FR3045165A1 - METHOD AND SYSTEM ON ULTRASOUND IMAGING CHIP - Google Patents

Abstract

The invention relates to a system-on-chip (1) for ultrasound imaging, comprising: - an acquisition unit (2) configured, for each transmission of a plurality of successive transmissions, to acquire the signals received by transducers as echoes of the program in question; an order sequencer (10) configured to order a flight time calculation module (8) and an image generation module (9) to reconstruct an image of an image zone by proceeding in a manner iterative for each transmission of said plurality of successive transmissions: ▪ calculating, for each point (P) of the image and for each transducer, the flight time corresponding to the signal path of the transmission passing through the point and reaching the transducer; ▪ calculating an intermediate image by summing, at each point of the image, the amplitude of the signal received by each transducer at the instant corresponding to the calculated flight time; Summing the calculated intermediate image to a cumulative image of the previous iteration to form a cumulative image of the iteration.

Description

METHOD AND SYSTEM ON ULTRASONIC IMAGING CHIP

DESCRIPTION

TECHNICAL AREA

The field of the invention is that of ultrasound imaging applicable to the non-destructive testing of parts or industrial installations to detect the presence of possible defects, to locate and size them, as well as to any type ultrasonic ultrasound imaging, especially in the medical field for the inspection of areas of interest in the human or animal body. The invention more particularly relates to an ultrasound imaging on-chip system for imaging the interior of an object using a multi-element probe and a synthetic focusing technique.

STATE OF THE PRIOR ART The purpose of the non-destructive test is to detect and characterize any defects present in mechanical parts.

Current ultrasound techniques use phased array sensors associated with instrumentation and data acquisition chains. Given the amount of data to be processed, the analysis of these data is generally performed offline.

Moreover, the algorithms for calculating imaging methods based on synthetic focusing, for example of the "Total Focusing Method" (TFM) type, require numerous operations on a large volume of data. This implies significant computation times, requiring a delayed processing.

However, the industrial constraints of characterization of mechanical parts in situ make it necessary to carry out the reconstruction of images during the measurement. It is then necessary to be able to process in real time (with a high frame rate and a low latency between the acquisition of the signals and the provision of the image) large amounts of data (which result from the large number of sensor elements, large depth of inspection and significant image resolution).

In this way, it is sought to embark on the measuring apparatus all the data acquisition and processing architecture of the sensors. However, to date, there is to the knowledge of the inventors no fully integrated implementation, from the acquisition of signals to the generation of images of the interior of the inspected part, which is compatible with a low-wattage device. consumption.

Jorge F. Cruza, Jorge Camacho, José M. Moreno and Carlos Fritsch's article entitled "Ultrafast hardware-based focal law calculator for automatic focusing" (NDT & E International 74 (2015) 1-7, a hardware implementation integrating within a dedicated circuit the realization of certain calculations with the signal acquisition function. These calculations are no longer performed offline, delayed by software processing, but directly within the dedicated circuit. This solution is however incomplete in that the calculations considered are confined to the calculations of flight times.

DISCLOSURE OF THE INVENTION The invention aims to respond to the need for a hardware system for performing the entire algorithmic chain of a synthetic focus, from the acquisition of signals to the generation of images. It also aims to allow to image in real time the interior of a room whatever the profile, including a piece of complex surface not known a priori. For this purpose, the invention proposes an ultrasonic imaging on-chip system of a region of interest of an object, comprising: a unit for acquiring ultrasonic signals from a probe comprising a plurality of ultrasonic transducers, the acquisition unit being configured to acquire the signals received by transducers as echoes of a sequence of successive transmissions of an ultrasound signal by at least one transducer of the probe; a unit for processing the acquired ultrasonic signals, which comprises: a module for calculating the time of flight; o an image development module; and o an order sequencer configured to instruct the time of flight calculation module and the image generation module to reconstruct an image of an imaged area by iteratively proceeding for each transmission of said program sequence. successive: calculating, by the flight time calculation module, for the transmission considered, for each point of the image and for at least one transducer, a flight time corresponding to the course of the signal of the emission passing through the point and reaching the at least one transducer; calculating, by the image forming module, an intermediate image by summation, at each point of the image, of the amplitude of the signal received by the at least one transducer at the instant corresponding to the time flight calculated by the flight time calculation module corresponding to the signal path of the emission passing through the point and reaching the at least one transducer; and when there is a cumulative image of a preceding iteration, at the summation, by the image processing module, of the intermediate image computed with the cumulative image of the preceding iteration to form a cumulative image of iteration.

Some preferred but non-limiting aspects of this system-on-a-chip are the following: the system-on-a-chip allows the imaging of a region of interest of an object separated from the probe by means of a coupling medium, the processing unit further comprises an interface determination module, and the command sequencer is configured to: o order the flight time calculation module and the image generation module to reconstruct an image of the imaged zone with consideration by the flight time calculation module of a propagation of the ultrasonic signal in the single coupling medium, the cumulative image of the last iteration formed by the image forming module constituting an image of area ; ordering the interface determination module to determine, from the surface image, an interface between the coupling medium and the object; o order the time of flight calculation module and the image development module to reconstruct an image of the imaged zone with consideration by the flight time calculation module of the interface determined by the determination module interface, the cumulative image of the last iteration formed by the image processing module constituting an image of the region of interest of the object; the interface determination module comprises a surface image analysis unit configured to determine a raw profile of the surface of the object and a programmable filtering unit of the raw profile; the interface determination module is embedded in a programmable microprocessor; the flight time calculation module comprises a flight time calculation unit and an interpolator configured to determine interpolated flight times from the flight times calculated by the flight time calculation unit; it further comprises a writing unit for transferring the signals acquired by the acquisition unit to a memory unit, the writing unit being configured to interleave, for each group of a plurality of groups of transducers , the signals acquired by the transducers of the group and for transferring the intertwined signals from the group to the memory unit; it further comprises a reading unit for reading the signals stored in the memory unit and a unit for filtering the signals read by the reading unit, the reading unit being configured to deinterlace the signals of each of the groups; and the filtering unit being configured to filter the signals of each of the groups in parallel; the filter unit is configured to implement a high-pass filtering of the signals read by the reading unit, to implement an interpolation of the signals read by the reading unit, to implement a calculation of the Hilbert transform of the signals read by the reading unit. The invention also relates to the method implemented by such a system-on-a-chip.

BRIEF DESCRIPTION OF THE DRAWINGS Other aspects, objects, advantages and characteristics of the invention will appear better on reading the following detailed description of preferred embodiments thereof, given by way of non-limiting example, and made in reference to the accompanying drawings in which: - Figures la and lb represent two ultrasonic imaging methods using a synthetic focus; FIG. 2 is a diagram of a system-on-a-chip according to the invention; FIG. 3a illustrates the various operations that can be implemented by the on-chip system according to the invention; FIGS. 3b and 3c illustrate the different types of image reconstruction performed during the implementation of the operations illustrated in FIG. 3a; FIGS. 4a and 4b are diagrams representing a time-of-flight calculation module and an image-building module of a system-on-a-chip according to the invention, as a function of the type of image construction performed during implementing the operations illustrated in Figure 3a.

DETAILED PRESENTATION OF PARTICULAR EMBODIMENTS

In the field of non-destructive ultrasonic testing, a phased array probe is used which includes a plurality of ultrasonic transducers arranged in array, for example linearly or matrixically. The transducers are designed to emit focused or non-focused ultrasound waves. The transducers are also designed to detect echoes of ultrasonic waves reflected on and in the object.

The phased array probe can be in direct contact with the object. In order to optimize the control in contact with a piece of complex geometry, it is possible to use a conformable probe that matches the shape of the part. The invention also advantageously extends to the imaging of a region of interest of an object separated from the probe by means of a coupling medium, such as water. This is particularly the case when immersion imaging is carried out for which the object and the probe are immersed in a fluid, often water, which serves as a coupling medium. This is also the case when the conformability with the surface of the part is obtained by means of a flexible membrane filled with a fluid and interposed between the probe and the object.

A synthetic focus is to coherently sum the signals received by the transducers to observe amplitude maxima at the location where the defects causing the detected echoes are actually located.

In particular, a FMC ("Full Matrix Capture") acquisition method is known which consists in emitting, in succession, transducers a spherical ultrasonic wave in the part to be imaged, and in recording in Nr transducers the echoes originating from the propagation of the waves emitted in the room. The method thus leads to the formation of Ne * Nr elementary signals Sij (t) with 1 <i <Ne and 1 <j <Nr.

A method of reconstruction by focusing in all FTP points then consists of coherently summing the received elementary signals. The coherence of the signals is determined from the theoretical travel times, called flight times, ultrasonic waves propagating in the room between a given transmitter transducer and a receiving transducer and passing through a point considered. This method is illustrated in FIG. 1a, where the presence of a medium coupling Mc between the probe S and the object O is found.

The FTP reconstruction method thus includes for each point P of the reconstruction zone Z, and for each transceiver pair ij, the calculation of the theoretical flight time, Τ ^ {Ρ) corresponding to the path between the transmitter i and the receiver j passing through the point P, ie rf + rf with the notations of Figure la.

1 J

Then, for each point P of the zone Z, we proceed to the summation of the amplitudes extracted from the signals S ^ ij at times t = TjJ (p) l which can be written

The focus can be realized by considering several modes of reconstruction which are characterized by different types of path (direct, echoes of corner, indirect echoes, ...) and several wave propagation modes (longitudinal or transverse).

Another imaging method known as PWI (for "Plane Wave Imaging") consists in carrying out successive emissions of a plane ultrasonic wave having L successive emission angles, each of these emissions being carried out simultaneously by transducers, and record in Nr transducers echoes from the propagation of waves emitted in the room.

The reconstruction method is illustrated in FIG. For each point P, and for each emission q of a plane wave, it comprises the calculation of the theoretical flight time, Tcorresponding to the course of the plane wave q between the probe and the receiver passing through the point P, ie Tp + TP with the notations of Figure lb. Then, for each point P of the zone Z, the amplitudes extracted from the signals Sq, {t) are summed at times t = Tp + TP, which can be written as

The invention relates to a system-on-chip ultrasound imaging of a region of interest of an object. The object is, for example, a mechanical part that one wishes to inspect by non-destructive testing or, in the medical context, a part of human or animal body that one wishes to control non-invasively.

The system-on-a-chip may be a dedicated ASIC-type circuit ("Application Specifies Integrated Circuit") or a Programmable Logic Device (PLD), for example an FPGA (Field-Programmable Gate Array), suitably configured to implement a real-time imaging method based on ultrasound signals from a multi-element probe.

This implementation makes it possible to carry out the entire algorithmic chain of a synthetic focus, from the acquisition of the signals until the generation of the images, with, if necessary, preprocessing of the signals and detection of the profile of the part.

With reference to FIG. 2, the system-on-a-chip 1 comprises an acquisition unit 2 for ultrasonic signals. The acquisition unit 2 is configured to acquire the signals received by a set of transducers of the probe as echoes of a sequence of successive transmissions of an ultrasound signal by at least one transducer of the probe.

The ultrasonic signal emitted may be a spherical wave emitted by a transducer in the context of the FTP reconstruction, the successive transmissions of the sequence being carried out by different transducers. It may also be a plane wave emitted by a set of transducers as part of the PWI reconstruction, the successive transmissions of the sequence being made with different emission angles.

Nr is the number of transducers whose signals received are acquired for each of the successive transmissions. In an exemplary embodiment, Nr = 64. With a sampling frequency of 50 Mhz, the system-on-chip 1 receives, for each of the successive transmissions, a stream of data having a peak rate of 6.4 Gb / s coming from the probe. The acquisition unit may in particular take the form of a plurality of buffer memories 2, each associated with one of the Nr receiving transducers.

As shown in FIG. 2, the data of the acquired signals that are previously stored in the buffer memories 2 can be transferred to a memory unit 4 by means of a write unit 3 and read from the memory unit 4 to the memory unit 4. The memory unit 4 is, for example, a volatile memory of the SDRAM type whose writing speed, for example 800 MB / s, is less than the speed of the stream coming from the source. probe. This is for example a memory type DDR (for "Double Data Rate"). The data of the acquired signals read from the memory unit 4 via the reading unit 5 can be subjected to a set of pre-treatments within a preprocessing unit 6 before being recorded in a plurality of buffers 7, each associated with one of the Nr receiving transducers.

In one possible embodiment, the write unit 3 is configured to optimize the storage in the memory 4 of the data of the acquired signals in order to allow the implementation of a parallelism during their readback and thus to minimize latency. For this purpose, the writing unit is configured to interleave the data of the acquired signals of a group of several receiving transducers to form interleaved data packets to be stored in the memory 4. For example, Ns = 8 groups can be considered. transducers. The interleaving of the data makes it possible to exploit the bandwidth of the memory 4 in writing (for example of 0.8 Gb / s) while maintaining a high clock frequency (for example of 10 ohmz).

In a possible embodiment, the preprocessing of the signals is directly integrated in the replay stream. For this purpose, the reading unit 5 is configured to deinterlace the data, and filtering can be performed in parallel by means of Ns filters 6, each able to process the data of the transducers of a group. The filtering may in particular comprise a high-pass filtering, for example by means of an infinite impulse response filter IIR, aimed at suppressing the low-frequency noise. The filtering may be accompanied by over-sampling by means of an interpolation filter, for example FIR finite impulse response, of the data making it possible to improve the signal-to-noise ratio of the final images and thus the detection level of the images. defaults. The interpolation allows for example to go from 2048 to 4096 samples.

The filtering can also be accompanied by the calculation of the Hilbert transform of each of the acquired signals in order to be able to subsequently make an envelope of the images. In such a case, when the following description refers to processing applied to an acquired signal, it is actually a processing applied to the analytical representation of said signal where the imaginary part corresponds to the Hilbert transform. Similarly, a reconstructed image then corresponds to the module of the analytical signal. The computation of the Hilbert transform can be carried out by means of a finite impulse response filter FIR which makes it possible to avoid having to implement a fast Fourier transform and a fast inverse Fourier transform that are too expensive in terms of of resources.

The on-chip system 1 according to the invention also comprises an acquired ultrasonic signal processing unit, which comprises a flight time calculation module 8, an image preparation module 9 and an order sequencer 10.

The command sequencer 10 can be connected to a memory 21, for example a volatile memory type DDR SDRAM, in which are stored all the computable data before acquisition, which saves computing resources. Can notably be stored in the memory 21, the coordinates of the transducers of the probe, the coordinates of the pixels of the image in the coupling medium, the coordinates of the pixels of the image in the object, parameters for the determination of the profile surface of the object. These various parameters can in particular be entered via a software user interface implemented on a remote computer able to communicate with the system-on-a-chip via a communication bus, such as an express PCI bus. . The user interface also makes it possible to display and analyze the reconstructed images by the system-on-chip according to the invention.

The command sequencer 10 is configured to instruct the flight time calculation module 8 and the image preparation module 9 to reconstruct an image of an imaged zone from the data of the acquired signals recorded in the Nr memories. buffer 7.

More particularly, the command sequencer 10 is configured to instruct the flight time calculation module 8 and the image preparation module 9 to proceed iteratively for each transmission of the sequence of successive transmissions: to the computation by the flight time calculation module 8, for the transmission considered, for each point of the image and for at least one receiving transducer, a flight time corresponding to the path of the passing transmission signal. by the point and reaching the at least one transducer; the computation, by the image forming module 9, of an intermediate image by summation, at each point of the image, of the amplitude of the signal received by the at least one receiving transducer at the instant associated with the flight time calculated by the flight time calculation module 8 corresponding to the signal path of the emission passing through the point and reaching the at least one transducer; and when there is a cumulative image of a preceding iteration, at the summation, by the image-forming module 9, of the intermediate image computed with the cumulative image of the preceding iteration to form an accumulated image of iteration. The image of the image area then corresponds to the cumulative image of the last iteration. It is understood that by cutting the construction of the image of the image area in sequence (iteration), it is possible to limit the amount of data to be stored in the system-on-a-chip. In particular, during one of the iterations, the flight time module calculates at most only the flight times necessary for the formation of the corresponding intermediate image. Such an embodiment makes it possible to minimize the necessary resources (calculation, storage) as well as the latency.

The flight time calculation module also advantageously reuses the flight times, making it possible to limit the necessary calculations. For example, the calculation of the flight time of the transducer i to the transducer) passing through a point P comprises the calculation of the propagation times rfetrj. The flight time rf may advantageously be reused to evaluate the flight time of the transducer i to the transducer j + k passing through the point P which is indeed written rf + Tj + k.

The flight time calculation module can also advantageously exploit the principle of reciprocity according to which, for a pair of i-receiver transmitter elements j, the amplitude extracted from the corresponding signals for a point P at the time t = Tij (P ) is equal to the amplitude extracted from the signals 5; i (t) corresponding to the point P at the time t = T) j (P). Therefore, the image generation module has doubled the amplitude Aj .. during the sequence corresponding to the emitter i ^ 7 (for all i <j <Nr) and ignore the amplitude Aj .. during of the sequence corresponding to 7 ^ the transmitter j (for all i <j). This principle applies for all couples with index i * j and finally allows to divide by almost 2 the number of flight times to calculate: (n (n + 1)) / 2 instead of n2 for example.

Taking the example of a FMC acquisition and an FTP reconstruction, and with reference to FIG. 1a, the command sequencer 10 orders the flight time calculation module 8, for each transmission i, to determine for each point P of the imaged zone, and in the absence of implementation of the principle of reciprocity, for each transducer j receiving the flight time TP + TP = Ti} (P). With the implementation of the principle of reciprocity, this calculation is carried out for each transducer] such that all i <j <Nr It will be noted that these flight times can be calculated by considering one or more types of path (direct, echoes of corner, indirect echoes, ...) and one or more wave propagation modes (longitudinal or transverse).

The order sequencer 10 then orders the image calculation module 9 to calculate an intermediate image I, corresponding to the coherent summation of the signals received by the transducers in reception for the emission in question. The intermediate image is thus written Il {j ^) = YljL1Sij (Tij {P) ') in the absence of implementation of the principle of reciprocity.

And, when there is a cumulative image of a previous iteration, that is to say when i> l, the order sequencer 10 then orders the image calculation module 9 to perform the summation of the calculated intermediate image I, in the cumulative image 1, the previous iteration is used to form a cumulative image of the iteration. At the end of the last iteration, the desired image is obtained:

Taking the example of a PWI reconstruction, the command sequencer 10 orders the flight time calculation module 8, for each emission of a plane wave q, to determine for each point P of the imaged zone, and for each each transducer j in reception, the flight time TPq + TP = Tqj (P). The order sequencer 10 then instructs the image calculation module 9 to calculate an intermediate image lq corresponding to the coherent summation of the signals received by the receiving transducers for the transmission of the considered plane wave. The intermediate image is written as Φ = Σ%, s,; (T,; (/>)). At the end of the last iteration, the desired image is obtained: / ^ Tt (f)).

When the object to be imaged is separated from the probe by a coupling medium such as water, whether in the context of an immersion imaging or by means of a flexible membrane probe allowing to marry the surface of the object, adaptive focusing is required. Indeed, the propagation of the ultrasonic waves in two media (coupling medium, object) characterized by different propagation velocities is accompanied by a refraction at the interface between these two media, that is to say at the level of the surface of the object.

With reference to FIG. 3a, in order to achieve such adaptive focusing, the system-on-a-chip is then configured to implement the following steps. The system-on-a-chip first acquires the "ACQ" of the recorded signals from the transducers. During a first iteration loop, and for each successive emission, it proceeds to the reconstruction "REC-I" of a surface image taking into consideration a propagation of the emissions in the single coupling medium to form a Intermediate surface image Is-int At the end of the different iterations, the surface image ls is obtained. Figure 3b illustrates in this connection the reconstruction "REC-I".

The system-on-chip 1 then proceeds to an "EXT" step of extraction of the profile of the object by analysis of the surface image ls. Then during a second iteration loop, and for each of the Emissions, it proceeds to the reconstruction "REC-0" of an image of the object taking into account the profile of the object previously extracted for to form an intermediate image of the object lo-int- At the end of Ne iterations, the image of the object lo is obtained. FIG. 3b illustrates in this connection the reconstruction "REC-0" with refraction of the ultrasonic waves at the points M of the surface S of the object because of the difference in the propagation velocities c1, c2 between the coupling medium Mc and the object O.

In order to enable the implementation of these steps, the order sequencer 10 of the system-on-chip according to the invention is configured to initially order the module 8 for calculating shutter time at the module 9 for elaboration. of images to reconstruct an image of the imaged zone with consideration by the flight time calculation module of a propagation of the ultrasonic signal in the single coupling medium. The cumulative image of the last iteration formed by the image processing module constitutes the surface image ls.

As shown in FIG. 2, the on-chip system 1 furthermore comprises an interface determination module 11 and the command sequencer 10 is configured to order, in a second step, this module 11 to determine, starting from the surface image ls, an interface between the coupling medium and the object. The interface determination module 11 may, for example, use software buried in an embedded microprocessor within the system-on-a-chip. It is in such a way as to make the determination and possible filtering of the interface parameterizable.

The command sequencer 10 is furthermore configured, once this interface has been determined, to order, in a third step, the flight time calculation module 8 and the image generation module 9 to reconstruct an image of the imaged zone with consideration by the module 8 for calculating the flight time of the interface determined by the interface determination module 11. The cumulative image of the last iteration formed by the imaging module 9 constitutes the image lo of a region of interest of the object.

FIGS. 4a and 4b show an exemplary possible embodiment of the module 8 for calculating flight time and the module 9 for producing images. More particularly, Figure 4a illustrates the reconstruction of the surface image ls while Figure 4b illustrates the reconstruction of the image of the object lo.

In these FIGS. 4a and 4b, the flight time calculation module 8 retrieves the parameters necessary for calculation from a focusing sequencer 12a, 12b (shown under the common reference 12 in FIG. 2) coupled to the memory 21 mentioned above. In FIG. 4a, the focusing sequencer 12a thus provides the coordinates of the transducers and the coordinates of the zone to be imaged to a unit 13a for calculating the flight times in the coupling medium. In FIG. 4b, the focusing sequencer 12b thus provides the coordinates of the transducers, the coordinates of the zone to be imaged and the coordinates of the surface of the object to a unit 13b for calculating the flight times in the object. The units for calculating flight times 13a, 13b are represented under the common reference 13 in FIG. 2. The flight times thus calculated are stored in Nr buffer memories 14, each associated with one of the receiving transducers for the one of the successive issues.

In one possible embodiment, the flight time calculation units 13a, 13b provide raw flight times for each point of a gate corresponding to the area to be imaged; this grid can then be interpolated by an interpolator 15, for example bilinear, to provide interpolated flight times in a finer grid than the initial grid. The interpolation is controlled by a scheduler 16 which receives the parameters necessary for the interpolation calculation from the focusing sequencer 12, and which provides interpolation coefficients to the interpolator 15 and raw flight time indices to be interpolated to the interpolator. buffers 14.

The interpolated flight times are then used as memory pointers for recovering the amplitudes of the signals acquired from the buffer memories 7. These signals are used by the imaging module 9 to form an image of the iteration considered. As shown in Figs. 4a and 4b, an intermediate image is calculated by a coherent summation unit 17 configured to add the amplitude of the receiving transducer signals to the interpolated flight times. This intermediate image is accumulated by an accumulator 18 in the cumulative image of the previous iteration to form the cumulative image of the iteration which is stored in a buffer memory 19. We find in this buffer memory 19, at the end iterations carried out for the different successive transmissions, the surface image Is in the case of Figure 4a and the image of the object lo in the case of Figure 4b.

A possible embodiment of the interface determination module 11 is detailed below. This module comprises a unit for analyzing the surface image 1s configured to determine the coordinates of the pixels of the surface image 1s belonging to the interface between the coupling medium and the object. This unit comes for example to browse the surface image in order to extract for each column of pixels, the index of the pixel having the largest amplitude. It is considered that the most "echoic" area of the surface image is represented by the interface. These indexes are then used to determine the coordinates of the points of the interface in the probe coordinate system, thus providing a raw profile of the surface of the object.

The interface determination module 11 may further comprise a programmable filtering unit of the raw profile. The filtering unit can notably implement a programmable median filtering to eliminate aberrations in the raw profile, followed by a programmable averaging filter and a programmable polynomial regression. These different operations are configurable so that they can be adapted to different types of object surfaces. The parameters are stored in the memory 21, and include as examples the size of the averaging window and the coefficients used for the regression.

The method of constructing the image of the object lo is identical to that used for the construction of the surface image ls. However, for the image of the object lo, the flight times are calculated by the module 13b, different from the module 13a, which takes into account the profile. Refraction occurs at the interface between the two media; thus, the path between the translator element and the pixel no longer returns to that of a simple Euclidean distance.

With reference to FIG. 3c, the Fermat principle makes it possible to determine that the coordinates of the point of the interface M, of coordinates (x, z) with z (x) the profile extracted by the unit 11, interface determination, where the ultrasonic path is refracted, are those for which the flight time is minimal.

For a point P (x2, Z2) of the object, one thus seeks the minimum of:

This minimum can be searched by the method of the descent of the gradient (k being the number of the iteration):

The computation of the first and second derivatives can be performed numerically according to:

With ε the digital step.

The choice of the initialization value for X ^ 0) may be that of the abscissa of the element of the transducer considered x1. Two iterations are sufficient to calculate the flight time. In order to be able to generate a flight time per clock stroke, and thus optimize the overall computation time, the entire calculation pipeline can be parallelized. The invention is not limited to the previously described system-on-a-chip, but also extends to a method implemented by such a system and in particular to an ultrasonic imaging method of a region of interest of a object implemented by a system-on-a-chip, comprising: - acquiring ultrasonic signals from a probe comprising a plurality of ultrasonic transducers, the acquisition being performed so as to acquire the signals received by transducers as echoes of a sequence of successive transmissions of an ultrasonic signal by at least one transducer of the probe; the iterative realization for each transmission of said sequence of successive transmissions: calculation for the transmission considered, for each point of the image and for each transducer, of a flight time corresponding to the signal path of the transmission; emission passing through the point and reaching the transducer; calculating an intermediate image by summing, at each point of the image, the amplitude of the signal received by each transducer at the instant corresponding to the calculated flight time corresponding to the signal path of the transmission passing through the point and reaching the transducer; and when there is a cumulative image of a previous iteration, summing the calculated intermediate image to the cumulative image of the previous iteration to form a cumulative image of the iteration.

Claims (11)

A system on a chip (1) for ultrasound imaging a region of interest (Z) of an object (O), comprising: - an acquisition unit (2) of ultrasonic signals from a probe comprising a plurality of ultrasonic transducers (ERi, ERj), the acquisition unit being configured to acquire the signals received by transducers as echoes of a sequence of successive transmissions of an ultrasonic signal by at least one transducer the probe; an acquired ultrasonic signal processing unit, which comprises: a module (8) for calculating flight time; a module (9) for producing images; and o an order sequencer (10) configured to instruct the flap time calculation module (8) at the image generation module (9) to reconstruct an image of an image area by iteratively proceeding to each transmission of said sequence of successive transmissions: to the computation, by the module (8) of calculation of flight time, for the emission considered, for each point (P) of the image and for at least one transducer (ERj ), a flight time [TP + TP; Tpn + TPA corresponding to the signal path of the emission passing through the point and reaching the at least one transducer; calculating, by the image forming module, an intermediate image by summation, at each point of the image, of the amplitude of the signal received by the at least one transducer at the instant corresponding to the time flight calculated by the flight time calculation module corresponding to the signal path of the emission passing through the point and reaching the at least one transducer; and when there is a cumulative image of a preceding iteration, at the summation, by the image processing module, of the intermediate image computed with the cumulative image of the preceding iteration to form a cumulative image of iteration. A system-on-a-chip system according to claim 1 for imaging a region of interest (Z) of an object (O) separated from the probe via a coupling medium (Mc), wherein the processing unit further comprises an interface determination module (11), and wherein the command sequencer is configured to: command the flight time calculation module (8) and the flight time module (9); developing images to reconstruct an image of the imaged zone with consideration by the flight time calculation module of a propagation of the ultrasonic signal in the single coupling medium, the cumulative image of the last iteration formed by the module developing images constituting a surface image (ls); instructing the interface determining module (11) to determine, from the surface image, an interface between the coupling medium and the object; instructing the flap time calculation module (9) in the image processing module to reconstruct an image of the imaged zone with consideration by the flight time calculation module of the interface determined by the module interface determination, the cumulative image of the last iteration formed by the image forming module constituting an image of the region of interest of the object (lo). The system-on-chip system of claim 2, wherein the interface determination module (11) comprises a surface image analysis unit configured to determine a raw profile of the surface of the object and a display unit. programmable filtering of the raw profile. The system-on-chip system of claim 3, wherein the interface determination module (11) is embedded in a programmable microprocessor. The system-on-chip system according to one of claims 1 to 4, wherein the flight time calculation module (8) comprises a flight time calculation unit (13a, 13b, 13) and an interpolator (15). configured to determine interpolated flight times from the flight times calculated by the flight time calculation unit. The system-on-chip system according to one of claims 1 to 5, further comprising a write unit (3) for transferring the signals acquired by the acquisition unit (2) to a memory unit (4), the write unit being configured to interleave, for each group of a plurality of groups of transducers, the signals acquired by the transducers of the group and to transfer the interlaced signals of the group to the memory unit (4). A system-on-a-chip system according to claim 6, further comprising a read unit (5) for reading the signals stored in the memory unit (4) and a filter unit (6) of the signals read by the storage unit. reading (5), the read unit being configured to deinterlace the signals of each of the groups and the filter unit being configured to filter the signals of each of the groups in parallel. 8. System according to claim 7, wherein the filter unit (6) is configured to implement a high-pass filtering of the signals read by the reading unit (5). 9. System according to one of claims 7 and 8, wherein the filter unit (6) is configured to implement an interpolation of the signals read by the reading unit (5). 10. System according to one of claims 7 to 9, wherein the filter unit (6) is configured to implement a calculation of the Hilbert transform of the signals read by the reading unit (5). 11. A method of ultrasonic imaging of a region of interest (Z) of an object implemented by a system-on-a-chip, comprising: the acquisition of ultrasonic signals from a probe comprising a plurality of ultrasonic transducers, the acquisition being performed in such a way as to acquire the signals received by transducers as echoes of a sequence of successive transmissions of an ultrasound signal by at least one transducer of the probe, iteratively being carried out for each emission of the sequence of successive transmissions: calculation for the emission considered, for each point (P) of the image and for at least one transducer (ERj), a flight time [TP + TP; TP + TP) corresponding to the signal path of the emission passing through the point and reaching the at least one transducer; calculating an intermediate image by summing, at each point of the image, the amplitude of the signal received by the at least one transducer at the instant corresponding to the calculated flight time corresponding to the signal path of the emission passing through the point and reaching the at least one transducer; and when there is a cumulative image of a previous iteration, summing the calculated intermediate image to the cumulative image of the previous iteration to form a cumulative image of the iteration.