WO2020233798A1

WO2020233798A1 - Ultrasound detection using phase information

Info

Publication number: WO2020233798A1
Application number: PCT/EP2019/063116
Authority: WO
Inventors: Relja BJELOGRLIC; Cengiz Tugsav KÜPÇÜ; Ralph Thorsten Mennicke
Original assignee: Proceq Sa
Priority date: 2019-05-21
Filing date: 2019-05-21
Publication date: 2020-11-26

Abstract

A method for ultrasound detection of a reflector (1) comprises the following steps: (A) Receiving at least one raw signal s_r from an ultrasound transducer (3). The at least one raw signal represents at least one ultrasound wave (4) reflected by the reflector (1). (B) Deriving a first signal s_i and a second signal s₂ from the at least one raw signal s_r. The first signal s_i has a first frequency spectrum, and the second signal s₂ has a second frequency spectrum different from the first frequency spectrum. (C) Determining an approximate location lapp of the reflector (1) from an envelope s_m of at least one of the signals. (D) Determining a refined location l_ref of the reflector (1) using an algorithm in which the refined location l_ref depends on the approximate location lapp and on a plurality of samples of the first signal s_i and/or of the second signal s₂.

Description

Ultrasound detection using phase information

Technical Field

The present invention relates to a method for ultrasound detection of a reflector, in particular in a surrounding medium which may e.g. be concrete, wood, as- phalt or ground.

Background Art

Ultrasound detection of a reflector is often used in non-destructive testing applications such as checking the integrity of concrete structures. Another application is directed to imaging the near-surface ground, e.g. for prospection in view of a planned excava- tion of ground. The reflector may for instance be a me- tallic structure, e.g. a rebar, or a void in the concrete or in the ground. In any case, the reflector has an acoustic impedance which differs from the acoustic imped- ance of the surrounding medium, i.e. there is a change of acoustic impedance at the reflector, also called an acoustic impedance contrast, leading to a generation of a reflected ultrasound wave from an incoming ultrasound wave. The reflected ultrasound wave is recorded by an ul- trasound transducer as a raw signal.

In the above-mentioned applications, it is of interest to infer information about the reflector. For instance it is useful to determine a location and proper- ties of the reflector, e.g. by discriminating between a reflector with a higher acoustic impedance compared to the surrounding medium, such as a metallic object, such as a rebar or a water pipe, in concrete, and a reflector with a lower acoustic impedance compared to the surround- ing medium, such as a void in concrete. Such information may be inferred by pro- cessing the raw signal. A method for obtaining polarity and phase information from the raw signal is described in US Patent 5,042,303. The raw signal is amplified with a specific time-dependent gain function, which is derived from the raw signal itself and is independent of the po- larity of the raw signal. Operations for deriving the specific gain function according to US Patent 5,042,303 comprise taking a modulus of the raw signal or generating a time-shifted version of the raw signal. However, the accuracy of the location of the reflector determined by conventional methods may not be sufficient for some ap- plications. Moreover, an inaccuracy in the location may lead to further inaccuracies when inferring other proper- ties of the reflector, such as the above-mentioned dis- crimination according to the acoustic impedance.

It is hence an object of the present inven- tion to provide a method for ultrasound detection of a reflector which yields an accurate location of the re- flector, and preferably a reliable discrimination between a reflector with a higher acoustic impedance compared to the surrounding medium, and a reflector with a lower acoustic impedance compared to the surrounding medium.

Disclosure of the Invention

The object is achieved by a method for ultra- sound detection of a reflector according to the invention comprising the following steps: At least one raw signal is received from an ultrasound transducer. The at least one raw signal represents at least one ultrasound wave reflected by the reflector. The ultrasound wave usually is a P-wave or an S-wave, i.e. particles of the medium oscillate parallel or perpendicular to the propagation direction, respectively. The reflector is particularly defined as above, namely as having a different acoustic impedance compared to a surrounding medium. The acoustic impedance of a medium is defined as the product of a den- sity of the medium and a speed of the ultrasound wave in the medium. Upon incidence of an ultrasound wave on the reflector, a change in acoustic impedance, also called an acoustic impedance contrast, leads to a generation of a reflected ultrasound wave. Typical reflectors in applica- tions, such as non-destructive testing or near-surface ground prospection, are metallic objects, such as a rebar or a pipe, or a void in concrete, in wood or in the ground .

Then a first signal and a second signal are derived from the at least one raw signal. The first sig- nal has a first frequency spectrum, and the second signal has a second frequency spectrum different from the first frequency spectrum. Such derivation of the first signal and the second signal may be achieved in different ways, such as using different raw signals representing differ- ent ultrasound waves with different frequency spectra, or such as deriving the first signal and the second signal from the same raw signal representing the same ultrasound wave but modifying the raw signal differently such that the resulting first signal and second signal have differ- ent frequency spectra, e.g. by filtering. Such different ways are dealt with in more detail below.

Further, an approximate location of the re- flector is determined from the envelope of at least one of the signals, i.e. from one or more of the raw signals, the first signal, and the second signal. For such deter- mination of the approximate location, several options ex- ist. In an embodiment, the approximate location of the reflector is determined by calculating a maximum of the envelope of at least one of the signals.

Finally, a refined location of the reflector is determined using an algorithm in which the refined lo- cation depends on the approximate location and on a plu- rality of samples of the first signal and/or of the sec- ond signal. The refined location is understood to be more accurate than the approximate location, and in particular closer to an actual location, i.e. the real location of the reflector within the surrounding medium. The samples represent values of the corresponding signal at defined times. In many cases, the signals will comprise time-dis- crete samples, in particular with a certain sampling rate .

The described method has several advantages over the prior art. Firstly, it leads to a more accurate location of the reflector due to iterative determination of the approximate location and the refined location de- pending on the approximate location. The iterative deter- mination also makes the method robust and reliable. Sec- ondly, the method makes use of different parts of infor- mation about the reflector comprised in the at least one raw signal: On the one hand, different frequency compo- nents are exploited by using the envelope of at least one of the signals as well as the first signal and/or the second signal. On the other hand, a plurality of samples of the first signal and/or the second signal is used, meaning that the determination of the refined location depends on non-local information in the signals. This again makes the refined location more accurate and relia- ble, and in particular less susceptible to noise within the raw data.

In a preferred embodiment, the exploitation of non-local information is done in the following way: A first phase signal is derived from the first signal, and a second phase signal is derived from the second signal. The refined location is then determined dependent on the first phase signal and the second phase signal. Phase signals may in general be considered to comprise non-lo- cal information. In a further embodiment, the refined lo- cation is determined by finding a zero point (zero cross- ing) of the difference between the first phase signal and the second phase signal. More details and variants for the determination of the refined signal are given below. The embodiment of the method for ultrasound detection of a reflector as described above uses the as- sumption that there is no dispersion of the at least one ultrasound wave in the surrounding medium, i.e. that dif- ferent frequency components of the at least one ultra- sound wave travel with the same speed. However, given that a dispersion relation for the ultrasound wave in the surrounding medium is known, it is also possible to cor- rect for the dispersion and still apply the described method .

In general, the refined location may also be determined dependent on zero points of the amplitudes of the first signal and the second signal. In an embodiment, a zero point of the difference between the amplitudes of the first signal and the second signal in the vicinity of the approximate location may be taken as the refined lo- cation .

Advantageously the method also comprises the step of emitting, by an ultrasound emitter, the at least one ultrasound wave with a peak frequency. The ultrasound emitter may be the same as the ultrasound transducer above. The peak frequency is understood to be the fre- quency component which is largest, i.e. has the largest energy, within the frequency spectrum of the ultrasound wave. In other words, the peak frequency is represented by a peak in the frequency spectrum; it is also known as the dominant frequency. The peak frequency may in partic- ular be defined or be set through an input signal to the ultrasound emitter. In particular, the peak frequency is between 20 kHz and 1 MHz, preferably between 20 kHz and 500 kHz.

Further, the method advantageously comprises the following step: A parameter indicative of the change of the acoustic impedance at the reflector is determined by evaluating one of the signals in a defined interval around the refined location of the reflector. As dis- cussed earlier, the change of the acoustic impedance at the reflector may be used to characterize the reflector.

In an embodiment, the parameter is indicative of the sign of the change of the acoustic impedance. In particular, such sign indicates whether the reflector has a higher or a lower acoustic impedance than the surround- ing medium. Hence the method advantageously comprises the following step: discriminating between a reflector having a higher acoustic impedance than a surrounding medium and a reflector having a lower acoustic impedance than the surrounding medium using the parameter. Thus it is for instance feasible to discriminate between a metallic ob- ject and a void in concrete or in the ground.

The method described above and detailed fur- ther below is preferably used for ultrasound detection of a reflector where the surrounding medium is one or more of concrete, wood, metal and ground.

The present invention further relates to: a data processing device comprising a processor configured to perform the described method; a computer program com- prising instructions which, when the program is executed by a computer, cause the computer to carry out the method; a computer-readable data carrier having stored thereon the computer program; as well as a system com- prising the ultrasound transducer and the data processing device .

Other advantageous embodiments of the method for ultrasound detection of a reflector are listed in the dependent claims as well as in the description below.

Brief Description of the Drawings

Embodiments of the present invention, aspects and advantages will become apparent from the following detailed description thereof. The detailed description makes reference to the annexed drawings, wherein the fig- ures show:

Fig. 1 a time series of a first signal and a second signal, also called an A-scan, according to an em- bodiment,

Fig. 2 the time series of Fig. 1 together with the envelope of the first signal,

Fig. 3 a time series of the first phase sig- nal and the second phase signal according to an embodi- ment,

Fig. 4 a time series of the difference be- tween the first phase signal and the second phase signal,

Fig. 5 a brightness plot of numerous enve- lopes of first signals s_m as the one of Fig. 2, also called a B-scan, together with the refined location of the reflector according to an embodiment, and

Fig. 6 a schematic view of a system for ul- trasound detection of a reflector according to an embodi- ment .

Detailed Description of the Drawings

Same elements are referred to by same refer- ence numerals across all figures. The method for ultra- sound detection of a reflector is illustrated by means of exemplary signals in Figs. 1-5. Further options and gen- eralizations of the method are explained on the way.

Fig. 6 schematically shows a system for ul- trasound detection of a reflector 1 within a surrounding medium 2 by means of ultrasound waves 4. The system ad- vantageously comprises an ultrasound transducer 3, a pro- cessor 5 and preferably a display 6. These elements are detailed later.

First signal and second signal Fig. 1 shows a time series of a first signal s₁ (solid line) and a second signal S2 (dashed line) ac- cording to an embodiment. The ordinate of the time series shows amplitudes of the first signal s₁ and the second signal s_å in arbitrary units. The abscissa indicates time in samples. It is straightforward to convert time to depth measured from the ultrasound transducer 3 in order to find the depth location of the reflector 1 assuming that the ultrasound velocity in the medium 2 surrounding the reflector 1 is known.

Information about the reflector 1, in partic- ular about the location of the reflector 1, is e.g. found between samples 200 and 300 in Fig. 1. In this sample range, the amplitude of the first signal s₁ and the sec- ond signal S2 are higher compared to the amplitudes at earlier times. This is due to the fact that an ultrasound wave 4 emitted at time sample 0 has travelled through the surrounding medium 2 to the reflector 1, it is reflected at the reflector 1 and travels (back) to the ultrasound transducer 3, where it arrives between samples 200 and 300. Consequently, the information in this sample range is going to be exploited by the method according to the invention in order to detect the reflector 1 and deter- mine its properties.

In the embodiment of Fig. 1, the first signal s₁ and the second signal S2 are derived from one raw sig- nal s_r from an ultrasound transducer 3, i.e. they stem from the same ultrasound wave 4 that has been reflected by the reflector 1 of interest. In particular, the first signal s₁ equals the raw signal S₁ = s_r, and the second signal S₂ is derived by applying a filter to the raw sig- nal s_r, in this case a high-pass filter. In general, the first signal s₁ and the second signal S2 may both be dif- ferent from the raw signal s_r due to an arbitrary deriva- tion step, which preferably comprises applying a linear filter. A requirement for the method according to the invention is that the first signal s₁ has a first fre- quency spectrum, the second signal S2 has a second fre- quency spectrum, and the two spectra differ from each other .

In this way, the first signal s₁ and the sec- ond signal S2 may contain slightly different information about the same reflector 1 and surrounding medium 2, which improves the determination of a refined location of the reflector 1 as well as advantageously the determina- tion of material properties.

Preferably, the method comprises the step of emitting the at least one ultrasound wave 4 with a peak frequency f_p. The peak frequency f_p is in particular be- tween 20 kHz and 1 MHz, preferably between 20 kHz and 500 kHz, more preferably between 20 kHz and 120 kHz. The ultrasound wave 4 is emitted by the ultrasound emitter, such as a piezoelectric transducer or a capacitive trans- ducer. The ultrasound emitter may advantageously be the same as the ultrasound transducer 3 receiving the re- flected ultrasound wave 4. In an embodiment, the at least one emitted ultrasound wave is a wavelet comprising one period of a sine oscillation with peak frequency f_p or a Ricker wavelet.

Alternatively, the first signal S₁ and the second signal S2 may be derived from more than one raw signal s_r. For instance, the first signal s₁ may represent a first ultrasound wave, and the second signal S2 may represent a second ultrasound wave.

In an embodiment, the method comprises emit- ting by at least one ultrasound emitter a first ultra- sound wave with a first peak frequency f¹ _r and a second ultrasound wave with a second peak frequency f² _r different from the first peak frequency f¹ _r. Preferably the peak frequencies fⁿ _r with n=l,2 again take on values within the ranges given above. The first signal s₁ is derived from a first raw signal corresponding to the first ultrasound wave, and the second signal S2 is derived from a second raw signal corresponding to the second ultrasound wave.

In the case of more than one ultrasound wave, a requirement for the method to work is that the more than one ultrasound waves are reflected by the reflector 1 and preferably have a similar travel path through the surrounding medium 2. Even if the travel path is not sim- ilar, i.e. longer or shorter, or simply different, it is also possible to apply the method after shifting at least one of signals s₁ or S2 in time, in order to compensate the difference in its time of arrival, similar as in the case of non-zero dispersion as mentioned above. Again this has the effect that the signals derived from the different ultrasound waves contain information about the same reflector 1 and surrounding medium 2. This may be exploited e.g. for a determination of further parameters, such as location and material properties, or for noise cancelling .

Coming back to the embodiment shown in Fig. 1, only one ultrasound wave 4 with a peak frequency f¹ _r is emitted. The second signal S2 is derived by applying a filter to the raw signal s_r, in particular a linear fil- ter. As described above, the first signal s₁ of Fig. 1 may be identical to the raw signal s_r.

Preferably, the filter has a lower transmis- sion at the peak frequency f¹ _r than at a frequency double the peak frequency, 2f¹ _r. In particular the transmission at the peak frequency f¹ _r is no more than 0.8 of the transmission at the frequency 2f¹ _P, preferably no more than 0.5, This has the effect that the filter damps the peak frequency f¹ _r compared to the second harmonic at 2f¹ _r. In the example of Fig. 1, the dominant frequency of the second signal S2 is higher than the peak frequency f¹r of the first signal S₁. This is for instance achieved by applying a high-pass filter with a corner frequency above the peak frequency f¹ _r. Determination of the approximate location

The method according to the invention com- prises the step of determining an approximate location l_app of the reflector 1 from the envelope s_m of at least one of the signals, i.e. of the at least one raw signal s_r or the first signal s₁ or the second signal S2. The en- velope S_m of an oscillating signal is as usual understood as a smooth curve outlining the extremes, i.e. local max- ima and minima, of the signal.

As an example, Fig. 2 shows the first signal s₁ and the second signal s_å of Fig. 1 together with the envelope s_m (dash-dotted line) of the first signal Si. Preferably, the approximate location l_app of the reflector 1 is determined by calculating the maximum rnax[s_m] of the envelope s_m of at least one of the signals, as for in- stance illustrated in Fig. 2.

The envelope s_m may be calculated in differ- ent ways. Preferably the envelope s_m is calculated by de- riving a first analytic signal s_1a of the first signal s₁. Herein, the analytic signal s_a(t) of a real-valued func- tion s(t) is defined as s_a(t) = s(t) + i H[s(t)] with i being the imaginary unit and H[s(t)] being the Hilbert transform of s(t). Such analytic signal s_a may also be expressed in terms of time-variant magnitude and phase angle: s_a(t) = s_m(t) exp(i phi(t)), wherein s_m(t) =

|s_a(t) I is the envelope, also called instantaneous ampli- tude, and the phase angle phi(t) = arg[s_a(t)] is also called the instantaneous phase.

A different approach (currently not claimed) for determining the approximate location of the reflector is to simply calculate the maximum absolute value of one of the signals, i.e. of the at least one raw signal s_r or the first signal S₁ or the second signal S₂. In the exam- ple of Fig. 2, the approximate location determined in this way would only slightly differ from the approximate location l_app as determined from the envelope, namely by roughly 5 time samples. Determination of the refined location

As a further step, the method according to the invention comprises determining a refined location l_ref of the reflector 1 using an algorithm in which the refined location l_ref depends on the approximate location l_app and on a plurality of samples of the first signal S₁ and/or of the second signal S₂.

In a preferred embodiment, the algorithm com- prises deriving a first phase signal phii from the first signal s₁ and a second phase signal phi₂ from the second signal S2, and determining the refined location l_ref de- pendent on the first phase signal phii and the second phase signal phi₂.

An example for the implementation of these steps is illustrated in Figs. 3 and 4. Fig. 3 shows a time series of the first phase signal phi₁ and the second phase signal phi₂ according to an embodiment. The de- picted phase signals are derived from the first signal s₁ and the second signal S2 shown in Fig. 1; their values vary over time between -pi and +pi .

Advantageously, the first phase signal phii is calculated from a first analytic signal s_1a of the first signal s₁, and the second phase signal phi₂ is cal- culated from a second analytic signal S2_a of the second signal S2. The analytic signal s_a may again be defined as above. In particular the first analytic signal s_1a and the second analytic signal S2_a may be calculated by using the Hilbert transform. In that case the phase signal is taken to be the instantaneous phase phi(t).

In a preferred embodiment, the refined loca- tion l_ref is determined by finding the position of a zero point of the difference between the first phase signal phii and the second phase signal phi₂- The zero point may be found by conventional methods, such as Newton's method or more sophisticated versions thereof. Fig. 4 shows an embodiment of finding a zero point of such difference between the first phase signal phii and the second phase signal phi₂ of Fig. 3. In Fig.

4, the sine of the difference is plotted: sin(phii - phi₂) . Due to the periodicity of the sine function, the phase jumps between -pi and +pi or vice versa, which would otherwise lead to spurious zero point, are automat- ically taken care of.

In general, further options for finding zero points of the difference exist: In an embodiment, the first phase signal phii and the second phase signal phi₂ are first unwrapped, i.e. phase jumps between -pi and +pi or vice versa are removed, and a smooth phase signal is generated, before taking the difference.

Advantageously, a zero point of the differ- ence of the two phase signals is only taken into account if the slope of the difference at the zero point is smaller than a threshold. This, again, has the effect of excluding spurious zero points of the phase difference which may result e.g. from phase jumps between -pi and +pi or from an incorrect unwrapping of the first phase signal phii or the second phase signal phi₂.

In an embodiment, a zero point is only taken into account if it is located within a defined range around the approximate location l_app. In particular, the defined range around the approximate location l_app may be defined as the range in which the envelope s_m of the at least one of the signals exceeds an amplitude threshold. The amplitude threshold is preferably chosen between 20% and 80% of the maximum of the envelope, in particular 50%. Alternatively, the range around the approximate lo- cation lapp may be defined in time domain between 1/f¹ _p or 1/(2f¹ _r) before and after the approximate location l_app.

Further it is advantageous if at least some of the samples for determining the refined location lref correspond to times differing by at least l/(4f¹ _r), advan- tageously by at least 1/ (2f¹ _r) , more advantageously by at least 1/f¹ _p. Such criterion implies that the refined loca- tion l_ref is determined from non-local information in the signals. Actually, a similar reasoning is also behind the idea of using the envelope s_m for determining the approx- imate location l_app and the idea of deriving analytic sig- nals s_a, which again imply non-local information. The use of non-local information in general makes the method less susceptible to noise and other flaws in the raw signal s_r. Hence the determination of the location of the re- flector as well as a subsequent determination of material properties are more reliable as well as more accurate and precise .

The rationale behind using a first signal s₁ and a second signal S2 having different frequency spectra for the determination of the refined location l_ref of the reflector may be understood as follows: When an incident ultrasound wave hits a point on the reflector 1 at a cer- tain time, this point may be considered as a virtual source which emits the reflected wave 4 at the certain time. All components, and in particular different fre- quency components, of the ultrasound wave 4 are emitted from that point at the certain time. Under the assumption of no dispersion, all frequency components have the same travel time from the virtual source to the ultrasound transducer 3 receiving the ultrasound wave 4. Conversely, the different frequency components as measured by the ul- trasound transducer 3 should exhibit the same phase at the location of the reflector 1, i.e. the location of the virtual source of the reflected ultrasound wave 4. In the embodiment above, the peak frequency f¹ _r, i.e. the funda- mental frequency of the ultrasound wave 4, and the fre- quency 2f¹ _r of its second harmonic are chosen as example frequencies since they are naturally among the strongest frequency components of the ultrasound wave 4, and hence lead to a robust working of the method.

Information about the acoustic impedance Besides retrieving a refined location l_ref of the reflector 1, it may also be of interest to character- ize the reflector 1 in terms of material parameters. The propagation of ultrasound waves in a medium heavily de- pends on the acoustic impedance of the medium as de- scribed above. In particular, a reflector 1 is character- ized by a difference between the acoustic impedance of the reflector 1 and the acoustic impedance of the sur- rounding medium 2. Hence it is possible to retrieve in- formation about the nature of the reflector 1, such as "harder" or "softer" than the surrounding medium 2, from reflected ultrasound waves. If the surrounding medium 2 is concrete or ground, "harder" materials, i.e. having a higher acoustic impedance, may be stones or metallic ob- jects such as pipes, and "softer" materials, i.e. having a lower acoustic impedance, may be voids, e.g. filled with air or water.

In an advantageous embodiment, the method comprises the step of determining a parameter indicative of the change of the acoustic impedance at the reflector 1 by evaluating one of the signals in a defined interval around the refined location l_ref of the reflector 1. The signal used for the evaluation may be any of the afore- mentioned signals; usually one choses the raw signal s_r or the first signal s₁. The defined interval may be cho- sen as a certain number of samples around the refined lo- cation lref. Advantageously, the defined interval is cho- sen dependent on the period T = l/f_p of the used signal, e.g. from l_ref to l/(2f¹ _r) above lref.

Advantageously, the parameter is indicative of the sign of the change of the acoustic impedance. In this way, the parameter may be used to discriminate be- tween reflectors that are "harder" and reflectors that are "softer" than the surrounding medium. In other words, this enables the method to further comprise the step of discriminating between a reflector having a higher acous- tic impedance than a surrounding medium and a reflector having a lower acoustic impedance than the surrounding medium 2 using the parameter.

In an embodiment, the parameter depends on the slope of the signal or on the integral of the signal within the defined interval. The slope of the signal or the integral of the signal within the defined interval may in particular be positive or negative depending on the signal and the refined location l_ref of the reflector 1. The sign depends on the reflector 1 being "harder" or "softer" than the surrounding medium 2 in view of the following consideration: Ultrasound waves incur a phase reversal, i.e. a phase change by pi, if reflected at a "softer" reflector, whereas they incur no phase change if reflected at a "harder" reflector.

Examples for the determination of the parame- ter indicative of the sign of the change of an acoustic impedance at the reflector 1 comprise the following:

- Determining an integral or a mean of the signal in the defined interval as defined above and tak- ing its sign.

- Determining a median of the signal in the defined interval and taking its sign.

- Determining the slope of the signal at the refined location l_ref of the reflector 1 and taking its sign: This may either done locally by taking the differ- ence between two samples on both sides of the refined lo- cation lref, or it may be achieved by more non-local in- formation, e.g. by performing a linear or non-linear re- gression of the signal within a certain interval around the refined location l_ref e.g. between 1/(2f¹ _r) or

1/ (4f¹ _r) above and below l_ref·

Displaying the retrieved information

In an embodiment, the method further com- prises the step of displaying on a display 6 the refined location of the reflector as a mark 7 within a B-scan showing multiple raw signals s_r or within an image of processed multiple raw signals. In particular, the multi- ple raw signals s_r may be processed with a SAFT (syn- thetic aperture focusing technique) algorithm, e.g. con- ventional SAFT, FT-SAFT, or TFM (total focusing method) . Such B-scan or other image of processed multiple raw sig- nals is a conventional way of displaying ultrasound data.

Fig. 5 shows an example of such B-scan as a brightness plot of numerous signals s_m including the one shown in Fig. 2. The grey-scale indicates, with its brightness, the envelope of the signals s_m. The ordinate indicates time in samples, and the abscissa gives the number of measurement, wherein subsequent measurements are taken along a line on the surface of the surrounding medium 2. Additionally, Fig. 5 shows the refined location lref of the reflector as the mark, in particular a line or mark 7, according to an embodiment.

In an embodiment, the color or the style of the mark 7 depends on the parameter, in particular the parameter indicative of the sign of the change of the acoustic impedance at the reflector. This supports the visual discrimination between a ("harder") reflector hav- ing a higher acoustic impedance than a surrounding medium and a ("softer") reflector having a lower acoustic imped- ance than the surrounding medium 2 by a user. In the ex- ample of Fig. 5, the refined location l_ref of the reflec- tor 1 is depicted as a white line 7, indicating a

"harder" reflector, in this case a metallic objects in concrete .

The described method for ultrasound detection of a reflector 1 may be used in various applications, such as for non-destructive testing, utility localization or near-surface ground prospection. In other words, the method is not intended for use in medical diagnostics; it is preferably used for material comprising less than 50% of organic tissue.

The method may be computer-implemented, and it may be automated. For that purpose, a data processing device comprises a processor 5 configured to perform the described method.

As an advantageous aspect, a system for ul- trasound detection of a reflector is provided which com- prises the following elements, see Fig. 6: an ultrasound transducer 3 as discussed above and the above-mentioned data processing device comprising a processor 5. Prefera- bly, the system further comprises a display 6, e.g. for displaying the data in a way similar to Fig. 5. The dis- play 6 may be arranged in the same device as the ultra- sound transducer 3 and the data processing device. Alter- natively, the elements may be separated into different devices. In particular the display 6 may be arranged re- motely, e.g. in a tablet computer, connected via a cable or a wireless connections to the data processing device.

While above there are shown and described em- bodiments of the invention, it is to be understood that the invention is not limited thereto but may be otherwise variously embodied and practiced within the scope of the following claims.

Claims

1. A method for ultrasound detection of a re- flector (1) comprising the steps of

- receiving at least one raw signal (s_r) from an ultrasound transducer (3) , the at least one raw signal (s_r) representing at least one ultrasound wave (4) re- flected by the reflector (1),

- deriving a first signal (s₁) and a second signal (S2) from the at least one raw signal (s_r) ,

wherein the first signal ( s₁ ) has a first frequency spectrum, and the second signal (S2) has a sec- ond frequency spectrum different from the first frequency spectrum,

- determining an approximate location (l_app) of the reflector (1) from an envelope (s_m) of at least one of the signals (s_r, s₁, S₂) ,

- determining a refined location (l_ref) of the reflector (1) using an algorithm in which the refined lo- cation (lref) depends on the approximate location (l_app) and on a plurality of samples of the first signal (s₁) and/or of the second signal (S₂).

2. The method according to claim 1,

wherein the algorithm comprises

- deriving a first phase signal (phii) from the first signal (s₁) and a second phase signal (phi₂) from the second signal (S₂) ,

- determining the refined location (l_ref) de- pendent on the first phase signal (phii) and the second phase signal (phi₂) .

3. The method according to claim 2,

wherein the first phase signal (phii) is cal- culated from a first analytic signal (s_1a) of the first signal (s₁) , and wherein the second phase signal (phi₂) is calculated from a second analytic signal (S2a) of the sec- ond signal ( S2 } ,

in particular the first analytic signal (s_1a) and the second analytic signal (S2a) being calculated by using a Hilbert transform.

4. The method according to claim 2 or 3, wherein the refined location (lref) is deter- mined by finding a zero point of a difference between the first phase signal (phii) and the second phase signal (phi₂) ,

in particular wherein the zero point is only taken into account if a slope of the difference at the zero point is smaller than a threshold.

5. The method according to claim 4,

wherein the zero point is only taken into ac- count if it is located within a defined range around the approximate location (l_app) ,

in particular wherein the defined range around the approximate location (lapp) is defined as the range in which the envelope (s_m) of the at least one of said signals (s_r, s₁, s₂) exceeds an amplitude threshold.

6. The method according to any of the preced- ing claims, comprising the step of

- emitting by an ultrasound emitter the at least one ultrasound wave 4 with a peak frequency fⁿ _p with n=1,...,

in particular the peak frequency fⁿ _p being be- tween 20 kHz and 1 MHz.

7. The method according to any of the preced- ing claims, wherein at least some of the samples for de- termining the refined location (Ire_f) correspond to times differing by at least 1/ (4f¹ _r) .

8. The method according to claim 6 or 7, wherein one ultrasound wave (4) with a peak frequency f¹ _r is emitted,

wherein the second signal (S2) is derived by applying a filter to the raw signal (s_r) , in particular a linear filter,

in particular wherein the first signal (s₁) is identical to the raw signal (s_r) .

9. The method according to claim 8,

wherein the filter has a lower transmission at the peak frequency f¹ _r than at a frequency double the peak frequency, 2f¹ _p,

in particular wherein the transmission at the peak frequency f¹ _p is no more than 0.8 of the transmission at the frequency double the peak frequency.

10. The method according to any of claims 1- 7, comprising the step of

- emitting by at least one ultrasound emitter a first ultrasound wave with a peak frequency f¹ _p and a second ultrasound wave with a peak frequency f² _r different from the first peak frequency f¹ _r,

wherein the first signal (s₁) is derived from a first raw signal corresponding to the first ultrasound wave, and the second signal (S₂) is derived from a second raw signal corresponding to the second ultrasound wave.

11. The method according to any of the pre- ceding claims comprising the step of

- determining the approximate location (l_apP) of the reflector (1) by calculating a maximum of the en- velope (sm) of at least one of the signals (s_r, s₁, S₂) , in particular wherein the envelope (s_m) is calculated by deriving a first analytic signal (s_1a) of the first signal (s₁) .

12. The method according to any of the pre- ceding claims, further comprising the step of

- determining a parameter indicative of a change of an acoustic impedance at the reflector (1) by evaluating at least one of the signals (s_r, s₁, S2) in a defined interval around the refined location (lref) of the reflector .

13. The method according to claim 12, wherein the parameter is indicative of a sign of the change of the acoustic impedance,

the method further comprising the step of

- discriminating between a reflector (1) hav- ing a higher acoustic impedance than a surrounding medium (2) and a reflector having a lower acoustic impedance than the surrounding medium (2) using the parameter.

14. The method according to claim 12 or 13, wherein the parameter depends on a slope of the signal (s₁, s₁, S₂) or on an integral of the signal (s_r, s₁, S2) within the defined interval.

15. The method according to any of the pre- ceding claims, further comprising the step of

- displaying on a display (6) the refined ID- cation (lref) of the reflector as a mark (7) within a B- scan showing multiple raw signals (s_r) or within an image of processed multiple raw signals,

in particular wherein the multiple raw sig- nals are processed with a SAFT algorithm.

16. The method according to one of claims 12- 14 and claim 15, wherein a color or a style of the mark (7) depends on the parameter, supporting a visual discrimina- tion between a reflector (1) having a higher acoustic im- pedance than a surrounding medium (2) and a reflector having a lower acoustic impedance than the surrounding medium (2) by a user.

17. The method according to any of the pre- ceding claims for ultrasound detection of a reflector (1) in a surrounding medium (2),

wherein the surrounding medium (2) is one or more of:

- concrete,

- wood,

- metal,

- ground.

18. A data processing device comprising a processor (5) configured to perform the method according to any of the preceding claims .

19. A computer program comprising instruc- tions which, when the program is executed by a computer, cause the computer to carry out the method according to any of claims 1-17.

20. A computer-readable data carrier having stored thereon the computer program of claim 19.

21. A system for ultrasound detection of a reflector, comprising

the ultrasound transducer (1) of claim 1, the data processing device of claim 18, and

the display (6) of claim 15.