JPH10142334A - Ultrasonic measuring device - Google Patents

Abstract

(57) [Summary] [PROBLEMS] To suppress a side lobe without increasing the number of elements of a transducer in ultrasonic measurement. SOLUTION: In an ultrasonic transmitter / receiver 10, an ultrasonic wave is transmitted using one or a plurality of transmitting elements 12a, and a reflected wave is received by a plurality of receiving elements 14a. The plurality of received signals thus obtained are subjected to filter processing by a plurality of types of imaging filters, and a plurality of imaging processing results are obtained. The plurality of imaging filters are set so that azimuth gain characteristics other than the main azimuth are different from each other. Integrated processing unit 2
Reference numeral 4 compares and compares a plurality of imaging processing results to determine and suppress side lobes.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an ultrasonic measuring apparatus for measuring an object by transmitting and receiving ultrasonic waves,
The present invention relates to an imaging process of an ultrasonic signal.

[0002]

2. Description of the Related Art Various types of devices have been known as object measuring devices for measuring an object three-dimensionally and imaging the object. Among them, measuring devices that use optical systems include:
Optical properties of the surface of the target object, measurement environment (eg, illumination)
There are restrictions on measurement conditions such as

On the other hand, a measuring device using sound, particularly a measuring device for an object using ultrasonic waves, can generally be reduced in size and cost as compared with a measuring device using an optical system. In particular, the ultrasonic measurement device is advantageous in that the speed of the ultrasonic wave is slower than that of light, and it is easy to directly measure and control the phase and the like of the signal. Because of these advantages, ultrasonic measuring devices are used in various fields (for example, in the field of security), and there is a demand for low-cost, high-accuracy, and highly reliable ultrasonic measuring devices.

In an ultrasonic measuring apparatus, an object is measured by transmitting and receiving ultrasonic waves (ultrasonic pulses) in an ultrasonic transducer having a plurality of transmitting elements and receiving elements. Specifically, an ultrasonic wave is transmitted from each transmitting element, and reflected waves reflected by the object are received by a plurality of receiving elements. In transmitting and receiving the waves, focusing by adjusting the phase of the ultrasonic wave and scanning of the ultrasonic beam are performed. That is, a large number of reception signals obtained by each transmission / reception wave are subjected to image formation processing based on the positional relationship between the transmission element and the reception element, and three-dimensional data (distance data or shape data) of an object existing in the measurement area Is obtained. Note that a three-dimensional image to be measured is formed and displayed as needed.

If the target is stationary within the measurement time, it is not always necessary to actually focus the ultrasonic waves, and the transmitting elements are separately driven based on the known synthetic aperture method to transmit each ultrasonic wave. On the other hand, if the reflected wave is received by each of the receiving elements, and the received signal group obtained by these is electronically converged and scanned inside the computer, the same three-dimensional image as described above can be obtained.

[0006]

In the object measurement using ultrasonic waves as described above, it is necessary to use a plurality of elements (a plurality of transmitting elements and a plurality of receiving elements). In this case, the number of elements is reduced. When the element density is reduced to a certain value, when the distance between the elements and the signal phase difference satisfy predetermined conditions, a wavefront equivalent to the target azimuth may be generated in addition to the target azimuth. In the measurement image, there is a problem that a "false image" is generated in addition to a target object image (generation of a grating lobe), and measurement reliability is reduced. For example, when a false image occurs when the ultrasonic measurement device is used in the security field, there arises a problem that the accuracy of discriminating an object is reduced.

In order to avoid this, it is necessary to arrange the ultrasonic elements densely. However, for that purpose, it is necessary to arrange an ultrasonic element having an extremely small aperture surface, so that the transmission output and the reception sensitivity are reduced. There is a problem that arises. Also,
When a limited number of elements are densely arranged, the size of the aperture formed by the element array constituted by those elements becomes small, so that the width of the ultrasonic beam to be formed and the resolution related to the azimuth deteriorate. There is also a problem.

Therefore, it can be said that it is ideal to arrange the ultrasonic elements densely and in large numbers. In this case, however, the number of wirings and circuits must be increased as the number of elements increases, so that the scale of the apparatus increases. This leads to an increase in the cost of the apparatus.

Conventionally, techniques for obtaining an ultrasonic image with less false images by sparsely arranging elements include a technique using a wideband signal (Japanese Patent Laid-Open No. 5-150043) and a technique using a plurality of independent arrays (see JP-B-59-16224)
However, the former requires a very special broadband device,
In the latter case, it is necessary to use a plurality of array devices, which leads to an increase in the number of elements and the scale of the device. Therefore, none of them can achieve the object of the present invention.

By the way, the present inventors have devised a method of arranging elements, and virtually obtain an ultrasonic measurement apparatus which acquires phase information equal to or less than the element arrangement density and thereby obtains an ultrasonic image with less false images. It is proposed in Hei 8-1712. The present invention mainly resides in the structure of an ultrasonic transducer, but the present invention mainly relates to an imaging process.

The present invention has been made in view of the above-mentioned conventional problems, and has as its object to provide an ultrasonic measurement method to which a new imaging processing method capable of reducing a false image even with a small number of ultrasonic elements has been applied. It is to provide a device.

Another object of the present invention is to provide an ultrasonic measuring apparatus capable of reducing forgery with high accuracy even if there are variations in elements and errors in the circuit.

[0013]

[Means for Solving the Problems]

(1) In order to achieve the above object, the present invention provides a method of transmitting ultrasonic waves from one or a plurality of transmitting elements and receiving reflected waves by a plurality of receiving elements, thereby receiving signals corresponding to the respective receiving elements. An ultrasonic transducer that outputs a group of signals; imaging processing means for obtaining a plurality of imaging processing results while applying a plurality of filters having different filter characteristics to the received signal group; and Integrated processing means for performing integrated processing on the image processing result.

Here, the respective filter characteristics are set so that the azimuth gain patterns are different from each other in an azimuth range other than the main azimuth, and the false image generation patterns are different from each other in each of the imaging processing results. I do.

According to the above configuration, an ultrasonic wave is transmitted toward an object using one or a plurality of transmitting elements, and reflected waves are received by a plurality of receiving elements, whereby a plurality of received signals are captured. . By applying a plurality of filters having different convergence and scanning states to these received signal groups,
Obtain a plurality of imaging results with changing the state of signal convergence,
A new reconstructed image is formed by integrating them.

The plurality of convergence and scanning states are realized by, for example, switching the characteristics of a filter for signal processing in a computer, and the characteristics of the plurality of filters are designed so as to be mutually complementary in azimuth. If a plurality of such different filters are prepared and a signal group observed through each filter is passed, a plurality of different imaging results can be obtained independently. Virtually, these imaging results have different azimuth gain patterns depending on the characteristics of each filter, that is, different false image generation patterns. Of course, the component of the main azimuth is the same in each imaging processing result. The plurality of imaging results are integrated by the integration processing means, and the reflected sound from the target object and the false image are discriminated and suppressed.

Therefore, according to the ultrasonic measurement apparatus of the present invention, the same effect as if a plurality of arrays having different apertures were formed can be virtually obtained on a computer, so that the performance can be improved. There is no need to actually increase the number of array elements or the scale of the device. In particular, the aforementioned Japanese Patent Publication No. Sho 59-1
Conventionally, in order to obtain image data with different apertures, including the technique disclosed in Japanese Patent No. 6224, it is necessary to actually provide a plurality of independent arrays, and as a result, the apparatus scale and hardware increase. There was a problem. But,
Since the present invention does not include any increase in the ultrasonic transducer, it is advantageous in terms of cost and performance. Further, in the present invention, a false image portion is determined from a plurality of imaging results, and only that portion can be suppressed. A highly accurate three-dimensional image can be obtained. As a result, without using a large amount of special elements,
An ultrasonic measuring apparatus using a relatively large element that is usually used can be configured. Therefore, there is an advantage that high measurement accuracy can be maintained while preventing a decrease in transmission output and reception sensitivity.

The present invention is also applicable to a method of performing focusing and scanning of an ultrasonic wave by electronic delay of a transmission signal and a reception signal, or a method of performing such focusing and scanning of an ultrasonic wave by a calculation inside a computer. Applicable. The ultrasonic measurement device according to the present invention is preferably applied to measurement of an object in the air, but can also be applied to measurement of an object in water.

(2) In a preferred aspect of the present invention, the imaging processing means includes a conversion means for converting each of the received signals from a time axis to a frequency axis, and a conversion means for each of the received signals on the frequency axis. Filter processing means for applying a plurality of filters;
Inverse transforming means for transforming each imaging processing result after the filter processing from a frequency axis to a time axis.

In a preferred aspect of the present invention, the image processing means applies a plurality of filters to the respective received signals on a time axis.

That is, the filtering process is executed on the frequency axis or on the time axis. According to the former, although Fourier transform and its inverse transform are required,
The filter processing itself can be easily realized based on, for example, a matrix operation, and the design of the filter is also easy. According to the latter, conversion on the frequency axis and vice versa are not required, but convolution of the filter coefficient group is required.

(3) In a preferred aspect of the present invention, the integration processing means multiplies the plurality of imaging processing results by each other to obtain a difference therebetween, or
A false image is suppressed by taking a ratio between each other. In other words, the results of each imaging process are different from each other because the false image occurrence patterns are different from each other.
The level of false images will be different. Therefore, the false image is reduced and suppressed by using such a difference.

(4) In order to achieve the above object, the present invention provides a method for transmitting ultrasonic waves from one or a plurality of transmitting elements and receiving reflected waves by a plurality of receiving elements, so that each of the receiving elements receives a reflected wave. An ultrasonic transducer for outputting a corresponding received signal group, and a plurality of filters having azimuth gain patterns complementary to each other in an azimuth range other than the main azimuth acting on the received signal group to perform a plurality of imaging processes. Imaging processing means for obtaining a result;
Integrated processing means for performing integrated processing for false image suppression on the plurality of imaging processing results.

According to the above arrangement, a plurality of imaging processing results are generated using a plurality of filters complementary to each other in an azimuth range other than the main azimuth, and the nonlinear side lobes based on the complementarity of the filters are generated. The processing is utilized to reduce side lobes.

In a preferred aspect of the present invention, the integration processing means includes a sum and difference calculation means for calculating a sum and a difference of the plurality of imaging processing results, and a difference calculation means for calculating a difference between the sum and the difference. And

In a preferred aspect of the present invention, the integration processing means includes a sum and difference calculation means for calculating a sum and a difference of the plurality of imaging processing results, and calculates a sum of the sum and the difference by α.
(Where 0 <α), an exponential operation means, and a difference operation means for calculating the difference between the sum raised and the difference raised to the power of α.

By calculating the sum and difference of the results of a plurality of imaging processes, a signal containing only side lobes and a signal containing side lobes and main lobes can be obtained theoretically. If a difference between those absolute values or a difference between values obtained by performing an exponent operation on the absolute values is calculated, a signal of only the main lobe can be basically obtained. Here, if the exponential calculation is performed, the forgery suppression effect can be further improved. Considering the calculation amount, for example, 0.5 or 1 is selected as α. Of course, theoretically, the effect can be enhanced as α is reduced.

It is desirable to perform the absolute value calculation before the sum and difference calculations and before the exponent calculation. However, even if the absolute value calculation is not performed before the sum and difference calculation, the same result can be obtained by performing the absolute value calculation after the sum and difference calculation. Also,
When the absolute value operation is performed before the sum and difference operation, the absolute value operation before the sum exponent operation is not required. That is, the absolute value operation may be performed only before the difference exponent operation.

In a preferred aspect of the present invention, the integration processing means includes an absolute value calculation means for calculating absolute values of the plurality of imaging processing results, and an absolute value calculation means for calculating the absolute values of the plurality of imaging processing results. Minimum value selecting means for successively comparing the measurement points and selecting the smallest one.

As described above, if side lobe reduction processing is performed using a plurality of complementary filters and further using non-linear conversion (absolute value calculation and exponential calculation), errors in ultrasonic wave transmitting / receiving elements can be reduced. It is possible to configure a high-accuracy ultrasonic measurement device that is not so affected.

In a preferred embodiment of the present invention, the minimum value calculating means selects, for example, the smaller one of the two values.
It is a circuit that performs calculations. The absolute value calculating means is, for example, means for calculating the square of each image forming processing result.

When α is 1, a circuit configuration utilizing the min operation can be realized, and a device which requires a smaller amount of operation and can expect rapid processing can be constructed.

It is desirable to provide an imaging means for forming a three-dimensional image of the observed object based on the result of the integration processing.

[0034]

Preferred embodiments of the present invention will be described below with reference to the drawings.

[First Embodiment] FIG. 1 shows the overall configuration of an ultrasonic measuring apparatus according to the present invention. This ultrasonic measuring device can be used, for example, as a security device for monitoring an intruder. First, the overall configuration of the device will be described.

The ultrasonic transducer 10 of this embodiment is composed of two transmission element arrays 12 parallel to each other and two reception element arrays 14 parallel to each other and orthogonal to them. The transmitting element array 12 includes a plurality of transmitting elements 12a, and the receiving element array 14 includes a plurality of receiving elements 14a. The transmission element array 12 includes a transmission element driving unit 16
Are driven simultaneously or sequentially. Thereby, an ultrasonic pulse is emitted.

The ultrasonic wave that has collided with the target and reflected is
The wave is received by each of the receiving elements 14a constituting the receiving element array 14. Thereby, for example, the same number of reception signals as the reception elements 14a are obtained.

These received signal groups are amplified and discretized in the received signal amplifying / converting section 18, and the processed received signal groups are sent to an image processing means 20 for electronically imaging them. In the present embodiment, the imaging processing means 2
0 is composed of a plurality of image processing units 22. Each imaging processing unit 22 has different convergence / scanning conditions (azimuth gain characteristics), and specifically has different imaging filters. Therefore, a plurality of different imaging results are obtained from the plurality of imaging processing units 22. The plurality of imaging results are shaped as a final imaging waveform by the next integration processing unit 24, and a three-dimensional image is formed by the imaging unit 26 that converts this into information in the distance direction.
The video display device 28 displays the three-dimensional image.

Incidentally, even if the arrangement of the transmitting element array and the receiving element array is devised, a false image can be suppressed to some extent. Regarding the method of this ingenuity, see the above-mentioned Japanese Patent Application
No. 1712.

Next, the principles of the image forming process and the combining process according to the present invention will be described.

The coupling process is generally realized by multiplying the output of each receiving element by a filter coefficient (filter characteristic) and calculating the sum. Here, the filter coefficient is set in consideration of the phase adjustment according to the position of each element and the azimuth to be imaged and the weight for each element, and this filter coefficient can be designed in advance. By preparing two types of filter coefficients or switching their values, it is possible to form different convergence / scan states for the same observation signal.

In general, an arbitrary imaging result includes not only a reflected sound from a target direction but also a signal from a direction corresponding to a side lobe component (false image). In order to correctly obtain an image of the target reflector, it is necessary to sufficiently suppress the side lobe components in all directions. However, when the large-diameter elements are sparsely arranged in the ultrasonic transducer 10, it is not easy to suppress the side lobe components. Therefore, even if it is difficult to suppress side lobes in individual imaging processes, a plurality of imaging filters with reduced side lobes in local orientations are prepared in parallel, and If the direction in which the side lobes are reduced in 22 is changed, the pattern of the false image included in each imaging processing result can be made different, and the false image can be discriminated from their comparison. In the design of the filter, the gain from the target direction (main beam direction) needs to be the same in each image processing unit.

Hereinafter, a specific method of forming a plurality of complementary filters having different side lobe patterns in an azimuth range other than the main beam azimuth will be described. Here, a description will be given using a discrete azimuth model for convenience.

The signal reflected in the discrete azimuth u [u = 1 to U] direction is defined as s _u (t), and its Fourier transform is defined as S _u (f). Here, t represents time, and f represents frequency. Direction number U
Is sufficiently large, an actual physical model can be sufficiently approximated by such discrete orientations.

Focusing on a certain set of transmitting and receiving elements, the received signal obtained by the combination is the element set number m
Directionally reflected signal s _u (t−τ _{m, u} ) with phase delay time τ _{m, u} determined by [m = 1 to M] and reflection direction u
For all directions. When the element output (received signal) describing the o _m (t), the above relationship can be written as follows.

[0046]

(Equation 1) When the result of the Fourier transform and O _m (f) for the element output o _m (t) expanded from the time axis to the frequency axis, of the formula is obtained.

[0047]

(Equation 2) On the other hand, the above relationship for all element sets can be represented by a matrix as shown in the following equation (3). Note that ^T below represents a matrix transpose.

[0048]

(Equation 3) As described above, the imaging process is realized by applying a filter to each element output O _m (f) and adding them to all the elements (phasing addition). This processing can be written as the following equation (7-1). As described later, in order to obtain a plurality of types of imaging processing results (imaging outputs), a plurality of types of filters E may be prepared. In addition,
E _M (f) shown in (8) is a filter characteristic applied to each element.

This equation (7-1) can be transformed into an equation (7-2) if equation (3) is taken into consideration, and specifically, can be rewritten as (7-3). . However, this is based on the relationship as described in (8) and (9) described later. In addition, the following () ⁺ represents a general inverse matrix of Moore-Penrose.

[0050]

(Equation 4) That is, the result of performing the imaging by the filter E is equivalent to the result of the addition to give a gain of G _u (f) for each direction of the reflected signal S _u (f). Of course, the reflected signals S _u (f) in each direction constitute a received signal as a whole,
Although they cannot be taken out individually, the relations of Expression (7-2) and Expression (7-3) are established on the virtual model.

T in the above equations (8) and (9) is a component S of the target direction u ₀ from the reflected signal vector S for each direction.
An operation vector for extracting only _u0 (f), and T =
[Δ ₁ ,..., Δ _{u0 -1} , 1, δ _{u0 + 1} ,..., Δ _U ] (only u _0th element is 1). Here, δ _{1 to} δ _U are set to 0 or a very small value. W in the equation is a weight matrix,

(Equation 5) Can be expressed in the form Here, the diagonal element W _u corresponds to each azimuth, and by changing the value of this weight, the azimuth gain distribution that suppresses a specific azimuth (specifically, G ₁ (f ) To G _U (f)). That is, by increasing the weight coefficient in a specific direction, a local suppression gain can be set for the direction corresponding to that position.

As described above, if G ₁ (f) to G _U (f) in G are obtained under the conditions required for T and W, the relations of equations (8) and (9) are obtained. , E in E ₁ (f) -E
_U (f) can be easily obtained.

Further, by creating another filter in which the positions of the weighting coefficients are interchanged, it is possible to easily design a plurality of complementary filter groups for suppressing different positions. Further, if these plural filters are used in parallel with the calculation software part of the same measuring device, a plurality of the above-mentioned imaging outputs can be obtained at the same time without increasing the hardware at all.

As a specific method for obtaining an appropriate filter coefficient, a weighted least square method, a nonlinear least square method, a neural network, or an arbitrary objective function minimizing method can be used.

Next, a method of designing a plurality of imaging characteristics for the same ultrasonic transducer will be described.

In FIGS. 2 and 3, (a) is a weighting factor distribution used in the design (diagonal element W in equation (10)).
_u ), and (b) shows virtual orientation characteristics (distribution of G _U (f) in equation (9)) obtained thereby.

The example shown in FIG.
It is designed to suppress 0 ° and the vicinity of 30 ° to 90 °, while the example of FIG.
It is designed with the aim of suppressing around ° to -10 ° and around 10 ° to 30 °. 2 and 3
As shown in (a) of FIG.
The main direction is given a weighting value of 1000.

If E ₁ (f) to E _U (f) in E actually used are determined from the distribution of the two kinds of diagonal elements W _U (and the above-described vector T), the local Two types of imaging filters in which the side lobes are sufficiently reduced in the azimuth can be designed.

Then, by comparing the outputs of the imaging filters with each other, it is possible to separate and determine the reflected signal from the target and the side lobe component.
This is because the signal from the target azimuth shows the same behavior in all filters, while the signal from other azimuths differs in magnitude depending on the coefficient value of each imaging filter.

In the present embodiment, such a determination process is executed by the integration processing unit in FIG. 1, and a specific example of the integration processing method will be described below.

As means for integrating a plurality of imaging results,
Examples of the method include a method of taking a product between the respective image forming signals and a method of discriminating and suppressing a reflected sound from a target object and a false image based on a variation between image forming outputs.

First, a method of obtaining a product between the image forming signals will be described. In an actual apparatus, the imaging output is given by the above equation (7-1). In order to explain the principle and effect of the false image suppression, the equation (7-1) obtained by modifying the equation (7-1) is used. ).

First, when equation (7-3) is rewritten into a time waveform by inverse Fourier transform, the following equation (11) is obtained. Here, g _u (t) is the result of inverse Fourier transform of _Gu (f), and * represents a convolution operation. _Gu is the maximum value of | G _u (f) | in the entire used frequency band. The approximation in the second row is established when the used signal bandwidth is wide to some extent. However, it is known that the approximation is sufficiently established in the measurement for detecting the reflection of the pulse signal as in the present invention.

[0064]

(Equation 6) Incidentally, G _u is in accordance with approximately the azimuth gain distribution shown earlier. That is, as can be seen from equation (11), the imaging signal in the time domain is expressed in a form of multiplying all the reflected signals s _u (t) by a certain azimuth gain _Gu and summing them.

Based on the above knowledge, the following two imaging outputs a (t) and b (t) are considered. Where G _u ^(a) , G
_u ^(b) corresponds to G _u in equation (11).

[0066]

(Equation 7) Here, when the product is taken between the two signals, the integrated output can be written as follows.

[0067]

(Equation 8) The first term of the equation (14) corresponds to the weighted value of G _k ^{( a)} × G _k ^(b) applied to the amplitude square value of the reflection signal s _k (t) in the k direction. Therefore, in order to obtain a reflected signal in the target azimuth, G _k ^(a) × G _k ^(b) must have a large value only at k corresponding to the azimuth and a small value in other azimuths. Should be designed. Here, G _k ^(a) , G _k
^(b) does not need to have low sidelobes, and if one takes a large value in a certain direction k, the other may be made small. Such complementary setting is as described in the above-described design example (FIGS. 2 and 3).

On the other hand, the second term of the equation (14) means an interference signal between different directions. For example, if there is a reflecting object at the same distance and at the same direction in _{u and v} directions where G _u ^(a) and G _v ^(b) take large values, the component due to their product is mixed into the signal from the target direction. However, the result of integration is deteriorated. However, if _Gu ^(a) and G _v ^(b) are not excessively concentrated and do not become large values, an average suppression effect can be obtained, and this problem can be avoided. Further, when the reflecting objects in the two directions exist in different distance ranges depending on the sensors, the overlapping portion on each waveform is reduced, and the value of the product is reduced. Therefore, also in this case, deterioration is avoided.

As described above, if the integration processing is performed based on a product between two filters having complementary azimuth characteristics, a false image component from an unnecessary azimuth can be suppressed.

Next, a description will be given of an integration method based on the amount of variation between the respective imaging outputs. The amount of fluctuation is given by the following difference output d (t).

[0071]

(Equation 9) Here, consider the property of Expression (15). First, assuming that the target direction is u = u ₀ , the design of the filter gives
Since G _u0 ^(a) and G _u0 ^(b) , the components related to the u ₀ direction in the difference output d (t) are as follows.

[0072]

(Equation 10) On the other hand, in directions other than the intended direction, it is possible to easily satisfy _Gu ^(a) > _Gu ^(b) or _Gu ^(b) > _Gu ^(a) by designing the filter. Therefore, the components related to these directions u (where u is a direction other than u ₀ ) are as follows.

[0073]

[Equation 11] From Equations (16) and (17), it can be seen that each azimuth component in the difference output d (t) produces a value only when a reflected signal exists in the side lobe. Therefore, only the false image component can be determined based on the value of d (t), and an appropriate suppression gain can be applied to the imaging signal based on the determination result. For example, if d (t) exceeds a certain small threshold, d
(t) is determined to be other than 0, and a certain amount of suppression is calculated by using equation (14).
It is simplest to multiply the product signal between the imaging outputs as shown in FIG.

In the above description, the difference output is expressed by the formula (1)
Although defined as 5), difference information between the absolute values of a (t) and b (t) can be used based on the same principle.

As a criterion for judging the amount of fluctuation, a criterion based on a ratio between a plurality of signals as described below, instead of the amount of difference as in equation (15), can be considered.

[0076]

(Equation 12) Here, consider the property of Expression (18). First, assuming that the target direction is u = u ₀ , G is determined by the filter design.
_u0 ^(a) = _Gu0 ^(b) . On the other hand, in directions other than the intended direction, _Gu ^(a) > _Gu ^(b) or _Gu
^(b) > G _u ^(a) . Therefore, the components of these orientations u in r (t) are as follows.

[0077]

(Equation 13) In other words, it is understood that each azimuth component in r (t) outputs 1 for a signal from the purpose, and a value other than 1 occurs when a reflected signal exists in the side lobe. Therefore, only the false image component can be determined based on whether or not r (t) is close to 1, and an appropriate suppression gain for the imaging signal is determined based on the determination result in the same manner as in the case of d (t). Can be applied. This determination method has a property that, unlike the above-described difference method, it does not depend on the magnitude of the reflected signal s _u (t) in each direction. Therefore, a threshold value for determining whether r (t) is close to 1 or an arbitrary determination function can be easily designed. For example, a threshold value is set near 1, and an appropriate suppression gain is applied when r (t) departs from 1 from that value. Further, as the determination function, a function may be prepared in which r (t) is input and the suppression amount is increased as the value is more distant from 1. For example, a function can be used in which the value decreases linearly or exponentially as the distance from the input 1 increases.
If an arbitrary function that satisfies such a condition is f (x), the output finally combined by this false image determination mechanism can be written as follows.

[0078]

[Equation 14] As is apparent from the above description, first, the inverse Fourier transform is performed on the two types of image outputs given by Expression (7-1), and the two types of image outputs after the inverse Fourier transform are compared with each other. If integrated, the side lobe signal component can be specified depending on the difference in the characteristics of the imaging filters.

In the above description, only two imaging outputs are dealt with for the sake of convenience. However, in actuality, even when three or more filters are used, the present integration method can be applied in exactly the same manner.

Next, the specific processing of the imaging processing section 22 shown in FIG. 1 will be described with reference to FIGS. The imaging processing unit 22 and the integration processing unit 24 can be realized by software, for example.

FIG. 4 shows the operation of the imaging processing unit 22 based on the above principle. In S101, first, a plurality of reception signals output from the plurality of reception elements 14a are fetched. These signals are signals on the time axis.
At 102, those signals are Fourier-transformed and converted into reception signals on the frequency axis.

In S103 and S105, two filter processes are performed on the same signal group using two types of imaging filters. That is, E1 · O and E2 · O
Is calculated. Here, a filter coefficient E1 [ _E1-1 (f), _E1-2 (f),..., E
_1-M (f)] and E2 [E _2-1 (f), E _2-2 (f),.
_2-M (f)] is determined in advance based on the above-described design concept and stored in the storage unit. S104 and S106
Then, for the filter processing result, that is, the imaging processing result,
An inverse Fourier transform from the frequency axis to the time axis is performed.
Then, in S107, the changed imaging processing result is output to the integration processing unit 24.

In the above embodiment, the filtering is performed on the frequency axis. However, the filtering can be performed on the time axis. FIG. 5 shows an operation example of the imaging processing unit 22 in that case.

In S201, a signal group is input as in S101. In S202 and S203, filter coefficients E1 [ _E1-1 (t), _E1-2 (t),..., E1 _-M (t)] and E2
[E _2-1 (t), E _2-2 (t),..., E _2-M (t)] are used, and two types of filtering are performed on the received signal group. Specifically, a convolution operation of the received signal group and the filter coefficient is performed. Incidentally, the filter coefficient E1 [E
_1-1 (t), _E1-2 (t), ..., E1 _{- M} (t)] and E2 [E
_2-1 (t), E _2-2 (t),..., E _2-M (t)], as in the above embodiment, so that the gain characteristics in the azimuth range other than the main azimuth direction are complementary. Should be set to In the convolution operation, the phase adjustment of each received signal may be performed at the same time, but a step for phasing addition may be separately provided. The two types of imaging results after the filtering process are described in S204.
Then, the image data is output to the integration processing unit 24, and a process for suppressing a false image is performed using the difference between the two image forming process results.

FIG. 6 shows an image of an iron ball having a diameter of 7 cm existing in the air measured by the ultrasonic measuring device having the double structure arrangement of the vibrating elements shown in FIG. The measurement is performed under the condition that the target is placed 1 m ahead of the ultrasonic transducer. A filter for realizing the characteristics shown in FIGS. 2 and 3 is used as a filter in the imaging processing unit. In addition, as a mechanism for determining a false image from the imaging results, a mechanism in which f (x) in Expression (20) is a higher-order function is used. As a comparison, FIG. 7 shows an imaging result obtained by imaging with a single filter. This is an imaging result obtained by simply phasing and adding observation signals.

A comparison between FIG. 6 and FIG. 7 shows that in the imaging processing using a simple single filter, a plurality of false images are generated around the image corresponding to the true target, whereas the imaging processing of the above embodiment is performed. According to the image processing, it can be seen that only the image corresponding to the true target is imaged.

[Second Embodiment] Next, a second embodiment will be described. The device configuration of this embodiment is the same as that shown in FIG. 1, but the processing content in the integrated processing unit 24 of FIG. 1 is different from that of the first embodiment. However, the basic principle of obtaining a plurality of imaging processing results by a plurality of complementary filters and reducing side lobes by integration processing is the same in the second embodiment as in the first embodiment.

Hereinafter, for convenience of explanation of the principle and configuration of the second embodiment, the image forming processing model and each parameter will be defined again.

FIG. 8 shows an array model in which only one-dimensional directions are extracted for simplification of the explanation regarding the image forming process. In FIG. 8, θ _b (b = 1 to B)
Is a target direction indicating a direction to be imaged, and x _k (K = 1)
To K) are virtual element arrangements formed by a combination of transmission and reception element arrangements, that is, arrangement coordinates of a synthetic aperture plane.

In this model, an imaging signal obtained by imaging processing such as a synthetic aperture is described in the frequency domain as follows.

[0091]

(Equation 15) Here, s (f) is an imaging output signal vector, o (f) is an observation signal vector, H (f) is a phasing matrix, and G is a weighting coefficient matrix relating to elements, and is defined as follows ( Where c is the speed of sound.

[0092]

(Equation 16) Next, the observation signal at each arrangement coordinate is described in the frequency domain as follows.

[0093]

[Equation 17] Here, a _d (f) is a so-called steering vector, and r (f) is a directional reflection signal vector, which is defined as follows.

[0094]

(Equation 18) Here, θ _d (d = 1 to D) is a discrete arrival direction of the reflected wave. The phasing matrix H (f) that scans the target direction is used.
Is a constant matrix uniquely determined by Expression (24).

Now, in such a model, in the multiplication process in the first embodiment, two types of load coefficient vectors g = [g ₁ ,..., G _k ] and h =
A plurality of imaging signals s ^(g) (t) and s ^(h) (t) are generated using [h ₁ ,..., H _k ], and are superimposed by, for example, multiplication in the time domain. is there.

Here, the reflected signal R _d (f) of the arrival direction θ _d
Is assumed to be a pulsed (zero-phased) narrow-band signal having a center frequency f ₀ and a bandwidth Δf, and its gain is P _d and its propagation delay time is τ _d , s ^(g) (t) and s ^(h) (t) is approximated as follows.

[0097]

[Equation 19] Further, assuming that the target object is localized in only one direction (d = d ₁ ), equations (30) and (31)
Can be written as in equations (32) and (33). The behavior when the target object exists in two directions will be described later.

[0098]

(Equation 20) Therefore, the value finally obtained is calculated by the above equation (3)
Given by the product of 2) and the absolute value of equation (33),

(Equation 21) Can be written. The right side of the equation (34) is | ga _d1 (f ₀ ) · h with respect to the amplitude square value of the azimuth signal envelope in the d ₁ azimuth.
a _d1 (f ₀ ) |. Therefore, in designing the filter, the weight may be designed to have a large value only in the front direction and a small value in the other directions. Here, | ga _d1 (f
₀ ) | and | ha _d1 (f ₀ ) | need not have low sidelobes, and if one has a large value, the other may be small. By using such complementarity, the degree of freedom in designing the load coefficient can be improved. As described above, such complementary design can be easily performed. FIG. 9 shows an example of two such complementary filters. g, h
Indicates each filter.

As described above, these processes are performed in two directions d ₁ and d ₂ where | ga _d1 (f ₀ ) | and | ha _d1 (f ₀ ) | In the case where there is a reflecting object at the same time, a component due to the product is mixed into a signal from a target direction, which results in deterioration of the integration result. However, when the reflecting objects in the two directions are present in different distance ranges depending on the sensors, the overlapping portions on the respective waveforms are reduced, and the value of the product is reduced. Therefore, the deterioration in this case is avoided. Also, if | ga _d (f ₀ ) | and | ha _d (f ₀ ) | are not excessively concentrated and do not become large values even when the reflecting objects in the two directions exist at the same distance. If
An average suppression effect can be obtained, and deterioration can be reduced.

As described above, theoretically, a low side lobe output represented by the equation (34) can be obtained. However, in actual operation, the directional characteristic ga is affected by an error in element variation and an error in a circuit. _d1 (f ₀₎ and ha _d1 (f _0), respectively results in amplitude and phase errors.

That is, the multiplication process in the first embodiment is as follows.
By reducing one of | ga _d1 (f ₀ ) | and | ha _d1 (f ₀ ) | and increasing the other, side lobes are obtained as their geometric mean values.
ga _d1 (f ₀ ) | and | ha _d1 (f ₀ ) | have a large effect on the smaller of the gains. Therefore, when such an error occurs, | ga _d1 (f ₀ ) · ha _d1 ( f ₀ ) | cannot be made sufficiently small. As a result, a signal other than the front component, that is, a signal containing many false images is generated, and there is a possibility that the imaging result will be significantly deteriorated.

Therefore, in the second embodiment, in order to avoid the deterioration due to the error, the multiplication processing in the first embodiment is extended to a non-linear side lobe canceller.

FIG. 10 is a block diagram showing the configuration of the second embodiment. First, the configuration will be described. The imaging processing means 50 includes a plurality (for example, 2
(Imaging) processing unit 22. Each imaging processing unit 22
Is composed of a plurality of coefficient multipliers that weight each received signal and an adder that performs addition while performing phase adjustment on the output signals. That is, each image processing unit 22 outputs an image processing result using each filter.

As shown in FIG. 1, the integration processing section 52 integrates the results of the respective image forming processes to reduce false images. Each imaging result is subjected to an absolute value calculation by an absolute value calculator 53, and the sum of the absolute values is calculated by a sum calculator 54.
On the other hand, the difference between these absolute values is calculated by the difference calculator 56. Non-linear operation is performed on the sum and difference by the non-linear operation unit 58. This non-linear operation includes an absolute value operation and an exponentiation of the power of α. The sum and difference signals subjected to the non-linear operation are sent to the difference calculator 60, and the difference between them is calculated. By this difference operation, only the signal of the main lobe is extracted. The output of the difference calculator 60 is sent to a calculator 62 for level adjustment. The calculator 62 is a circuit that multiplies the 2 / α power of the difference calculation result by 1/4. In an actual device, the arithmetic unit 62
Need not necessarily be provided. Further, the calculation of the absolute value can be omitted depending on the case. α may be configured to be variably set.

The processing shown in FIG. 10 can be written as follows. Here, fα (t) is the final output using the nonlinear parameter α (the output of the arithmetic unit 62).

[0106]

(Equation 22) Here, regarding the two absolute value α power parts in the equation (35), the first term corresponds to the Primary signal shown in FIG.
The term corresponds to the reference signal shown in FIG. The reason why the whole is raised to the power of 2 / α is that the final output in the main lobe is used as the squared value of the signal amplitude.

More specifically, the side lobes | ga _d1 (f ₀ ) | and | ha _d1 (f
₀ ) | is formed complementarily, 2 in the equation (35)
The two magnitude squared parts can be written as:
First, the first term (primary signal) is

(Equation 23) Becomes Next, the second term (reference signal) is

(Equation 24) Becomes Looking at equation (36), the first term part (p
The primary signal (rimary signal) approximately includes a side lobe component and a main lobe component. On the other hand, looking at the equation (37), the (35) second term (referenc
It can be seen that the e signal) extracts only the side lobe component in the first term and directs a blind spot to the main lobe component. FIG. 11 shows these relationships. As described above, the processing expressed by the equation (35) approximately includes the output (prim) including the side lobe component and the main lobe component.
ary signal), only its side lobe component (reference
Signal) can be interpreted as being estimated / subtracted. That is, it can be said that this is a side lobe canceling process in the nonlinear region.

The multiplication processing in the first embodiment is represented by the following equation (3)
This is equivalent to the case where α = 2 in 5). That is, under certain conditions, equation (35) in the present embodiment corresponds to the multiplication process in the previous embodiment, but in the second embodiment, by reducing the nonlinear parameter α, The side lobe suppression performance can be further improved without changing the output in the lobe. This is given by the equation (38) between α and the array output fα (t).
Is mathematically established. Where α ₁ ,
α ₂ is an arbitrary positive real number.

[0109]

(Equation 25) This equation (38) indicates that the array output can be further reduced by decreasing the parameter α. This means that even if an error occurs in the directivity characteristic due to an element error or the like, the side lobe can be made equal to or less than the array output of the multiplication processing in the first embodiment. It can be said that it is processing. Incidentally, the signal in the main lobe is designed such that the two directional characteristics ga _d1 (f ₀ ) and ha _d1 (f ₀ ) are equal, so that | s ^(g) (t) | = | s ^{(h )} (t ) |, Which corresponds to the case where the equal sign on the right side in Expression (38) is satisfied. Therefore, the signal of the main lobe is not affected by the magnitude of α.

If α is set to a non-integer value or a value other than 0.5, a non-integer power operation or a power root operation other than the square root is required, which is not preferable in numerical calculations. However, in order to obtain the side lobe suppression performance higher than the multiplication array processing, α needs to be 2 or less.

Considering the relationship between the two, in the following, α
Consider the case of = 1 as an example. The specific calculation is as follows.

[0112]

(Equation 26) Here, the variables x and y are set to arbitrary values of 0 or more, and min
If [x, y] is defined as a function that takes the smaller of x and y,

[Equation 27] Is established. By substituting this relationship into equation (39), the following equation is obtained.

[0113]

[Equation 28] This calculation means finding the square value of a plurality of array output signals and selecting the smaller one of the magnitude relations, and has the advantage that very simple signal processing can be realized.

FIG. 13 shows an example of the configuration of the integration processing unit 52 for realizing this. 2 from the imaging unit
The two signals are squared by a square calculator 64, and they are m
It is input to the in-operation unit 66. The min calculator 66 is a circuit that outputs the smaller of the input signals. Note that an absolute value calculator may be provided instead of the square calculator 64, and in this case, there is an advantage that the signal level does not need to be adjusted by performing the square calculation after the min calculation.

The side lobe suppression processing of this embodiment (FIG. 1)
0) can be roughly divided into two structures. That is, a "complementary directivity forming process" immediately after the array signal and a "nonlinear sidelobe canceling process" in which only the side lobe component is subtracted in the non-linear region from the side lobe and the main lobe. Here, if only the side lobe cancel operation is performed in the non-linear region, it can be realized by simply making the conventional side lobe canceller (side lobe canceller in the linear region) nonlinear. However, the above-mentioned "complementary directional pattern formation processing" and "non-linear side lobe cancellation processing"
Is more stable against errors in directional characteristics due to element errors and the like.

In the following, this property will be described using a comparative example.

The specific configuration of the nonlinear sidelobe canceller alone can be shown as in FIG. 12 (comparative example). For convenience, the nonlinear parameter α is set to 1. In FIG. 12, the output of each imaging processing unit 22A is input to two non-linear calculators 58, and the difference between the outputs is calculated by a difference calculator 60. The calculator 62A is a circuit that calculates the 2 / α power of the output of the difference calculator 60 for level adjustment.

Here, the load coefficient vector w ⁽¹⁾ = [w _{1
^{(1), ···, w k}} (1)] and _{^{w (2) = [w 1}} (2), ···,
w _k ⁽²⁾ ], these weighting factor vectors are not complementary, w ^{(1 )} forms the main lobe (including side lobes ⁾ for the target orientation, and w ⁽²⁾ is w ⁽¹⁾ It is assumed that only the side lobe is estimated and the main lobe is designed to have no gain. Specifically, it can be written as the following equation. Here, it is assumed that the gain in the main lobe of w ⁽¹⁾ is normalized.

[0119]

(Equation 29) An error caused by variation between elements mainly affects amplitude characteristics and phase characteristics of directional characteristics. Focusing on this point, comparing this comparative example with the second embodiment, the following differences can be seen.

First, consider a case where an amplitude error exists in the directivity characteristics. In the case of only the nonlinear side lobe canceller in FIG. 12, two directional characteristics | w ⁽¹⁾ a _d1 (f ₀ ) | and | w
⁽²⁾ There is a possibility that an amplitude error occurs independently at a _d1 (f ₀ ) |. Here, assuming the error distribution and uncorrelated distribution with mean 0, if the dispersion and sigma ^2, error variance in the final array output can be estimated to be 2 [sigma] ^2.

On the other hand, when the directional pattern forming method of the second embodiment is used in the preceding stage, the output of the directional pattern forming method is equal to that of the smaller output (see equation (41)), and the error in the final array output is obtained. The variance remains at σ ² . this is,
It is smaller than the case of FIG.

Next, consider a case where a phase error exists on the directional characteristics. In the case of using only the nonlinear side lobe canceller, there is a possibility that a phase error occurs independently in the two directional characteristics. Here, when the phase difference between them is shifted by about π / 2, the subtraction is not established, and this greatly affects the side lobe suppression amount.

On the other hand, when the directional pattern forming method of the second embodiment is used, pr is obtained from the relationship of equations (36) and (37).
The same phase error component generated in the imary signal is approximately r
This will also occur in the eference signal. Therefore, by subtracting them, the phase error component is reduced,
The side lobe does not rise.

As described above, when used in combination with the complementary directional pattern forming method, robustness against both the amplitude and phase errors in the directional pattern caused by the variation in the elements is more effective than when only the nonlinear side lobe canceller is used. It can be seen that it is.

As described above, if non-linear side lobe canceller processing is performed on the outputs of two weight coefficient vectors having complementary directional characteristics, it is possible to suppress false image components from unnecessary directions. In addition, it is possible to configure an imaging device that is not easily affected by errors of the elements. In the above description, only two imaging outputs are dealt with for the sake of convenience. However, the present integration method can be applied in exactly the same way when actually using three or more weight coefficient vector groups.

FIG. 14 is a photograph of a human model consisting of only a real-sized upper body actually existing in the air by the ultrasonic measuring apparatus having the double structure arrangement shown in FIG. The target is located about 2 m from the front of the array. As the directivity of the load coefficient vector in the image forming processing unit, the one shown in FIG. 9 was used. As a mechanism for determining a false image from the results, a nonlinear sidelobe canceller (Equation (3)
5)), the parameter α was set to 0.5, 0.7 and 1. For comparison, an imaging result in the case of performing integration by multiplication processing is also shown. This corresponds to the case where the parameter α is set to 2 in the nonlinear sidelobe canceller (Equation (35)).

The simple multiplication (when α = 2: see the lower right of FIG. 14) and the method of the second embodiment (α = 0.5, 0.
7, 1: upper left, upper right and lower left, see FIG. 14). In the image forming process by the simple multiplication, a plurality of false images are generated around the image corresponding to the true target. On the other hand, in the ultrasonic measurement device having the imaging method of the present application, it can be seen that only the image corresponding to the true target is formed.

[0128]

As described above, according to the ultrasonic measuring apparatus of the present invention, it is possible to virtually obtain the same effect on a computer as if a plurality of transmitting / receiving arrays having different apertures were configured. Therefore, it is not necessary to actually increase the number of elements of the transmission / reception array and the scale of the apparatus in order to improve the performance. Further, since a false image can be determined from a plurality of imaging processing results and only that part can be suppressed, even in a transmission / reception array in which a small number of elements in which a false image is likely to appear are sparsely arranged, a relatively false image can be obtained. It is possible to obtain a three-dimensional image with little accuracy. As a result, without using a large amount of special elements,
Since it is possible to configure an ultrasonic measurement device using a relatively large-diameter element that is usually used, it is possible to prevent a decrease in transmission output and reception sensitivity and reduce the device scale while maintaining high azimuth resolution. it can. Further, according to the present invention, it is possible to configure a device which is hardly affected by variations in elements and errors in circuits.

[Brief description of the drawings]

FIG. 1 is a diagram showing an embodiment of an ultrasonic measuring device according to the present invention.

FIG. 2 is a diagram showing characteristics for designing an imaging filter.

FIG. 3 is a diagram showing characteristics for designing an imaging filter.

FIG. 4 is a diagram illustrating a first embodiment of an imaging processing unit.

FIG. 5 is a diagram illustrating a second embodiment of the image forming processing unit.

FIG. 6 is a diagram showing a measurement result of the device of the present embodiment.

FIG. 7 is a diagram showing measurement results of a conventional device.

FIG. 8 is a diagram showing a one-dimensional model in an image forming process.

FIG. 9 is a diagram showing two directional characteristics (filter characteristics) set mutually in the image processing unit.

FIG. 10 is a block diagram illustrating a configuration of an imaging processing unit and an integration process according to a second embodiment.

11 is a diagram showing a primary signal and a reference signal when the two filters shown in FIG. 9 are used.

FIG. 12 is a diagram showing a configuration of a comparative example in which only a non-linear side lobe canceller structure is introduced.

FIG. 13 is a diagram illustrating a configuration example of an integration processing unit using a min operation.

FIG. 14 is a diagram showing an imaging result of a human model.

[Explanation of symbols]

Reference Signs List 10 ultrasonic transducer, 12 transmitting element array, 14
Receiving element array, 16 transmitting element driving section, 18 received signal amplifying / converting section, 20 imaging processing means, 22 imaging processing section,
24 integrated processing unit, 26 imaging unit.

Claims (12)

[Claims] An ultrasonic transducer for transmitting ultrasonic waves from one or a plurality of transmitting elements and receiving reflected waves by a plurality of receiving elements to output a received signal group corresponding to each of the receiving elements. An image processing means for obtaining a plurality of image processing results while applying a plurality of filters having different filter characteristics to the received signal group; and An ultrasonic measurement device, comprising: processing means. 2. The apparatus according to claim 1, wherein each of the filter characteristics is set so that the azimuth gain patterns are different from each other in an azimuth range other than the main azimuth, and the false image generation patterns are mutually different in each of the imaging processing results. An ultrasonic measuring device characterized by being different. 3. The apparatus according to claim 1, wherein the imaging processing unit includes: a converting unit configured to convert each of the received signals from a time axis to a frequency axis; On the other hand, an ultrasonic measurement apparatus comprising: a filter processing unit for applying a plurality of filters; and an inverse conversion unit for converting each imaging processing result after the filtering from a frequency axis to a time axis. 4. The ultrasonic measuring apparatus according to claim 1, wherein said imaging processing means applies a plurality of filters on said received signal on a time axis. 5. The ultrasonic measurement apparatus according to claim 1, wherein the integration processing unit suppresses a false image by multiplying the plurality of imaging processing results by each other. apparatus. 6. The ultrasonic measurement apparatus according to claim 1, wherein the integration processing unit suppresses a false image based on a difference between the plurality of imaging processing results. apparatus. 7. The ultrasonic measurement apparatus according to claim 1, wherein the integration processing unit suppresses a false image based on a ratio of the plurality of imaging processing results. apparatus. 8. An ultrasonic transducer that transmits ultrasonic waves from one or a plurality of transmission elements and receives reflected waves by a plurality of reception elements, thereby outputting a reception signal group corresponding to each reception element. Image processing means for applying a plurality of filters having azimuth gain patterns complementary to each other in an azimuth range other than the main azimuth to the received signal group to obtain a plurality of imaging processing results; and An integrated processing means for performing an integrated process for suppressing a false image on an image processing result. 9. The apparatus according to claim 8, wherein the integration processing means includes: a sum and difference calculation means for calculating a sum and a difference of the plurality of imaging processing results; and a difference for calculating a difference between the sum and the difference. An ultrasonic measuring device, comprising: a calculating means. 10. The apparatus according to claim 8, wherein the integration processing means includes: a sum and difference calculation means for calculating a sum and a difference of the plurality of imaging processing results; An ultrasonic measurement apparatus comprising: (a) an exponent calculating means for raising the power; and a difference calculating means for calculating a difference between the sum raised and the difference raised to the power of the α power. 11. The apparatus according to claim 8, wherein the integration processing means calculates absolute values of the plurality of imaging processing results, and calculates absolute values of the plurality of imaging processing results. An ultrasonic measurement apparatus, comprising: a minimum value selecting unit that suppresses a false image by successively comparing each measurement point and selecting a minimum one. 12. The ultrasonic measurement apparatus according to claim 1, further comprising: an imaging unit configured to form a three-dimensional image of the observed object based on a result of the integration processing. .