficient vector based on the sampling index includes: According to the sampling rate of the two-dimensional plane receiving array and delay vector Calculate the sampling index corresponding to the delay ; right Round to get an array ,calculate and The difference , then the interpolation coefficient vector is , superscript is transposed.

5. The time-domain beamforming acceleration method based on sparse matrix according to claim 4, characterized in that: The method further comprises: after calculating the sampling index, determining whether the sampling index size satisfies the requirement that it is greater than zero and less than or equal to the number of sampling points of a single channel of the two-dimensional plane receiving array. If the threshold range is not met, the corresponding delay vector is excluded to update the delay vector. .

6. The time-domain beamforming acceleration method based on sparse matrix according to claim 1, characterized in that: The step of calculating the block sparse matrix corresponding to each group of pixels according to the interpolation coefficient vector and the delay vector includes: Construct a block sparse matrix to store the signal interpolation coefficients and the corresponding delays, and calculate the non-zero element values ​​in the block sparse matrix as ,in is the interpolation coefficient vector, is the delay vector, with superscript is transposed, and the number of rows of the block sparse matrix corresponding to each group of pixels is the number of pixels in each group, and the number of columns is ,in is the number of channels of the two-dimensional planar receiving array, is the number of sampling points of a single channel of the two-dimensional planar receiving array.

7. The time-domain beamforming acceleration method based on sparse matrix according to claim 1, characterized in that: The step of combining the block sparse matrices corresponding to all groups of pixel points into an overall sparse matrix includes: The block sparse matrices corresponding to all groups of pixels are stacked and combined in the row direction to obtain the overall sparse matrix , The dimension of ,in is the total number of pixels, The number of columns for each block sparse matrix.

8. The time-domain beamforming acceleration method based on sparse matrix according to claim 1, characterized in that: The one-dimensional matrix is ​​transformed into a three-dimensional space and then reconstructed to complete the beamforming calculation of the time domain multi-channel signal, including: The one-dimensional matrix is ​​reconstructed into a three-dimensional matrix through dimension transformation. The dimension of the three-dimensional matrix is ,in , and The three-dimensional imaging space is set as , and The number of imaging points in three directions.

9. A time-domain beamforming acceleration device based on sparse matrix, characterized in that: include: Three-dimensional imaging space block module, block sparse matrix building module, overall sparse matrix building module and time domain beamforming calculation module; The three-dimensional imaging space block module is used to expand the three-dimensional pixel point spatial distribution in the set three-dimensional imaging space into one dimension and then averagely block the pixel points to obtain a plurality of groups of pixel points; The block sparse matrix construction module is used to calculate the delay vector according to the distance vectors from each group of pixels to the two-dimensional plane receiving array and the two-dimensional plane transmitting array respectively; calculate the sampling index according to the sampling rate and the delay vector of the two-dimensional plane receiving array, and calculate the interpolation coefficient vector according to the sampling index; calculate the block sparse matrix corresponding to each group of pixels according to the interpolation coefficient vector and the delay vector; The overall sparse matrix construction module is used to combine the block sparse matrices corresponding to all groups of pixel points into an overall sparse matrix; The time domain beamforming calculation module is used to obtain the time domain multi-channel signal transmitted by the two-dimensional planar transmitting array and expand it into a one-dimensional arrangement, and calculate the one-dimensional matrix according to the overall sparse matrix and the one-dimensionally arranged time domain multi-channel signal, including: obtaining the time domain multi-channel signal transmitted by the two-dimensional planar transmitting array, expanding the two-dimensional time domain multi-channel signal into one dimension in a row-first manner, and performing beam domain signal calculation ,in is the calculated one-dimensional matrix, is the overall sparse matrix, To obtain the time domain multi-channel signal; then the one-dimensional matrix is ​​transformed into a three-dimensional space to complete the beamforming calculation of the time domain multi-channel signal.

Images