A method for scene holographic imaging based on ultrasound

Abstract

The present invention discloses a method for scene holographic imaging based on ultrasound, which uses a plurality of ultrasound transmitting probes arranged in a two-dimensional array to form a transmitting array, and transmits ultrasound signals in turn in time division, and receives them after being reflected by an object by a receiving array arranged in a two-dimensional array. According to the delay of the reflected signal, the position of the reconstruction surface is determined, and the echo signal from the reflection point on the specific reconstruction surface to each transmitting probe is solved by a compressed sensing reconstruction algorithm to solve the sparse reconstruction optimization problem, and obtain multiple sets of optimized target reflection point positions and signal strength data, and obtain target point cloud images. And by changing the position and superimposing multiple sets of target reflection point positions and signal strength data, point cloud data is obtained, and the above point cloud data is fitted to obtain the position and shape of the target, forming a holographic imaging of the target, and the method can realize ultrasonic holographic imaging of near-field targets.

Description

Technical Field

[0001] The present invention relates to an ultrasonic imaging method, and more particularly, to an ultrasonic-based scene holographic imaging method. Background Art

[0002] Ultrasonic positioning technology calculates the position of the target by measuring the arrival time (TOA) or time difference of arrival (TDOA) of the ultrasonic wave from the transmitter to the receiver. Compared with other positioning technologies, ultrasonic positioning has the following advantages: (1) Ultrasonic wave propagation speed is low, and it can directly track and locate targets at a short distance, with high resolution; (2) Ultrasonic wave is insensitive to external interference and can be used in harsh environments; (3) Ultrasonic wave is easy to transmit in a directional manner, has good directionality, and its intensity is easy to control; (4) Ultrasonic sensor has a simple structure, high cost performance, and is easy to miniaturize and integrate.

[0003] Ultrasonic imaging is currently used mainly in medical imaging and metal flaw detection. Its principle is to use ultrasonic sound beams to scan objects and obtain the internal structure of objects by receiving and processing reflected signals. However, in near-field imaging, near-field targets have rich position reference information, and the resolution of beamforming methods is difficult to meet the requirements. Summary of the invention

[0004] The present invention overcomes the deficiencies of the prior art and provides an ultrasound-based scene holographic imaging method, which can realize ultrasonic holographic imaging of near-field targets.

[0005] A scene holographic imaging method based on ultrasound, specifically comprising the following steps:

[0006] S1: A transmitting array is formed by using multiple ultrasonic transmitting probes arranged in a two-dimensional square array, which transmit ultrasonic signals in turn at different times. After being reflected by an object, the ultrasonic signals are received by a receiving array arranged in a two-dimensional square array.

[0007] S2: After receiving the reflected signal, the receiving element determines the position of the reconstruction plane according to the delay of the reflected signal;

[0008] S3: For the echo signals from the reflection points on the specific reconstruction surface to each transmitting probe, the sparse reconstruction optimization problem is solved by the compressed sensing reconstruction algorithm, and multiple sets of optimized target reflection point positions and signal strength data are obtained;

[0009] S4: changing the positions of the transmitting array and the receiving array, repeating the above steps S1 to S3, and reconstructing on different reconstruction planes to obtain multiple sets of target reflection point positions and signal strength data;

[0010] S5: superimpose the target reflection point position and signal strength data obtained from each optimization, and then set a threshold. The signal strength higher than the threshold is determined as a real reflection point. The above real reflection points constitute the point cloud data of the scene.

[0011] S6: Fit the point cloud data of the scene composed of real reflection points to obtain the position and shape of the target, and form a holographic imaging of the target.

[0012] This method uses a plurality of ultrasonic transmitting probes arranged in a two-dimensional square array to form a transmitting array, and transmits ultrasonic signals in turn in time division. After being reflected by the object, the receiving array arranged in a two-dimensional square array receives it. According to the delay of the reflected signal, the position of the reconstruction surface is determined. For the echo signal from the reflection point on the specific reconstruction surface to each transmitting probe, the compressed sensing reconstruction algorithm is used to solve the sparse reconstruction optimization problem, and the optimized multiple groups of target reflection point position and signal strength data are obtained, and the target point cloud image is obtained. And by changing the position and superimposing the multiple groups of target reflection point position and signal strength data, the point cloud data is obtained, and the above point cloud data is fitted to obtain the position and shape of the target, forming a holographic imaging of the target.

[0013] A further technical solution is that S3 is specifically:

[0014] S31: Calculate the distance between a grid point pi on the reconstruction surface and the receiving probe on the measurement surface, obtain the array manifold vector corresponding to the grid point, and further obtain the relationship between the array manifold vector matrix and the receiving array element strength matrix and the error matrix;

[0015] Y=AX+E1

[0016] in:

[0017] A is the array manifold vector matrix, specifically A=[a(1),a(2),...,a(i),...,a(N)], where a(i) is the array manifold vector;

[0018] X is the signal strength matrix of the reconstruction points, specifically X = [s(1), s(2), ..., s(k), ..., s(K)];

[0019] s i (k) is the signal strength at the i-th reconstruction point at time k, specifically s(k) = (s1(k), s2(k), ..., s i (k),...,s N (k)) T , where T is the period of the ultrasonic wave;

[0020] Y is the receiving element strength matrix, specifically Y = [r(1), r(2), ..., r(k), ..., r(K)];

[0021] r l (k) is the signal strength of the lth receiving element k at the moment, specifically r(k) = (r1(k), r2(k), ..., r l (k),...,r M (k)) T , where T is the period of the ultrasonic wave;

[0022] E1 is the error matrix.

[0023] S32: Use the measurement matrix M to compress the information of the receiving array element strength matrix Y, and obtain the compressed matrix Z. Since the number of receiving array elements is large and the number of sampling grid points on the reconstruction surface is small, information compression can effectively reduce the amount of calculation. The compression equation is as follows:

[0024] Z=MY=MAX+E2=HX+E2

[0025] in:

[0026] Z is a compressed matrix, which has the same number of columns as the receiving element strength matrix Y, but has a smaller number of rows;

[0027] M is the measurement matrix, which satisfies the restricted isometry property (RIP);

[0028] E2 is the error matrix, E2 = ME1, where E1 is the error matrix;

[0029] H is the sensing matrix, H=MA, where A is the array manifold vector matrix.

[0030] S33: Construct a sparse reconstruction optimization model for the compressed matrix Z:

[0031]

[0032] in:

[0033] ||·|| F represents the F norm of the matrix;

[0034] ||·|| 0,2 l represents the matrix 0,2 norm, that is, first find the 2 norm in columns, then find the 0 norm in rows;

[0035] Nr is a set positive integer threshold.

[0036] In the above sparse reconstruction problem, since the reconstruction is for a three-dimensional target, each reconstructed surface can only reflect part of the target's reflected signal, and the reconstruction error will be relatively large. 0,2The norm reflects the row sparseness of the matrix X. Since each row of X corresponds to a reflected signal, the reconstruction error needs to be taken as the optimization target.

[0037] The model can be viewed as solving a multiple-measurement vector (MMV) problem, where l 0,2 The norm characterizes sparsity.

[0038] Since the l0 norm is an NP-hard problem, a continuous function is needed to approximate the l0 norm. 1,2 Norm or weighted l 1,2 The norm rewrites the model, and the rewritten model is:

[0039]

[0040] S34: Using the zero-attraction adaptive filtering algorithm to solve the rewritten formula, multiple sets of optimized target reflection point positions and signal strength data are obtained, that is, the target point cloud image at position 1 is obtained.

[0041] A further technical solution is that, in S31, the array manifold vector is calculated as follows:

[0042]

[0043] in:

[0044] a(i) is the array manifold vector;

[0045] f(·) is the ultrasonic attenuation function;

[0046] ril is the distance from the i-th (i=1,2,…,N) reconstruction point to the l-th (l=1,2,…,M) array element;

[0047] c is the speed of sound in air;

[0048] T is the period of the ultrasonic wave.

[0049] A further technical solution is that the array composed of the plurality of transmitting probes is composed of 5 transmitting probes, which are distributed as follows: 4 at the edge and 1 at the center, and the 4 transmitting probes at the edge are inclined toward the center at 15° to 30°. The transmitting array arranged in this way achieves a balance between efficiency and accuracy.

[0050] A further technical solution is that the transmitting probe transmits ultrasound in a wide beam scanning manner, which can increase the detection speed.

[0051] Compared with the prior art, the present invention effectively utilizes a large amount of receiving array element information to establish an optimization model for near-field acoustic holographic imaging using multi-angle scanning information, thereby overcoming the problem that the resolution of the beamforming method is difficult to meet the requirements of near-field imaging. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 A schematic diagram of a flow chart of a scene holographic imaging method based on ultrasound;

[0053] Figure 2 To reconstruct the scene and move the location diagram;

[0054] Figure 3 This is a diagram showing the arrangement of the transmitting probe and receiving array elements;

[0055] Figure 4 Schematic diagram of the reconstruction surface position;

[0056] Figure 5 A schematic diagram of a reconstruction surface and a receiving surface of a scene holographic imaging method based on ultrasound;

[0057] Figure 6 A holographic image of the target. DETAILED DESCRIPTION

[0058] 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 embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0059] Reference Figure 1 , a scene holographic imaging method based on ultrasound, specifically comprising the following steps:

[0060] First, build a specific imaging scene, such as Figure 2 As shown, two standard threaded steel bars with a diameter of 6 mm are arranged in parallel, and then a device equipped with an ultrasonic transmitting probe and a receiving array element is placed 3 meters away from the plane where the two standard threaded steel bars are located.

[0061] S1: A transmitting array is composed of multiple ultrasonic transmitting probes arranged in a two-dimensional square array. Ultrasonic signals are transmitted in turn at different times, and then received by a receiving array after being reflected by an object.

[0062] like Figure 3 As shown, the ultrasonic transmitting array includes 5 transmitting probes, the central frequency of the transmitting probes is 40kHz, and the diameter is 11mm (the diameter of the piezoelectric piece is 8mm).

[0063] The transmitting probes are distributed in a two-dimensional square array at the edge and center of the array. The distance between the central transmitting probe and other transmitting probes is 200 mm. The four transmitting probes at the edge are tilted 15° toward the center and use a wide beam scanning with a scanning range of 60°.

[0064] The five transmitting probes use a time-sharing method for detection. The transmission interval is 2ms, the detection distance is 3m, and the ultrasonic round-trip delay is about 2ms. The five transmitting probes use a time-sharing method for detection, and it takes about 10ms to collect one frame.

[0065] The receiving array adopts a two-dimensional square array with an array element spacing of 22.7 mm and an array element size of 16×16;

[0066] S2: After receiving the reflected signal, the receiving element determines the position of the reconstruction plane according to the delay of the reflected signal, such as Figure 4 As shown, the specific position is on the steel bar diameter plane and on the plane 2mm away from the steel bar diameter plane;

[0067] S3: For the echo signals from the reflection points on the specific reconstruction surface to each transmitting probe, the sparse reconstruction optimization problem is solved by the compressed sensing reconstruction algorithm, and multiple sets of optimized target reflection point positions and signal strength data are obtained;

[0068] like Figure 5 As shown in the figure, the measurement surface corresponds to the plane where the receiving array is located, the circle represents the receiving probe, there are two reconstruction surfaces, and the grid represents the potential target surface reflection points. The number and layout of the reconstruction surface grid points are different from those of the receiving array.

[0069] The specific steps are:

[0070] S31: Calculate the distance between a grid point pi on the reconstruction surface and the receiving probe on the measurement surface to obtain the array manifold vector corresponding to the grid point.

[0071] The specific calculation formula is as follows:

[0072]

[0073] in:

[0074] a(i) is the array manifold vector;

[0075] f(·) is the ultrasonic attenuation function;

[0076] ril is the distance from the i-th (i=1,2,…,N) reconstruction point to the l-th (l=1,2,…,M) array element;

[0077] c is the speed of sound in air;

[0078] T is the period of the ultrasonic wave.

[0079] After calculating the array manifold vector, the relationship between the array manifold vector matrix and the receiving array element strength matrix and error matrix is ​​obtained:

[0080] Y=AX+E1

[0081] in:

[0082] A is the array manifold vector matrix, specifically A=[a(1),a(2),...,a(i),...,a(N)];

[0083] X is the signal strength matrix of the reconstruction points, specifically X = [s(1), s(2), ..., s(k), ..., s(K)];

[0084] s i (k) is the signal strength at the i-th reconstruction point at time k, specifically s(k) = (s1(k), s2(k), ..., s i (k),...,s N (k)) T , where T is the period of the ultrasonic wave;

[0085] Y is the receiving element strength matrix, specifically Y = [r(1), r(2), ..., r(k), ..., r(K)];

[0086] r l (k) is the signal strength of the lth receiving element k at the moment, specifically r(k) = (r1(k), r2(k), ..., r l (k),...,r M (k)) T , where T is the period of the ultrasonic wave;

[0087] E1 is the error matrix.

[0088] S32: Use the measurement matrix M to compress the information of the receiving array element strength matrix Y, and obtain the compressed matrix Z. The compression equation is as follows:

[0089] Z=MY=MAX+E2=HX+E2

[0090] in:

[0091] Z is a compressed matrix, which has the same number of columns as the receiving element strength matrix Y, but has a smaller number of rows;

[0092] M is the measurement matrix, which satisfies the restricted isometry property (RIP);

[0093] E2 is the error matrix, E2 = ME1, where E1 is the error matrix;

[0094] H is the sensing matrix, H=MA, where A is the array manifold vector matrix.

[0095] S33: Construct a sparse reconstruction optimization model for the compressed matrix Z:

[0096]

[0097] in:

[0098] ||·|| F represents the F norm of the matrix;

[0099] ||·|| 0,2 l represents the matrix 0,2 norm;

[0100] Nr is a set positive integer threshold;

[0101] The sensing matrix H=MA, where A is the array manifold vector matrix.

[0102] Taking the reconstruction error as the optimization target, use l 1,2 norm to rewrite the model, and the rewritten model is:

[0103]

[0104] S34: Using the zero-attraction adaptive filtering algorithm to solve the rewritten formula, finally obtaining multiple sets of optimized target reflection point positions and signal strength data, that is, obtaining the target point cloud image at position 1.

[0105] S4: Change the position of the transmitting array and the receiving array, such as Figure 2 As shown, with the middle point of the line connecting the middle points of the two threaded steel bars as the center and 3m as the radius, the device equipped with the ultrasonic transmitting probe and the receiving array element is moved upward by 15° from the initial position 1 to position 2, and moved downward by 15° from the initial position 1 to position 3. After reaching the corresponding position, the above steps S1 to S3 are repeated, and reconstruction is performed on different reconstruction planes to obtain multiple sets of target reflection point positions and signal strength data;

[0106] S5: The signal strengths of the specific reflection points obtained from each optimization are superimposed, and a threshold N is set, where the value of N is half of the sum of the maximum and minimum signal strengths of the specific reflection points obtained from each optimization. The signal strengths higher than the threshold N are identified as real reflection points, and the real reflection points are combined to form the point cloud data of the scene.

[0107] S6: Fit the point cloud data of the scene composed of real reflection points to obtain the position and shape of the target and form a holographic image of the target, such as Figure 6 shown.

[0108] Although the present invention is described herein with reference to the illustrative embodiments of the present invention, it should be understood that those skilled in the art can design many other modifications and implementations, which will fall within the scope and spirit of the principles disclosed in the present application. More specifically, within the scope disclosed in the present application, multiple variations and improvements can be made to the components and / or layout of the subject combination layout. In addition to the variations and improvements made to the components and / or layout, other uses will also be apparent to those skilled in the art.

Claims

1. A method for scene holographic imaging based on ultrasound, characterized in that: Use the following steps: S1: A transmitting array is formed by using multiple ultrasonic transmitting probes arranged in a two-dimensional square array, which transmit ultrasonic signals in turn at different times. After being reflected by an object, the ultrasonic signals are received by a receiving array arranged in a two-dimensional square array. The transmitting array is composed of 5 transmitting probes, which are distributed as follows: 4 at the edge and 1 at the center, and the 4 transmitting probes at the edge are inclined toward the center at 15° to 30°; S2: After receiving the reflected signal, the receiving array element determines the position of the reconstruction plane according to the delay of the reflected signal; S3: For the echo signals from the reflection points on the specific reconstruction surface to each transmitting probe, the sparse reconstruction optimization problem is solved by the compressed sensing reconstruction algorithm, and multiple sets of optimized target reflection point positions and signal strength data are obtained; S4: changing the positions of the transmitting array and the receiving array, repeating the above steps S1 to S3, and reconstructing on different reconstruction planes to obtain multiple sets of target reflection point positions and signal strength data; S5: superimpose the target reflection point position and signal strength data obtained from each optimization, and then set a threshold. The signal strength higher than the threshold is determined as a real reflection point. The above real reflection points constitute the point cloud data of the scene. S6: Fit the point cloud data of the scene composed of real reflection points to obtain the position and shape of the target, and form a holographic imaging of the target.

2. The method for scene holographic imaging based on ultrasound according to claim 1, characterized in that: The S3 is specifically: S31: Calculate the distance between a grid point pi on the reconstruction surface and the receiving probe on the measurement surface, obtain the array manifold vector corresponding to the grid point, and further obtain the relationship between the array manifold vector matrix and the receiving array element strength matrix and the error matrix; Y=AX+E1 in: A is the array manifold vector matrix, specifically A=[a(1),a(2),...,a(i),...,a(N)], where a(i) is the array manifold vector; X is the signal strength matrix of the reconstruction points, specifically X = [s(1), s(2), ..., s(k), ..., s(K)]; s i (k) is the signal strength at the i-th reconstruction point at time k, specifically s(k) = (s1(k), s2(k), ..., s i (k),...,s N (k)) T , where T is the period of the ultrasonic wave; Y is the receiving element strength matrix, specifically Y = [r(1), r(2), ..., r(k), ..., r(K)]; r l (k) is the signal strength of the lth receiving element k at the moment, specifically r(k) = (r1(k), r2(k), ..., r l (k),...,r M (k)) T , where T is the period of the ultrasonic wave; E1 is the error matrix; S32: compressing the receiving array element strength matrix Y using the measurement matrix M, and obtaining a compressed matrix Z; Z=MY=MAX+E2=HX+E2 in: Z is the compressed matrix; M is the measurement matrix, which satisfies the restricted isometry property (RIP); E2 is the error matrix, E2=ME1, where E1 is the error matrix; H is the sensing matrix, H = MA, where A is the array manifold vector matrix; S33: construct a sparse reconstruction optimization model for the compressed matrix Z; in: ||·|| F represents the F norm of the matrix; ||·|| 0,2 l represents the matrix 0,2 norm; Nr is a set positive integer threshold; Taking the reconstruction error as the optimization target, we use l 1,2 Norm or weighted l 1,2 Norm rewrite model; S34: Using the zero-attraction adaptive filtering algorithm to solve the rewritten formula, multiple sets of optimized target reflection point positions and signal strength data are obtained, that is, the target point cloud image at position 1 is obtained.

3. The method for scene holographic imaging based on ultrasound according to claim 2, characterized in that: In S31, the array manifold vector is calculated as follows: in: a(i) is the array manifold vector; f(·) is the ultrasonic attenuation function; ril is the distance from the i-th (i=1,2,…,N) reconstruction point to the l-th (l=1,2,…,M) array element; c is the speed of sound in air; T is the period of the ultrasonic wave.

4. The method for scene holographic imaging based on ultrasound according to claim 1, characterized in that: The transmitting probe transmits ultrasound in a wide beam scanning manner.

Images