System and method for forming an image of ultrasonic reflectance

Abstract

The object image value for each location in the ultrasonic reflectance image is calculated by summing the image values ​​from images associated with selected intersections in different rows and columns within the array of ultrasonic transducers. The image value for each of these intersections is determined by transmitting an ultrasonic transmission signal from the ultrasonic transducer to a first subsampled selection of first sampling locations along the column of intersections, and at least partially simultaneously receiving reflections from a second subsampled selection of second sampling locations along the row of intersections. The image values ​​for different locations are calculated by combining the vertical resolution from the subsampled transmitted ultrasonic signal and the horizontal resolution from the subsampled received ultrasonic signal. The range of subsampling locations for each intersection has a large overlap with the subsampling range for adjacent intersections, and the subsampling locations for different intersections exist at least partially scattered amongst each other within that overlap. The extension of subsampling locations across a large overlapping range ensures high resolution in the intersection image, and the use of different sampling points within that overlap reduces artifacts in the summed image.

Description

Technical Field

[0001] The present invention relates to ultrasonic imaging using an array of ultrasonic transducers.

Background Art

[0002] A large-area two-dimensional array of piezoelectric columns on a substrate for use as an ultrasonic transducer is described in a poster presentation entitled "Flexible ultrasound transducer array concept scalable to large areas: first 128 element 7 MHz linear array" in IEEE IUS 2021 by van Neer et al. The technology for manufacturing such an array is described in a poster presentation entitled "Technology for Large Area and Flexible ultrasound patches" in IEEE IUS 2021 by van Peters et al. The array comprises small-sized uniform pillars (e.g., 40×40×80 micrometers) with a small pitch (e.g., 60 micrometers) of the pillars, and the array can have a large size (e.g., 150×150 mm, with more than 6 million pillars). The array can be used to form a three-dimensional image or one or more two-dimensional images of reflections from positions within an object space having a distance in a direction perpendicular to the array. A flexible substrate that makes the array flexible can be used. Due to the flexibility of the substrate, the array can be used as a wearable device on a human body. This enables acquisition of medical-purpose images during normal daily activities of a patient.

[0003] However, using large arrays requires considerable time to perform the measurements and calculations necessary for imaging. Unlike reading data from a memory array or optical images from an optical sensor array, sensing signals from individual receiving elements within an ultrasonic receiving array must be performed as a function of time. This may need to be repeated sequentially for many different signals to gather sufficient information for imaging, and the digitized signals from many individual receiving elements must be combined to obtain high three-dimensional resolution.

[0004] The considerable time required to acquire images limits its usefulness. For example, when the device is used as a wearable imaging device, it cannot be expected that the person will remain still during measurement. In addition, long processing times may be required to perform the necessary signal processing. While this can be reduced by using more sophisticated processing circuits, this may hinder wearable use. The time required for such measurements can be reduced to some extent by receiving ultrasound in parallel from rows at multiple locations within the array, and / or transmitting it in parallel from different locations. While more complex wiring than that of a simple array allows for a further increase in the amount of parallel processing, it is desirable to avoid excessive complexity.

[0005] Another method to reduce the time required to acquire images is to use spatial subsampling, for example, by processing only the signals received in a subset of locations within the array. Arrays with a pitch of 50–500 micrometers can produce much higher resolution than is required for many purposes, such as acquiring large-scale medical images, so the loss of resolution due to subsampling may not be a problem. However, spatial subsampling can lead to image artifacts such as aliasing, or the need to suppress such artifacts at the expense of resolution. [Overview of the project]

[0006] In particular, the objective is to provide efficient ultrasonic imaging using an ultrasonic transducer array with a large number of transducers.

[0007] The method described in claim 1 is provided for forming an object image of ultrasonic reflectance in object space using an array of rows and columns of ultrasonic transducers defining adjacent sensing surfaces in object space. In the method, the object image value for each position in the image is calculated by summing the image values ​​based on the intersection for the position determined for a plurality of intersection locations associated with each combination of rows and columns in the array. The image values ​​based on the intersection for the position in the image are determined using transmissions from subsampled locations (preferably in the column of each intersection location) from a range of positions that do not exist farther from the row of each intersection location than the column of each intersection location, to provide at least a composite directional sensitivity in the column direction of the array, and reception or resulting reflection from subsampled locations from a range of positions in the row, to provide a composite directional sensitivity in the row direction of the array. At least the ranges of adjacent intersections overlap with each other, for example, such that the range is 2, 3, 4 or more times the distance between the intersections. The sampling locations used for different intersections within the overlap are at least partially scattered with each other.

[0008] Image values ​​based on intersections calculated for different positions in object space can constitute a complete image as a function of those positions. However, although the intersection images used are spatial images of the same object (the same part of the object), the object images based on intersections determined for different intersection locations will be different because different subsamplings in both the row and column directions are used for the images based on different intersections. Subsampling can be very deep; for example, only one of 10 positions along a row may be used and offset by an average from the row. Subsampling based on random selection may also be used.

[0009] Deep subsampling saves computation time but introduces sampling artifacts. These sampling artifacts are mitigated by using the sum of image values ​​based on the intersections, combined with different sampling selections that are at least partially scattered from overlapping ranges for different intersections. In this way, sampling artifacts in images based on different intersections are at least partially compensated for by each other in the summed image.

[0010] Preferably, scattered subsampling locations are selected to avoid matches between subsampling locations and between locations at subsampled offsets from rows relative to neighboring intersection locations (e.g., between subsampling locations in columns). However, even with partial matches, subsampling artifacts in images based on different intersections will at least partially compensate for each other in the aggregated image. Deep subsampling makes it easier to avoid such matches.

[0011] Independent random selection of sampling locations for different intersection locations can result in at least partially scattered sampling locations. This can lead to some agreement between sampling locations for different intersection locations, although this rarely occurs with deep subsampling.

[0012] In embodiments, the first and second selections are aperiodic selections. This selection may be used to further reduce strongly visible subsampling artifacts. In further embodiments, the distance values ​​between consecutive sampling locations of the first sampling location and / or between consecutive sampling locations of the second sampling location are distributed over a predetermined range of distance values, which includes a plurality of distance values. Thus, strong local sampling density fluctuations can be prevented. The distance values ​​may be distributed similarly to a pseudo-random selection from a predetermined range of distance values. In further embodiments, the lower limit of the predetermined range of distance values ​​is at most half of the mean of the distance values.

[0013] In one embodiment, the intersections are located in a two-dimensional periodic grid of locations having periods smaller than the first and second ranges. This provides uniform intersection image determination. However, in other embodiments, non-periodically located intersections may be used to reduce artifacts.

[0014] In the embodiment, the mean subsample ratio of the first and second selections is at most half the density of intersections along the columns and / or rows of each intersection location. Thus, deep subsampling can be used. Normally, this causes strong artifacts, but the use of summation of intersection images based on different selections from overlapping ranges makes it possible to use such deep subsampling with less artifacts.

[0015] In some embodiments, the sum of image values ​​based on intersections relative to position may be a weighted sum. Therefore, for example, intersection images for intersections located far from the object spatial position where the image value is calculated may be given less weight, or no weight at all. This can be used to reduce noise.

[0016] In a further embodiment, the weighted sum defines a window of intersection locations in which the first and / or second ranges completely overlap with at least one of the first and / or second ranges among the intersection locations in the window. This allows for optimized artifact reduction. The least significant of the intersection locations in the window may be the intersection location closest to the object spatial position where the image value is calculated. Preferably, intersection locations whose ranges do not overlap with the least significant of the intersection locations are not weighted. Thus, noise can be reduced. [Brief explanation of the drawing]

[0017] These and other purposes and advantageous embodiments are illustrated in the detailed description of the exemplary embodiments with reference to the following figures.

[0018] [Figure 1] A top view of an ultrasound imaging device is shown. [Figure 2a] This shows the imaging system. [Figure 2b] The signal flow model is shown. [Figure 2c] This shows a matrix with a reduced number of elements. [Figure 3] This shows a linear excitation / reception pattern. [Figure 4] A side view of a wearable ultrasound imaging device is shown. [Figure 5] The circuit diagram of the array element is shown. [Figure 6] A flowchart for image conversion is shown. [Figure 7] A flowchart for image calculation is shown. [Figure 8] An embodiment having a receiver with multiple lines will be illustrated as an example.

[0019] Detailed description of exemplary embodiments FIG. 1 shows an example of an ultrasonic imaging device having a two-dimensional array 10 of ultrasonic transducer elements where, when viewed from a position in the z-direction perpendicular to the array 10, the array extends horizontally in the x-direction and y-direction. Imaging using the ultrasonic imaging device involves both ultrasonic transmission and ultrasonic reception by the array. Each ultrasonic transducer element can be an ultrasonic receiver, an ultrasonic transmitter, or both. Although the layout of the array may appear similar to that of an optical image sensor, it should be noted that ultrasonic imaging is more complex than optical imaging. In addition to the fact that the array is used for both transmission and reception, the signals at each array element must be processed as time-dependent ultrasonic signals.

[0020] The array 10 is located on a substrate, which can be a flexible substrate. The array 10 can have a rectangular shape with sides having a length of several hundred millimeters or more. At the same time, the array elements can be provided at a much smaller pitch of less than 100 micrometers (e.g., 60 micrometers), resulting in thousands of ultrasonic transducer elements along the sides of the array and millions of ultrasonic transducer elements within the array.

[0021] The ultrasonic imaging device can be used as a wearable ultrasonic imaging device, for example, to be worn on a human or animal body to capture a 3D image of the ultrasonic reflectivity within a depth range under the area of the body where the device is worn. The device can be used to acquire an image over the entire area, or an image over a much smaller sub-area, or both. For the purpose of acquiring an overview image over the entire area, the number of array elements is redundantly large. In principle, for example, deep sub-sampling at a rate of one-tenth in both the x-direction and y-direction may be sufficient for the overview image.

[0022] Array 10 comprises rows and columns of array elements 100 of uniform size. The ultrasonic imaging device is designed to generate ultrasonic signals from array elements 10 at selectable locations within the array, or from a selectable group of such locations, and to output electrical signals excited by reception of ultrasonic waves at selectable locations within the array, or at a selectable group of such locations.

[0023] Row selection circuit 12, and first column access circuit 14 and second column access circuit 16 are located adjacent to array 10 on a substrate. In the illustrated embodiment, first column access circuit 14 can be used for signals received from array elements, and second column access circuit 16 can be used to supply signals for transmission by array elements. Although an embodiment having two column access circuits 14, 16 is shown for illustrative purposes, it should be noted that one column access circuit combining the functions of first column access circuit 14 and second column access circuit 16 may be sufficient for both signal transmission and reception.

[0024] Array 10 may comprise conductive row lines 24 coupled to row selection circuit 12, and conductive column lines 23b coupled to column access circuits 14, 16. Row selection circuit 12 has a row address input 120 and an output coupled to row lines 24. Each row line 24 connects row selection circuit 12 to the selection input of array elements of each row of array 10. Column access circuits 14, 16 each have an input and an output coupled to column line 23b. Column access circuits 14, 16 have column address inputs 140, 160 for ultrasonic excitation signals and ultrasonic detection signals, a signal input 162, and a signal output 142. Each column line 23b connects column selection circuits 14, 16 to the signal connection of array elements in each column of array 10.

[0025] Figure 2a shows an imaging system comprising a processing system 40, a wearable ultrasonic imaging device 42 having an array as described in Figure 1, a timing circuit 43, a signal generator 44, and a detection circuit 45. Together with the row selection circuit 12 and the second column access circuit 16, the signal generator 44 forms an excitation circuit configured to induce ultrasonic emission from a selectable first ultrasonic transducer in the array. Together with the row selection circuit 12 and the first column access circuit 14, the detection circuit 45 forms a receiving circuit configured to receive an ultrasonic measurement signal from a selectable second ultrasonic transducer in the array. As can be noted, the excitation circuit and the receiving circuit may share some components.

[0026] The processing system 40 has an output coupled to the address input of the wearable ultrasound imaging device 42 in order to supply column addresses to the column access circuit of the wearable ultrasound imaging device 42 and row addresses to the row selection circuit of the wearable ultrasound imaging device 42.

[0027] Furthermore, the processing system 40 has a control output coupled to a timing circuit 43. The timing circuit 43 has outputs coupled to a signal generator 44 and a detection circuit 45. The generator 44 has an output coupled to the wearable ultrasound imaging device 42 to supply a transmit signal to the signal input of the first column access circuit of the wearable ultrasound imaging device 42. The detection circuit 45 has an input for receiving a receive signal from the first column access circuit of the wearable ultrasound imaging device 42.

[0028] During operation, the signal generator 44 generates signals for transmission by array elements in rows and columns of an array selected by the processing system 40. The detection circuit 45 detects signals received by array elements in rows and columns of an array selected by the processing system 40. The timing circuit 43 controls the signal generation time for transmission and the time reference for detection. The detection circuit 45 samples and digitizes the signals received from the array elements, and may preferably be configured to simultaneously sample and digitize signals received from multiple columns from the same row, with the time reference defining the relationship between the sampling time and the generation of the transmission signal. The detection circuit 45 has an output coupled to the processing circuit 40 to supply the sampled digitized signals.

[0029] Although the input to the detection circuit 45 is illustrated as a single line, it should be understood that the lines to the detection circuit 45 may represent a single signal input used to receive signals from individual selected column lines, or multiple signal inputs for receiving and detecting signals from multiple selected column lines in parallel. In one embodiment, the detection circuit 45 has a multiplex input for receiving received signals in parallel from all columns of the array 10. In another embodiment, the detection circuit 45 has a multiplex input for receiving received signals in parallel from a selectable subset of columns of the array 10.

[0030] Similarly, although the output of the signal generator 44 is illustrated as a single line that may represent a single signal output, in some embodiments, this line may represent multiple signal outputs when used to supply signals to individual selected lines.

[0031] The processing system 40 is programmed to control the operation of the imaging system. In operation under the control of the processing system 40, the processing system 40 calculates an image of ultrasonic reflectivity in object space, i.e., space containing points within a range of z-direction distances from the xy-plane of the array 10, which may be more commonly referred to as the sensing plane. Reflections can originate from within objects located in object space. If the wearable ultrasonic imaging device 42 is worn on the body of a person or animal, reflections originate from within the body at locations where the device is worn to a certain depth within the body.

[0032] The processing system 40 can be programmed to calculate a three-dimensional image of reflectance, or a two-dimensional image of reflectance in a cross-section of an object having a virtual plane in object space or a curved virtual surface in object space. The position in object space is denoted by "r" = (rx, ry, rz), and the location on the sensing surface is denoted by p = (px, py), in which case p = (px, py, 0) corresponds to object space coordinates. As used herein, a "reflectance image" is a set of image values ​​(or their Fourier transform) corresponding to reflectance at a position in object space, in the sense that the set of image values ​​provides information about the position dependence of the actual reflectance in object space, but not necessarily in terms of actual reflectance values ​​or without any distortion; that is, the amplitude scale or spatial scale may differ from the actual reflectance, and deviations due to the approximate nature of the image calculation are not excluded.

[0033] Signal flow for image computation Figure 2b shows a signal flow model for determining the image Im(r) from the reflectance from the object space. Image determination involves transmitting a signal from the array element 100, receiving the reflected signal in the array element, and calculating the resulting image value Im(r) from the received signal.

[0034] The excitation circuit is represented by a signal generator E1, a vector response filter V(j,r), and spatial sampling functions Sv(j) and Av. The receiving circuit is represented by spatial sampling functions Au and Su(j), a vector response filter U(j,r), and a signal detector D.

[0035] In Figure 2b, Av and Au represent sampling resulting from the fact that array 10 contains individual array elements. Sv and Su represent subsampling within these elements, which will be considered later. Av and Au can be absorbed by Sv and Su. Ignoring Sv and Su initially, the transmit signal for acquiring Im(r) can be expressed as follows: e0 = V(r)*s

[0036] Here, s is a timing control signal and can be expressed as a time-domain signal or a Fourier transform-domain signal (this dependency remains implicit), where "*" indicates convolution in the time domain or multiplication in the Fourier transform domain. "s" is used to allow distinction between different reception time delays and can be, for example, a pulse signal. Formally, V(r) is a vector having vector components for each of the transmitter elements in array 10, where each component may represent a time-dependent response function applied to s to obtain a signal for each array element, depending on the position "r" where the image value Im(r) is determined. "s" can also be selected depending on the position "r", but this is not necessary, and therefore, any dependency on "r", if present, is not shown.

[0037] Furthermore, the determination of Im(r) involves a similar convolution or product of the vector D*U(r) and the measured vector signal vs, as well as a scalar product. Im(r)=D(r)*U(r)*vs

[0038] In this specification, vs is formally and in principle a vector having components for each of the array elements, representing the received signal in the time domain or the Fourier transform domain. U(r) is a vector of time-dependent response functions. D is an operation applied to the scalar product U(r)*vs to select a distance by distinguishing signal values ​​with different time delays. In embodiments where s is a pulsed signal, D may correspond to selecting a value of U(r)*vs at a given time delay. If the signal e is more complex, D may be more complex accordingly. D may depend on the position r at which the image value Im(r) is determined. However, any dependence on r, if any, is not shown.

[0039] The reflectivity in object space relates the received signal vs. the transmitted signal. vs=R'*e0

[0040] Here, R' is a matrix having rows and columns, where both the number of rows and columns are equal to the number of transmitter and receiver elements in array 10 (i.e., not equal to the number of rows or columns in array 10), and "*" indicates convolution in the time domain or multiplication in the Fourier transform domain. The matrix R' can be modeled as follows: R' = integral of r' and r'' in physical space Tu(r')*R(r'-r'')*Tv(r'')

[0041] Here, r' and r'' are positions in object space, R represents the local reflectance (a continuous function) in object space, and Tv and Tu are vectors representing ultrasonic propagation from a location on the array to a location on the array, and ultrasonic propagation to a location on the array, respectively, via the local position in object space. Tv and Tu are functions of position r in object space and travel time. Formally, the product of Tu and Tv with R involves a convolution in object space (or Fourier space). However, if we can assume that the reflectance R is local and instantaneous, the product with Tv will be a simple product, or more precisely, the following: R' = integral of r' in physical space Tu(r')*R(r')Tv(r')

[0042] Therefore, in summary, the image value Im(r) calculated from vs is equal to the following: Im(r) = D*U(r)*R'*V(r)*s

[0043] In principle, vectors U(r) and V(r) are chosen so that together they compensate for the effects of Tu and Tv. That is, Im(r) = D*R(r)*s

[0044] Therefore, in principle, the calculated image value Im(r) of an image corresponds to the reflectance R(r) at position "r" in object space and can be determined by transmitting e0 and receiving vs. Preferably, U and V compensate for the difference in propagation travel time, represented by at least Tu and Tv. Thus, the image value approximates R(r). Optionally, U(r) and V(r) also compensate for the effect of distance on the signal amplitude.

[0045] In principle, the correspondence between R(r) and Im(r) applies to an infinite sensing surface with infinitely small array elements. In reality, both are finite. Consequently, Im(r) corresponds to a broadband pass-space filtered version of R, where some high frequencies are suppressed due to the limited size of the sensing surface.

[0046] As described above, vectors U(r) and V(r) are preferably selected, at least together, to compensate for the difference in propagation travel time represented by Tu and Tv. Preferably, U and V provide at least a delay time that changes in the opposite manner to the travel time represented by vectors Tu(r') and Tv(r'). The component Tu(r',p) of vector Tu(r') corresponds to the propagation from position r' in object space to location p of the receiving element. This is a function of p-r'. The travel time of the ray between r' and p is |p-r'| / c, where "c" is the speed of sound. U(r) may be selected such that the component U(r,p) of vector U(r) changes as -|rp| / c. In the time-dependent Fourier transform, this corresponds to the phase factors exp(if|p-r'| / c) and exp(-if|rp| / c), i 2 A similar relationship, =-1, also applies to V(r).

[0047] In this approximation, the dependence of the travel time difference (|rp|-|p-r'|) / c on p is given by p*(e(r)-e(r')) / c, where p*(e(r)-e(r')) is the scalar product of p and the orthogonal projections of the unit vectors e(r) and e(r') in the directions of r and r'. This approximation applies when the range of propagation distances |rp| and |r'-p| is small compared to the object distances |r| and |r'|. In this approximation, the Fourier transforms of the time dependence of the components U(r,p) and V(r,p) of the vectors U(r) and V(r) provide a phase factor exp(ik*p), where k=2*PI*fe(r) / c, and "f" is the time frequency. Similarly, in this approximation, the Fourier transforms of the time dependence of Tu and Tv include a phase factor exp(-ik*p).

[0048] Subsampling Subsampling may be used to reduce the time and memory space required for measurement and calculation. With respect to U and V, subsampling can be modeled as matrix multiplication in vector space [USu], [SvV] using diagonal matrices Su, Sv with coefficients equal to 1 or zero. Thus, the image values ​​acquired by transmission and reception can be represented by the response matrix RT as follows: Im(r)=D(r)*[U(r)Su / <su>]*RT*[SvV(r) / <sv>]*e

[0049] Here <su>This is the mean value of Su as a function of position, and is also called the sampling ratio of Su. Similarly, <su>This is the mean value of Sv as a function of position.

[0050] Subsampling introduces a deviation in the correspondence between the calculated image Im(r) and R, i.e., an artifact. As is well known, one-dimensional periodic subsampling results in aliasing, i.e., signal spectrum folding, where higher frequency components in the unsubsampled image become near-zero frequency components in the subsampled image. The image Im'(r) obtained by subsampling can be expressed by the following error term compared to the image obtainable without subsampling. Im'(r) = Im(r) + errU + errV + errUV Herein lies the following: errU=D*U(r)(Su / <su>The integral of r' in -I)*Tu(r')R(r')*s errV = D * R(r') * Tv(r') * (Sv / <sv>The integral of r' in -I)V(r)*s errUV = D * U(r)(Su / <su>I)*R(r’)*Tv(r’)(Sv / <sv>The integral of r' in -I)*V(r)*s

[0051] In the approximation where the dependence of the travel time difference changes as p*(e(r)-e(r')) / c, and in the Fourier transform of the time dependence, U(r)(Su / <su>-I)*Tuは、(Su / <su>-I) corresponds to the spatial Fourier transform, and similarly, Tv(r')(Sv / <sv>-I)*V(r) is (Sv / <sv>This corresponds to the spatial Fourier transform of -I). Therefore, in the case of periodic sampling, an effect similar to one-dimensional aliasing occurs. However, as will be considered, sampling does not have to be limited to periodic sampling. In principle, any choice of location p can be used as a subsampling point, or even as an image point based on random selection.

[0052] Intersection image In principle, U and V can be vectors having the same number of non-zero response function components as the array elements 100 in array 10. This corresponds to using all array elements 100 to create a transmit beam focused on different positions r in object space, and further, using all array elements 100 to create a selective receive pattern focused on different positions r.

[0053] The actual implementation of the signal flow model may involve consecutive measurements using different (groups of) array elements 100 consecutively as transmitters and receivers. This can lead to excessive measurement time.

[0054] In this embodiment, the image is calculated using transmitter elements located at different offsets relative to rows on array 10 and receiver elements located in a single row (where jx and jy are used to indicate selected row and column numbers in matrix 10, and j indicates a combination of jx and jy), with U and V selected to provide resolution in the x and y directions, respectively.

[0055] In a preferred embodiment, transmitter elements at different offsets relative to a row reside in a single column. This provides optimal results. However, if necessary, transmitter elements at different offsets may reside within a range of columns selected such that they do not reside further from the reference column than they do in a single row. For example, transmitter elements at different offsets relative to a row may be one or fewer array elements from the intersection column. In another embodiment, transmitter elements at different offsets reside along a line traversing the row direction at an angle that deviates from the column direction by 30 degrees or less, for example. Transmitter elements are said to be along such a line if they are the nearest transmitter elements adjacent to the line of the row.

[0056] In the following, a preferred embodiment is described as an example in which transmitter elements with different offsets reside in the same column. In this case, U and V can be seen as samples of a single row in jy and a single column in jx, respectively, of the more general functions U' and V', which depend on two-dimensional jx and jy. The access structure of the wearable ultrasound imaging device allows for simultaneous reception of signals in a single row.

[0057] Images acquired using transmitter elements in a single column and receiver elements in a single row of array 10 can be associated with an "intersection," i.e., a combination of row jy and column jx in array 10. Such an image is referred to as an "intersection image" Im(j,r). However, in principle, the image value of such an intersection image Im(j,r) does not depend on the intersection "j," and an intersection image Im(j,r) for any intersection "j" can be defined across the entire object space for all positions "r" in that space at any distance from the intersection "j."

[0058] However, if different subsampled selections from the transmitter elements in column jx and the receiver elements in row jy of array 10 are used for different intersections, the intersection images Im(r,j) for different intersections j will be different from each other, although they approximate the same "true" image. Generally speaking, in a plot of slices of Im(j,r) at a constant distance from the surface of array 10, the term errU corresponds to artifacts along horizontal lines, the term errV corresponds to artifacts along vertical lines, and the term errUV corresponds to artifacts along slope lines. Since deep subsampling may be used, e.g., subsampling less than one out of ten locations on array 10, the deviations can be considerable.

[0059] In this embodiment, the image value of the aggregated image Im(r) is calculated as the sum of the image values ​​of such intersection images Im(j,r) (preferably a weighted sum, weighted by a weighting coefficient w(j)). The sum of j in Im(r) = w(j)Im(j,r)

[0060] The summation of j is limited to intersection images for selection of intersections only, for example, intersections on a grid where the number of columns and rows of array 10 between adjacent intersections is fixed. Different selections of subsampling locations on array 10 are used to calculate the image value of the same position "r" in object space for different intersection images Im(j,r). As a result, different deviations resulting from subsampling for different intersection images are summed up, and the deviations of the different intersection images may cancel each other out or at least not be the same for each intersection image. This can be used to reduce the net deviation in the summed image compared to the deviation resulting from subsampling in the individual intersection images.

[0061] Figure 2c shows a portion of the array 10 having array elements 100 (labeled only) only at selected locations used to calculate image Im(r) from intersection image Im(j,r). Note that the array may be further extended in all directions from the portion shown.

[0062] All array elements in all rows are shown only for the selected column 120 containing the intersection, and similarly, all array elements in all columns are shown only for the selected row 110 containing the intersection. Only array elements in the selected row 110 are used as receivers for the intersection, and only array elements in the selected column 120 are used as receivers for the intersection. Array elements are not needed or do not exist at other locations in the array 10. In some embodiments, such array elements do not exist in the array. However, it should be noted that in other embodiments, if the intersection is used to compute an image of a wider area, for example, more array elements may exist at all locations in the array to enable the computation of a local image using all array elements in the selected area.

[0063] The number of array elements between consecutive intersections is merely an example for illustrative purposes. In reality, there may be array elements between consecutive intersections, ranging from 10 to 100. All intersection images can be calculated according to the same method, with the distinction made only in the selection of sampled locations within array 10 used to calculate different intersection images. In embodiments, the general vectors U(r) and V(r) used in this calculation are the same for all intersection images, but for each different intersection, a different selection of components of these vectors is used according to the selection of sampled locations within array 10 for each of these intersections. In a preferred embodiment, this selection is first made so that the image value of each intersection image Im(j,r) is obtained by transmitting ultrasound only from the location of array element 100 in column jx and using only the array element in row jy. Secondly, for each intersection image, the Im(j,r) selection is made from the locations in this row and column used to calculate the intersection image.

[0064] Each intersection image is based on a transmission from an array element at a subsampled location within a range in a column of array 10, and a reception from an array element at a subsampled location within a range in a row of array. The subsampled locations include locations at the edges of the range. The row and column ranges for two of the intersections 130, 132 are shown as two-dimensional crosshair windows having horizontal ranges 140a, 142a and vertical ranges 140b, 142b on array 10. These intersection ranges are large enough to overlap each other. As shown, each range of the crosshair windows 140a, b, 142a, b for intersections 130, 132 includes its intersection. Preferably, the intersection of each range 140a, b, 142a, b for each intersection 130, 132 is at least near the intersection 130, 132.

[0065] For example, an array may have a size of 150 × 150 mm, contain 5 to 20 array cells per mm, and have a length of 20 to 100 mm. Using intersection images acquired by transmit and receive with array elements distributed over a much wider range of locations than the distance between intersections has the effect of providing high resolution to the intersection images.

[0066] Figure 3 similarly illustrates intersection locations 50 along a single horizontal line 52. Different intersection locations are associated with their respective vertical lines 53a-e through the intersection locations. An image for each intersection location 50 is formed using excitation signals from the array elements 100 in a first aperiodically subsampled selection of array element locations along its vertical lines 53a-e and received signals from the array elements 100 in a second aperiodically subsampled selection of locations along the horizontal line 52.

[0067] For one intersection location 50 of the vertical lines, Figure 3 shows a horizontal range 54 along a horizontal line 52 between the leftmost and rightmost locations in a second selection of that intersection location 50. This horizontal range 54 includes the intersection location 50 and extends beyond adjacent intersection locations 50 in the row direction. Thus, for at least the nearest adjacent intersections in the row direction, the horizontal ranges of locations where signals are received for the intersections overlap. The locations in the first sampled selection similarly extend over a similar vertical range including the intersection location 50, preferably over the vertical range of at least the nearest adjacent intersections in the column direction.

[0068] For each intersection location 50 labeled "j", the ultrasonic signal time-dependent signal e0(i,j) is excited at a selected location labeled jy along the vertical line 53a~e of intersection location j=(jx,jy). Similarly, for each intersection location j, the ultrasonic signal time-dependent signal vs(jx,jy) is received, and the selected location labeled jx along the horizontal line 52 of intersection location j is used to calculate the image value Im(r,j) for the intersection. To form the image value of the aggregated image, the same "r" image values ​​of intersection images from different intersections are summed up.

[0069] In the combined image, the deviation due to subsampling is the sum of the deviations errU, errV, and errUV in the intersection image. errU and errV appear as deviations along the x and y axes of the image, respectively. errUV appears as a deviation along the slope of the xy-plane of the image. The formulas for errU and errV in the combined image are given by the sum of the sampling functions, which is the sampling function Su(j) / <Su(j)> , Sv(j) / <Sv(j)> It is the same as the case of the intersection image, except that it replaces the previous one. For example, the following: The integral of r' in errU=D*U(r)[w(j)(Su(j) / <Su(j)> -I)) sum of j]*Tu(r')R(r')*s

[0070] w(j)Su(j) / <Su(j)> If the sum of j better approximates a constant function, i.e., if the sampling function is equal to 1 for more locations on array 10, the size of the deviation decreases. When deep subsampling is used, w(j)Su(j) / <Su(j)> Since the value of is far from a constant function, the deviation of individual intersection images will be large.

[0071] However, as the formula for errU shows, the use of different subsampling selections for different intersections can be used to improve the approximation of the constant function, so that the term w(j)Su(j) / from different intersection images can be used.<Su(j)> These are added together. The more scattered the subsampling locations are, the less the selection of sampling locations for different intersection images "j" will match, and the smaller the deviation due to subsampling in the combined image will be.

[0072] (Su(j) / <Su(j)> The variance of the sum of j in (-1) can be used as a measure for approximating a constant function. For a single intersection image, if Su(j) is equal to 1 for m subsampled selections out of n possible sampling locations, then (Su(j) / <Su(j)-1> The variance of is 1 / <Su(j)> The variance is -1. For N overlapping intersection images where the sampling points do not coincide, the variance is (1 / N* <su>-1 (No matching sampling points exist, N* <su>(This means that is less than or equal to 1), i.e., the variance decreases with the number of intersections N that have overlapping ranges. Since adjacent intersections in both the x and y directions are counted for N in this context, it should be noted that a rapidly increasing gain can be realized with a two-dimensional array of intersections.

[0073] If necessary, the same applies to the deviation errV. For errUV, the formula is more complex, but it has also been found that reducing the agreement between Su and Sv in different intersection images reduces errUV for the combined image. In both cases, the gain is optimal when the direction in which the transmitting element is located is perpendicular to the row direction.

[0074] Therefore, in order to allow the net deviation of the aggregated image to be smaller than the deviations in the individual intersection images due to subsampling, the ranges on array 10 associated with adjacent intersections 130, 132 overlap, and the subsample locations of different intersections within that overlap are at least partially scattered amongst themselves. Locations in the first set are said to be scattered among locations in the second set if different locations in the first set lie between different pairs of consecutive locations in the second set, without precluding that some pairs of consecutive locations in the second set have multiple locations in the first set between them, or that some pairs have no locations in the first set between them at all. Locations in the first set and the second set are generally scattered if they are distributed across the same range of locations.

[0075] The range associated with intersection 130 may overlap with multiple ranges associated with nearby intersection locations, for example, with windows associated with N neighboring intersections in both the x and y directions, e.g., N=2 or N=4. Preferably, the range of the closest adjacent pair of intersections is at least twice the distance between the intersections. More preferably, it is a significantly reduced resolution of 3, 4 or more times. A sliding window may be used, where each intersection is the same size and the relative location to the intersection is the same, but the sampling location within the window to the intersection may be different.

[0076] The number of subsample locations within each range may be, for example, one-fifth of the total number of subsample locations within the range, or less, for example, one-twentieth. Preferably, the subsampling rate is large enough, on average, that at least two subsample locations for an intersection exist between that intersection and each of its nearest adjacent intersections. Preferably, at least 32 subsample locations are used within each range. Larger numbers such as 128 or 256 may be used within each range.

[0077] If the subsampling locations are distributed across the entire sampling range, the effect of subsampling on resolution is limited. The width of the zero-frequency peak of the absolute square of the Fourier transform of the sampling functions Su, Sv can be used as a measure of both resolution and the variance of the subsampling locations. Preferably, Su, Sv are selected such that this width does not exceed twice the width without subsampling.

[0078] In this embodiment, the selection of sampling locations in the horizontal and vertical ranges 140a,b of the window for intersection 130 is made such that none of the sampling locations in the horizontal and vertical ranges 140a,b of window 140 coincide with any sampling locations for other intersections 132 that have horizontal and vertical ranges 142a,b windows (e.g., 142) that overlap with those for intersection 130. This can be easily achieved when the subsampling rate is low. The sampling locations within each range are scattered with the sampling locations in overlapping ranges for other intersections. This reduces the net deviation of the aggregated image. For example, deep sampling using less than one-tenth of the locations within the sampling range makes such selection possible even when there are many overlapping windows.

[0079] However, even if some of the sampling locations within window 140 coincide with sampling locations within windows having overlapping ranges 142a,b, the net deviation of the combined image may be smaller than the deviation of the intersection image, unless all sampling locations within that overlap coincide. That is, at least some of the sampling locations within a window are scattered with the sampling locations within the overlapping window, while other parts of the sampling locations within a window may coincide with the sampling locations within the overlapping window. Preferably, at least half of the sampling locations within window 140 do not coincide with the sampling locations within the overlapping window. Similarly, it is not necessary to use all possible sampling locations within the ranges 140a,b, 142a,b of at least one window.

[0080] Non-periodic subsampling Preferably, the selection is aperiodic, i.e., the selected sampling locations for the intersection do not exist at integer multiples of the base distance. Aperiodic selection has the advantage of blurring the aliasing effect on the remaining net deviation in the aggregated image and reducing the size of the deviation at the image location r where the size of the deviation is largest.

[0081] To characterize the aperiodicity, Su- <su>and / or Sv- <sv>A location-dependent Fourier transform may be used. Subtraction of Su and Sv prevents a peak at zero frequency corresponding to a sample without deviation. In an approximation where the dependence of travel time difference varies as p*(e(r)-e(r')) / c, the successive operations U(r)SuTu yield a deviation corresponding to the convolution of R by this Fourier transform. In the case of periodic subsampling, the amplitude of this Fourier transform has a sharp peak in the sampling function, and this peak causes aliasing. In contrast, in the case of aperiodic subsampling, the amplitude is distributed pseudo-continuously over a range of frequencies. One measure of aperiodicity attributable to the received sample Su, Q, is the highest peak value, excluding the peak at zero frequency, when the square of the absolute value of the Fourier transform of the sampling function Su is divided by the maximum value of the square of the absolute value of the Fourier transform of a periodic subsampling function having the same number of subsample locations (same subsampling rate). Preferably, aperiodic subsampling with a Q of less than 0.5, more preferably with a Q of less than 0.1, is used. If necessary, similar criteria can be used for Sv. Sampling selection that satisfies such criteria can be easily achieved by performing deep subsampling. <su>If the value is less than 0.5, then a nearly arbitrary random selection of sample locations is sufficient.

[0082] Therefore, using subsamples dispersed over a much wider range than the distance between intersections helps provide high resolution to the intersection images, the measurement time required to use such a wider range is reduced by subsampling, and the resulting artifacts are reduced by aggregating non-periodic subsamples and intersection images obtained with significantly different subsampling selections.

[0083] In one embodiment, aperiodic selection can be obtained by using frequency modulation, where consecutive sampling locations p(m) indexed by m are selected according to p(m+1)=p(m)+r(m). Herein, the frequency r(m), i.e., the distance between consecutive sampling points, is selected such that the selected distances are distributed over a predetermined distance range (frequency modulation range), preferably excluding distances below that predetermined distance so that consecutive sampling locations within that distance are prevented. The frequency r(m) can be randomly selected over the frequency modulation range according to a predetermined probability distribution. Thus, the average <su>This is the mean of a probability distribution. For example, a probability distribution is constant within a limited range at its center, and other... <su>(It can be zero in some locations.)

[0084] If the sampling selection for each intersection is strictly periodic (p(m+1)=p(m)+d, where "d" is the sampling period), then even if the Q value is equal to 1, the net deviation in the aggregated image can be reduced by selecting the sampling locations for calculating different intersection images in different ways, for example by using different values ​​of p(m)mod d for different intersection images. However, it has been found that using aperiodic subsampling provides better suppression of deviation.

[0085] Embodiments have been described in which intersection images are calculated along a single row and column, respectively, in reception and transmission. However, it should be noted that a third selection of multiple rows may be used instead of a single row. This can be seen as including multiple intersections in the same column and different rows, except that the same sampling locations are used in the columns. The different selection may be used for different rows in the third selection, preferably such that at least some of the sampling locations are scattered with the sampling locations in the different second selection, as well as the sampling locations in other intersection images. Alternatively, this can be seen as a single intersection image with additional rows in addition to the row that defines its intersection.

[0086] Using the same sampling location in the columns for multiple rows, instead of using different selections in the columns for each row, may increase the deviation of the combined image, but it is still better than using a single intersection image.

[0087] intersection In the illustrated embodiment, for example, a regular grid of intersections may be used, having intersections that are consecutive at distances of 10 to 100 array elements along the rows and columns. When a regular grid of intersections is used, the selected row 110 and selected column 120 form a regular grid of horizontal and vertical lines. Figure 2c shows an example where one of the four rows and columns is the selected row 110 and selected column 120 where the intersections are located, but other (generally larger) distances between the selected row 110 and selected column 120 may be used. However, a regular grid is not necessary. Instead, an irregular grid may be used. This can further reduce the deviations caused by subsampling.

[0088] It should be noted that intersection images are not simply stitched together multiple individual image patches that overlap with intersection image patches of at most partially adjacent intersections. In principle, intersection images each define image values ​​within the same image space, and summing intersection images acquired with significantly different subsampling selections is used to reduce subsampling artifacts in high-resolution images acquired with reduced measurement time.

[0089] The weights may be independent of the intersection. Alternatively, the weight coefficient w(j) may be used to assign variable weights to different intersection images, for example, depending on the position "r" in object space from which the aggregated image is calculated. In other embodiments, the weight w(j) may be used, for example, as a decreasing function of the distance between the image position r and the intersection location, or at least a non-increasing function, to assign less weight to different intersections or no weight at all as the distance from the image point increases. Alternatively, the image value of an intersection image for an intersection "p" far from the image point "r" may simply be ignored when calculating the aggregated image for that image point. For example, the weight may be non-zero (e.g., constant) only for the intersection closest to the image position from which the image value is calculated, and for all other intersections whose range overlaps with the subsample range. This may be used to limit noise when reducing subsampling artifacts.

[0090] The weights can be chosen so that the sum of their intersections equals 1 in order to normalize the aggregated image. However, since an unnormalized image is still an image, normalization is not necessary. A useful image can be obtained even if the weights w(j) are omitted.

[0091] If the array size is large and the row and / or column lengths are not significantly less than, or even greater than, the distance to the object position "r", then using a weighted sum of intersection images to compute the sum image can be helpful in giving the intersection image a decreasing weight as the distance between the intersection and "r" increases. In principle, using a sum of intersection images representing the same object space makes it easier to explain the fact that the most reliable intersection image is obtained for the intersection at the shortest distance to "r".

[0092] Implementation of a device array Figure 4 shows a cross-sectional view of the side view of the device. In an exemplary example, each array element 100 includes a portion of a layer 20 of piezoelectric material, and the portion of layer 20 in the array element forms a piezoelectric pillar 20a that is laterally separated from the portion of layer 20 in other array elements by a space 21 which may contain air or non-piezoelectric filler material. For example, each piezoelectric pillar 20a may have a height of 80 micrometers and a square cross-section with sides of 40 micrometers in length. The continuous array elements 100 may be provided with a pitch of 60 micrometers.

[0093] Space 21 can be created, for example, by starting from a continuous layer by hot embossing using a lithographic pattern stamp. A conductive layer 27 is provided at the top of the piezoelectric pillar 20a, extending over the array 10. Furthermore, each array element 100 has its own electrode 23a formed in the electrode layer 22 below the piezoelectric pillar 20a to apply an electric field through its piezoelectric pillar 20a to the electrode 23a and the conductive layer 27.

[0094] The substrate on which the array 10 is located may be a flexible carrier foil 26. Optionally, an encapsulation layer 28 is provided on top of a conductive layer 27. Column selection circuits 12 and column access circuits 14, 16 are located adjacent to the array 10 on the flexible carrier foil 26.

[0095] Row lines 24 may be provided below the electrode layer 22 coupled to the row selection circuit 12. Column lines 23b may be provided to the electrode layer 22 coupled to the column access circuits 14, 16.

[0096] Each array element 100 may include an access transistor having a channel formed in the transistor layer 25, the channel being coupled between the electrode 23a of the array element 100 and the row line 23b of the row of the array 10 in which the array element 100 is located. The row line 24 of the row of the array 10 in which the array element 100 is located may be connected to the gate of the access transistor of the array element 100, or may form the gate itself.

[0097] Figure 5 shows an exemplary circuit diagram of an array element 100 (shown by a dashed rectangle) coupled to row 24 and column 23b. The array element 100 comprises a piezoelectric pillar 20a, an electrode 23a, and an access transistor 32. The piezoelectric pillar 20a is located between the electrode 23a and a conductive layer 27, which may be at ground potential. The channel of the access transistor 32 is connected between the electrode 23a and column 23b of the column of the array 10 in which the array element 100 is located. The gate of the transistor is connected to row 24 of the array 10 in which the array element 100 is located.

[0098] In some embodiments, all array elements 100 may be designed and wired to function as both an ultrasonic transmitter and an ultrasonic receiver. Alternatively, different array elements 100 may be designed and wired to implement only one of the two functions: transmitting or receiving.

[0099] At an abstract level, it should be noted that the circuitry of the ultrasonic imaging device 42 has some similarities to DRAM memory circuits or pixel-based display control circuits. However, its use differs, at least in that time-dependent signals are used for generating and measuring ultrasonic signals.

[0100] process Figure 6 shows a flowchart of an embodiment of the operation of the imaging system. In the first step 61, the processing system 40 selects an intersection location (e.g., 130, Figure 2c) indexed by j=(jx,jy). In the second step 62, the processing system 40 provides an address to the first column access circuit to enable the selection signal from the signal generator 44 to pass through column line 23b corresponding to intersection location 130.

[0101] In the third step 63, the processing system 40 causes a row selection circuit to apply a selection signal to row line 24 so as to select a row corresponding to a location in the first selection of locations subsampled from a vertical range 140 with respect to the intersection location 130. While the selection signal is applied, the processing system 40 causes a signal generator 44 to generate an ultrasonic excitation signal according to the e0(t) component corresponding to the selected array element, and the first column access circuit passes this signal to column line 23b. Under the control of the row line signal, this excitation signal is selectively passed to the piezoelectric pillars of the array element 10 in the selected row and column. For example, the excitation signal may be a pulse-modulated signal, and the delay corresponding to the selected component of e0(t) may be considered separately.

[0102] Next, in the fourth step 64, the processing system 40 causes the row selection circuit to apply a signal to the row control unit to the row line 24 in order to select the row corresponding to the row in which the horizontal range 140 of the selected intersection 130 is located. Furthermore, the processing system 40 allows the second column access circuit to pass selected received time-dependent signals from array elements in the column corresponding to the location in the second selection of subsampled locations from the horizontal range 140 for the intersection location 130 to the detector circuit 45. In response, the detector circuit 45 obtains digitized signal values ​​of a series of time points from the transmission of the ultrasonic excitation signal. Preferably, signals from multiple columns are passed to the detector circuit 45 simultaneously and processed by the detector circuit 45. In an embodiment, the second column access circuit is configured to pass signals simultaneously from at least the horizontal range of the selected intersection. In another embodiment, the second column access circuit may be configured to pass signals simultaneously from programmed selected columns corresponding to the subsampled locations in the row. Still in the fourth step 64, the processing system 40 reads the digitized signal from the detector circuit 45 and stores the signal in memory (not shown).

[0103] The fourth step 64 may be performed starting immediately after the ultrasonic excitation signal of the third step 63 has finished. The fourth step 64 may continue during the time interval in which a measurable reflection of the ultrasonic excitation signal is received. If signals received from only a single column can be processed by the detector circuit 45 at one time, steps 63 through 64 may be repeated for consecutive locations in the second selection of subsampled locations from the horizontal range 140 for the intersection location 130.

[0104] In the fifth step 65, the processing system 40 tests whether all locations in the first selection of subsampled locations from the vertical range 140 for the intersection location 130 have been used. If not, in step 65a, the processing system 40 selects the next location from the first selection and repeats from the second step 62 for the next location. If all locations in the first selection 52 have been used, the processing system 40 performs the sixth step 66 to test whether the previous step has been performed for all intersection locations 130. If not, in step 66a, the processing system 40 selects the next intersection location 132 and repeats from the second step 62 for the next intersection location. In the sixth step 66, if it is determined that the previous steps have been performed for all selected intersection locations, the processing system 40 performs the seventh step 67, and the processing system 40 calculates the image.

[0105] The seventh step of image calculation effectively involves applying U(j,r) and V(j,r) to each image position "r", applying D to the result to obtain the intersection image value for each intersection "j", followed by the (weighted) summation of the intersection image values ​​obtained for the same image position "r" in the intersection images for different intersection j. The effect of V(j,r), which formally represents a vector with components for different array elements representing the signals transmitted from these array elements, can be synthesized by summing the signals received in response to transmissions by individual array elements. Similarly, the effect of U(j,r) can be synthesized by summing the signals received by individual array elements. Since V(j,r) does not change with the x component of "r" and U(j,r) does not change with the y component of "r", part of the calculation of the intersection image value for different positions "r" can be shared, which can reduce the computational complexity of the intersection image value.

[0106] Figure 7 illustrates an example of an image calculation that may be used in the seventh step 67. In this example, a position-dependent intersection image is calculated first and then summed. In the first step 71, for each intersection location j, a composite signal resulting from a combination of transmissions from a first selection of locations among subsampled locations from a vertical range 140 relative to intersection location j is synthesized for each location in a second selection of subsampled locations from a horizontal range 140 relative to intersection location j.

[0107] As is known, from the perspective of signal processing flow, the excitation signal for measuring reflection at position r in object space is defined by the response function vector V(j,r), where the vector components correspond to locations in the array and define the time- or frequency-domain filtering operation. Only the selected components of V corresponding to the first subsampling selection of locations in the array are used, and different first subsampling selections are used for different intersections. Similarly, only one of the subsampling selections of locations in the array is used.

[0108] Therefore, the processing system 40 synthesizes the composite reflections for each location in the second selection by applying the components of the first directional response function vector V(j,r) to all selected measured reflected signals acquired for that location in the second selection in response to transmissions from different locations from the first selection of location 50 for the intersection location, and summing the results.

[0109] The measured reflected signal and the combined reflected signal are time-dependent, and the combined reflected signal can be calculated for multiple time points. Alternatively, Fourier transform domain calculations may be used to account for time dependence, including a phase factor that treats the Fourier transform domain signal as a complex number to compensate for the travel time difference in the beam direction from different locations 50 from a first selection.

[0110] In the second step 72, each intersection location processing system 40 synthesizes the reflected signal in the object space direction from the synthesized reflected signal to the location from the second selection for the intersection location.

[0111] As you know, from the perspective of signal processing flow, the received signal for measuring reflection at position r in object space is defined by the response function vector U(j,r), where the vector components correspond to the positions in the array and define the time- or frequency-domain filtering operation. Only the selected components of U corresponding to the second subsampling selection of locations in the array are used, and different second subsampling selections are used for different intersections.

[0112] Therefore, the processing system 40 synthesizes the reflected signal in the object space direction by applying the components of the second directional response function vector U(j,r) to all the synthesized reflected signals obtained for the locations in the second selection for the intersection location and summing the results. The synthesized reflected signal is time-dependent and can be computed in the Fourier transform domain.

[0113] Since the first selection involves selecting multiple locations in the same column, the first step 71 provides an increase in the combined directional sensitivity when the directional component along the column direction changes. Similarly, since the second selection involves selecting multiple locations in the same row, the second step 72 provides an increase in the combined directional sensitivity when the directional component along the row direction changes. The first and second steps 71 compensate for directional-dependent propagation time differences, however the combined signal may still depend on the time from transmission of the ultrasonic excitation signal, for example, due to variations in reflectivity in object space as a function of distance to the array.

[0114] In the third step 73, for each intersection location, the processing system 40 uses the synthesized signal for different combinations of beam direction and reception direction to obtain a selected time-delayed reflection measurement from the excitation of the ultrasonic signal. This corresponds to the effect of signal processing operation "D".

[0115] In principle, the image values ​​of a three-dimensional image of ultrasonic reflections in object space can be calculated using this method. Optionally, the calculation may be limited to locations on the surface in object space. In the fourth step 74, the processing system 40 constructs an aggregate image of reflections using the results of different intersection locations. Optionally, the results may be obtained along a conformal surface that is conformal to the surface of the array at a distance d from the surface of the array, using the results of the third step 73 for locations within the conformal surface. Another option is to determine the results of a two-dimensional slice of object space extending across the array.

[0116] As mentioned above, the example in Figure 7 serves primarily as an illustrative solution for the implementation of the calculation. In practice, different implementations may be used. For example, some or all of the calculations of time and / or spatial position dependence may be replaced by calculations in the Fourier transform domain. As another example, the sequence of steps may be modified, for example, by synthesizing the signal for the receiving direction first and then the signal for the beam direction later, or by selecting the signal for the time delay earlier and then synthesizing the signals for the receiving direction and / or beam direction later. As yet another example, the first and second steps 71, 72 may be performed as part of the steps shown in Figure 6, before all signals have been recorded for the intersection locations, although the necessary signals have been recorded. For example, if the second step 72 is performed first, it may be performed before the signals using all locations 50 from the first selection are used, or steps 71, 72 may be performed before all intersection locations are used.

[0117] In some embodiments, the total measurement time can be reduced by receiving responses to the same transmitted signal across multiple rows. The transmission time interval during which the array elements are excited to transmit ultrasound is typically much shorter than the reception time interval during which the reflection of that ultrasound is received. In most embodiments, it is possible to change the selected row multiple times during the reception time interval to provide time-multiplexed outputs of reflections across different rows.

[0118] Therefore, as illustrated in Figure 8, in response to the same transmitted signal, signals for different intersections along the same column 82 may be received by receivers in different rows 84. Receiving signals for different intersections along the same column 82 at receivers in different rows 84 has the disadvantage that the same vertical subsampling selection is used for different intersections, which renders the suppression of subsampling artifacts ineffective. However, in many applications, some reduction and suppression of subsampling artifacts may be acceptable.

[0119] Alternatively, this can be seen as an embodiment in which receivers in multiple rows 84 are used for the same intersection 80 to obtain a portion of the resolution in the column direction, thereby reducing the number of column-direction locations that themselves are required to provide the same resolution. Thus, multiple rows 84 of the receivers are used for one intersection such that one or more of these rows 84 do not cross a column 82 at the intersection 80.

[0120] In this embodiment, the locations of array elements used for receiving in different rows 84 are used together with the locations of array elements used for transmitting in different rows of column 82 to combine the resolution in the column direction. In this embodiment, the resolution in the column direction can be described as the product of a transmit function and a receive function, which are the result of a combination using array elements for transmit and a combination using array elements for receive, respectively.

[0121] Similar to embodiments having a single row, the response function U(j,r) may be selected to compensate for the time difference in travel from the position of the array element used for reception (which is now in a different row) to the position "r" in object space where the image is formed.< / su> < / su> < / su> < / sv> < / su> < / su> < / su> < / sv> < / sv> < / su> < / su> < / sv> < / su> < / sv> < / su> < / su> < / su> < / sv> < / su>

Claims

1. A method for forming an object image of ultrasonic reflectance in an object space using an array of rows and columns of ultrasonic transducers that define sensing surfaces adjacent to the object space, wherein the method is - The calculation includes determining the object image value for each position in the image of the ultrasonic reflectance within the object space by summing the image values ​​based on the intersections for the positions determined for each of the multiple intersection locations associated with each combination of rows and columns in the array, The image value based on the intersection for each of the aforementioned multiple intersection locations is ●Transmitting an ultrasonic transmission signal from the ultrasonic transducer to a subsampled first selection of first sampling locations that are not located further from the row of each intersection location than the column of each intersection location, wherein the first selection extends over a first range of offsets from the row of the first sampling location. ● Receiving the reflection of the ultrasonic transmission signal from the object space in a second selection of subsampled second sampling locations along the row of each intersection location, or in a second selection of second sampling locations in a third selection of subsampled rows including the row of each intersection location, wherein the second selection or the second selection extends over a second range of the second location along the row of each intersection location, and at least a portion of the reflections in the second selection or the second selection are received simultaneously. ● For each of the intersection locations, the image value based on the intersection, which is associated with the reflectance at the position in the object space, is calculated by synthesizing the reflections received in the second selection or a plurality of the second selections in response to the ultrasonic transmission signal in the first selection, - The first range for each intersection location has a first overlap with the first range for one or more of the intersection locations, and at least a portion of the first locations in the first range for one or more of the intersection locations are at least partially scattered among the first locations in the first range for each intersection location. A method wherein the second range or the second plurality of ranges for each of the intersection locations has a second overlap with the second range for one or more of the plurality of intersection locations, and at least a portion of the second locations in the second range for one or more of the plurality of intersection locations are at least partially scattered among the second locations in the second range for each of the intersection locations.

2. The method according to claim 1, wherein the first range for each of the intersection locations extends at least to the offset of the nearest intersection location from each of the intersection locations, and / or the second range for each of the intersection locations extends at least to the nearest row from each of the intersection locations.

3. The method according to any one of the prior claims, wherein the first selection and the second selection are aperiodic selections.

4. The method according to claim 3, wherein the distance values ​​between consecutive sampling locations among the first sampling locations and / or between consecutive sampling locations among the second sampling locations are distributed over a predetermined distance range including a plurality of distance values.

5. The method according to claim 4, wherein the lower limit of the predetermined distance range is at most half of the average of the distance values.

6. The method according to any one of the prior claims, wherein the intersection is located in a two-dimensional periodic grid having a period smaller than the first range and the second range.

7. The method according to any one of the prior claims, wherein the mean subsample ratio of the first selection and the second selection is at most half the density of the intersections along the column and / or row of the respective intersection locations.

8. The method according to any one of the prior claims, wherein the sum of image values ​​based on the intersection with respect to the position is a weighted sum.

9. The method according to any one of the prior claims, wherein the weighted sum defines the window of the intersection location such that the first range and / or the second range all overlap with at least one of the first ranges and / or the second ranges among the intersection locations in the window.

10. A computer program product comprising a program of instructions for a programmable processing system, wherein, when executed by the programmable processing system, the computer program product causes the programmable processing system to perform the method according to any one of the prior claims.

11. An ultrasound imaging system, - An array of rows and columns of ultrasonic transducers defining the sensing surface, - An excitation circuit configured to induce ultrasonic emission from a selectable first ultrasonic transducer in the array, - A receiving circuit configured to receive an ultrasonic measurement signal from a selectable second ultrasonic transducer in the array, - A processing system, and the processing system is - The system is configured to calculate the object image value for each position in the ultrasonic reflectance image within the object space by summing the image values ​​based on the intersections for the determined positions, for each of the multiple intersection locations associated with each combination of rows and columns in the array. The processing system determines the image value based on the intersection for each of the multiple intersection locations. ● Causing the ultrasonic transducer to transmit an ultrasonic transmission signal from a subsampled first selection of first sampling locations that are not located further from the column of each intersection location than the row of each intersection location, wherein the first selection extends over a first range of offsets from the row of each intersection location. ● Receiving the reflection of the ultrasonic transmission signal from the object space in a second selection of subsampled second sampling locations along the row of each intersection location, or in a second selection of second sampling locations in a third selection of subsampled rows including the row of each intersection location, wherein the second selection or the second multiple selections extend over a second range of the second location along the row of each intersection location, and at least a portion of the reflections in the second selection or the second multiple selections are received simultaneously. ●For each of the intersection locations, the determination is made by calculating an image value based on the intersection, which is associated with the reflectance at the position in the object space, synthesized from the reflections received in the second selection or a plurality of the second selections in response to the ultrasonic transmission signal in the first selection. - The first range for each intersection location has a first overlap with the first range for one or more of the intersection locations, and at least a portion of the first locations in the first range for one or more of the intersection locations are at least partially scattered among the first locations in the first range for each intersection location. - An ultrasonic imaging system in which the second range or the second plurality of ranges for each of the intersection locations has a second overlap with the second range for one or more of the plurality of intersection locations, and at least a portion of the second locations of the second range for one or more of the plurality of intersection locations are at least partially scattered among the second locations of the second range for each of the intersection locations.

12. - The excitation circuit comprises a signal generator and a first access circuit coupled between the signal generator and a row of the array to generate ultrasonic waves having a time dependency defined by the signal generator from the first selectable location in the array, wherein the first access circuit is configured to select the location of the first ultrasonic transducer under the control of a processing circuit. - The ultrasonic imaging system according to claim 11, wherein the receiving circuit comprises a detection circuit and a second access circuit coupled between the array column and the detection circuit, the detection circuit is configured to detect the time dependence of the ultrasonic measurement from the selectable second ultrasonic transducer in the array, and the second access circuit is configured to select the location of the second ultrasonic transducer under the control of the processing circuit.

13. The ultrasonic imaging system according to claim 11 or 12, wherein the first range and / or the second range for each of the intersection locations extends, respectively, along the column and / or row of each of the intersection locations, to at least the nearest intersection location from each of the intersection locations.

14. The ultrasound imaging system according to any one of claims 11 to 13, wherein the first selection and the second selection are aperiodic selections.

15. The ultrasonic imaging system according to any one of claims 11 to 14, wherein the sum of image values ​​based on the intersection with respect to the position is a weighted sum.

Images