JPH08322836A - Ultrasonic diagnostic device - Google Patents

Abstract

PURPOSE: To provide highly accurate phase matching by multiplying the window function of a valid time length to reception signals, obtaining an autocorrelation coefficient, optimizing the valid time length of the window function based on the arithmetic result of an evaluation function to an autocorrelation coefficient value, multiplying the optimized window function and the reception signals and performing a mutual correlation arithmetic operation. CONSTITUTION: In the case of computing an autocorrelation function, in data transfer among a central arithmetic processing circuit 71, a storage circuit 75 and an address decoder 72 and an arithmetic operation inside the central arithmetic processing circuit 71, for the Fourier transformed result of the same reception signals, complex conjugate data are outputted through a data bus DB to a selector 197 and the selector 197 successively stores input in a shift register 194. In the case of computing a mutual correlation function, for the Fourier transformed result of the different reception signals, the complex conjugate data are outputted to the selector 197. The central arithmetic processing circuit 71 corrects values stored in the internal register of a reception delay circuit and a transmission circuit corresponding to the delay time calculation program of a program storage circuit 74.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a medical ultrasonic diagnostic apparatus, and more particularly to improvement of aberration correction of its phasing circuit.

[0002]

2. Description of the Related Art An ultrasonic diagnostic apparatus is widely known, which transmits an ultrasonic wave into a subject and displays a tomographic image within the subject based on a signal received by an echo that is reflected and returned. The device uses a technique called electronic focusing of received signals.

As shown in FIG. 2, assuming a reflection source at a point 20 of a predetermined depth in the subject, a reflected wave 201 is generated by the receiving element 10.
Suppose that a spherical wave arrives at. Assuming that the sound velocity is constant between each position of the arrayed receiving elements and the reflection source, the difference in the propagation time of ultrasonic waves can be calculated from the spatial distance of the reflection source assumed to be each element. A delay amount 21 is given to the signal obtained by each receiving element so as to cancel the time difference of the sound waves arriving from the reflection source, and the reflection signals are aligned in the same phase as 22 and then added by the adder 16. The signals from the reflection source have the same phase of interference and strengthen each other, and the other signals are significantly weakened by the interference of different phases, which causes resolution in the array direction of the receiving elements.

However, since various tissues such as fat layers and muscles exist in the human body, an error occurs in the determination of the propagation time difference based on the uniform sound velocity assumed in the electronic focus. For example, the sound velocity of the fat layer is about 1470
m / s, while muscle and liver are 1540 m / s
It is before and after, and unless the error of the propagation time due to the individual difference of the human body such as the thickness of the fat layer is not corrected, sufficient convergence of the electronic focus cannot be achieved and the tomographic image becomes unclear.

As a method for solving this problem, a method using a correlation operation is disclosed in US Pat. No. 4,876,614. This method gives a delay time corresponding to a predetermined reception focus to the reception signal of each receiving element on the assumption that the sound velocity in the subject is uniform, and a part of each reception signal after giving each delay amount. Is extracted as a region for correlation calculation, and the time difference to be corrected is estimated from the result of cross-correlation calculation between adjacent elements. In this method, the portion corresponding to the same reflection source in each received signal has a reflection waveform shape with a very high correlation, and if the time difference with the highest correlation is taken as the correction amount, a more focused tomographic image is obtained. Is based on the assumption that

[0006]

However, the ultrasonic diagnostic apparatus configured as described above has the following problems.

That is, when the cause of the disturbance of the propagation time of ultrasonic waves is considered to be diffraction at the boundary 30 between layers having different sound velocities as shown in FIG. 3, a significant waveform deformation may occur depending on the boundary shape. As a typical example, there is a part that forms an acoustic lens based on the difference in sound velocity in a part of the boundary surface, and it can be considered that the focal length of the lens is sufficiently large with respect to the distance between the boundary surface and the receiving element. The case where the wave when it disappears affects the received signal can be mentioned. In such a situation, the wavefront propagating from the reflection source does not simply propagate with temporal distortion, but has a similar waveform 33 with different amplitudes at another temporal position. Propagate with distortion.

In such a case, in addition to the waveform group 320 including time error, an additional waveform group 330 caused by the wave 33 is generated.
Are included in the time domain of the waveform for which the correlation calculation is performed. Generally, since the time difference between the waveform group 320 and the waveform group 330 is usually different for each received signal, when the cross-correlation calculation is performed with the time length including the waveform group 330,
Due to the correlation of the additional waveform portion of the waveform group 330, an error occurs in the time difference calculation of the main waveform group 320.

As a result, an appropriate time difference can be provided due to a significant waveform distortion only by the waveform cutting out of each received signal by the time gate processing which has been performed for a predetermined fixed time length to detect the time error and the correlation calculation processing. I couldn't.

The present invention has been made in view of the above circumstances, and an object thereof is to provide an ultrasonic diagnostic apparatus capable of obtaining phase adjustment with extremely high accuracy.

[0011]

Of the inventions disclosed in the present application, a representative one will be briefly described below.
It is as follows.

That is, in the ultrasonic diagnostic apparatus having means for correcting the delay time given to each of the transmitted and received signals based on the time difference obtained by the cross-correlation calculation between the received signals, the ultrasonic diagnostic apparatus is effective for the received signals. Means for obtaining an autocorrelation coefficient by multiplying the window function of the time length, means for optimizing the effective time length of the window function based on the calculation result of a predetermined evaluation function for this autocorrelation coefficient value, and each received signal A means for performing the cross-correlation operation using the window function optimized for each and the result of multiplication of the received signal is provided.

[0013]

In the ultrasonic diagnostic apparatus configured as described above, first, after the time delay processing performed under the assumption that the sound velocity is constant, the reception waveform corresponding to each reception element obtained by cutting the reception waveform by the time gate is used. An autocorrelation function can be calculated to detect similar waveform portions in the received signal caused by the diffraction of the sound field.

In this case, since the autocorrelation function does not have the maximum for each carrier wave period by performing the quadrature detection processing by the carrier wave frequency, the similar portion can be easily detected. Then, if the result of the autocorrelation calculation shows a value equal to or greater than a predetermined threshold value at a time position away from the position where the time difference is 0, it is determined that a significant similar waveform may overlap the time difference portion. , The effective time length of the window function by which the received waveform before the correlation calculation is multiplied is shortened by a predetermined width.

By optimizing the time length of the window function performed in this way, the waveform for which the cross-correlation calculation is performed is determined for each received signal, and the time difference to be corrected is obtained from the correlation calculation values. That is, the time difference to be corrected is appropriately obtained by the cross-correlation calculation between the reception signals with the window function weights.

Therefore, the influence of the similar waveform portion (false image) caused by the diffraction of the sound field on the correlation calculation value can be eliminated when the cross-correlation calculation is performed by such processing.
Extremely accurate phasing can be obtained.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of an ultrasonic diagnostic apparatus according to the present invention will be described below with reference to the drawings.

First, referring to FIG. 1, an electroacoustic transducer group 1
Each element of 0 is connected to each of the transmission / reception separation circuits 11. The transmission / reception separation circuit 11 connects the output of the wave transmission circuit 12 to the electroacoustic transducer group 10 and connects the output of the electroacoustic transducer group 10 to the reception delay circuit 13. The wave transmission circuit 12 is controlled by a command signal CNT2 from the control circuit 70. Reception delay circuit 1
The output of 3 becomes the input of the selector group 14. Reception delay circuit 1
3 is controlled by a command signal CNT1 from the control circuit 70.

The selector group 14 receives the reception signal delayed by the reception delay circuit 13 as an input of the adder 16. This selection is determined by the command signal CNT3 from the control circuit 70.
The adder 16 outputs a phasing addition signal to the display circuit 18, and the display circuit 18 sequentially converts different raster (vector) data into a luminance signal and displays a tomographic image such as a B mode image.

An output group Si composed of respective outputs from the selectors constituting the selector group 14 becomes an input of the quadrature detection circuit 15. Complex signal outputs I and Q of the quadrature detection circuit 15
Becomes an input of the time gate circuit 17. The time gate circuit 17 is controlled by the command signal GTCTL from the control circuit 70. The output of the time gate circuit 17 becomes the input of the correlation calculation circuit 19. The control circuit 70 controls 12, 13, 14, and 17. The correlation calculation circuit 19 uses the signal line group CORCT.
L performs bidirectional data exchange and control with the control circuit 70.

Next, the quadrature detection circuit 15 and the time gate circuit 1
7, the respective configurations of the correlation calculation circuit 19 and the control circuit 70 will be described in more detail. First, the quadrature detection circuit 1
5, the output Si from the selector group 14 shown in FIG.
Is an input of the selector 150. The selection is determined by the control signal CH.

The output of the selector 150 is the frequency mixer 15
3 and 154 are input. The local oscillators 151 and 152 output sine wave signals whose phases are different from each other by 90 degrees, and are input to the frequency mixers 153 and 154. Frequency mixer 15
The outputs of 3, 154 are low-frequency pass filters 155, 156.
Will be input.

The outputs I and Q of the low frequency pass filters 155 and 156 are input to the time gate circuit 17. In the time gate circuit 17, the outputs I and Q become inputs to the analog-digital converters 171 and 172. The outputs of the analog-digital converters 171 and 172 are the storage circuits 191 and 192 that form the correlation calculation circuit 19.
Are connected to the data buses 191D and 192D attached to the. The sampling operation of the analog-digital converters 171 and 172 is controlled by the counter 173. Counter 173
Is controlled by the command signal GTCTL to start and end counting.

The data buses 191D and 192D are storage circuits 191, 192, an analog-digital converter 171, and
172 and a selector 193 are commonly connected. The memory circuits 191 and 192 have write permission commands WEN1 and WE.
It is controlled by N2, read permission commands REN1, REN2, and address command ADR. The selector 193 outputs the command signal Sl
Accordingly, the selection output result of the data buses 191D and 192D is input to the selector 197. The selector 197 is the selector 19
3 is input and the data bus DB is input, and one of them is input to the shift register 194 by the command signal S2. The shift register 194 inputs a time-series signal from the selector 193 by the shift pulse signal SP1, and outputs the contents of the register in parallel to the fast Fourier calculation circuit 195 according to the output permission command OEN.

The fast Fourier calculation circuit 195 outputs the result of the Fourier transform of the contents of the shift register 194 to the shift register 196. The shift register 196 receives the output of the fast Fourier calculation circuit 195 in parallel according to the input permission command IEN, converts the parallel signal into serial by the shift pulse signal SP2, and outputs the serial signal to the selector 73.
The selector 73 connects either the output of the shift register 196 or the input / output of the data bus DB to the data input / output of the central processing unit 71 by the command signal S3. Data bus DB
Is connected to the data output of the program storage circuit 74.

The program memory circuit 74 uses the address command ADR.
And the read permission command REN4. The data output of the memory circuit 75 is also connected to the data bus DB. The memory circuit 75 is controlled by an address command ADR, a read permission command REN3, and a write permission command WEN3. The output of the address data MADR of the central processing unit 71 is output to the address analysis circuit 72. The write-read command RW is also output to the address analysis circuit 72.

The address analysis circuit 72 outputs control data S1 to 3, REN1 to 4, WEN1 to 3, CH, according to the address data MADR output from the central processing unit 71, the write-read command RW, and the output of the data bus DB. AD
Outputs R, SP1, 2, IEN, OEN. further,
The address analysis circuit 72 receives the control signal DRADR to the wave transmission circuit 12 and the control signal CHAD to the reception delay circuit 13 in FIG.
The control signal CSEL is output to the R and selector group 14. C
HADR, DRADR, and CSEL form command signals CNT1, CNT2, and CNT3, respectively, together with the data bus DB.

Next, the operation of the ultrasonic diagnostic apparatus configured as described above will be described below.

First, the functions of the respective parts in the case of operations other than the correction processing will be described with reference to FIGS. 1 and 7.

First, the delay time group calculated so as to focus at a predetermined position in the subject is set in the wave transmission circuit 12 by the control circuit 70 in FIG. 1 according to the command signal CNT2. The delay time group should be obtained in advance by calculating the propagation time using the geometrical distance between each element of the electroacoustic transducer group 10 and the transmission focus, for example, assuming that the sound velocity in the subject is constant. You can

These delay time groups have been converted into clock count value data suitable for control by the wave transmission circuit 12, and are stored in the program storage circuit 74 of FIG. Before the wave transmission operation is started, the central processing unit 71 sets the data storage address of the program memory circuit 74 and the address corresponding to the register (not shown) in the wave transmission circuit 12 as the address data MADR to the address decoder 72 before the wave transmission operation is started. It is realized by outputting. At this time, the address decoder 72 controls the control commands DRADR, S3, REN4, ADR according to the address data MADR.
Is output. DRADR designates which register in the wave transmission circuit 12, and S3 indicates that the selector 73 is the data bus D.
Specify to select B. DRADR and the data bus DB are command signals CNT2 for the wave transmission circuit 12 of FIG.

Inside the wave transmission circuit 12, the wave transmission start trigger pulse of the excitation pulse signal for each electroacoustic transducer is controlled by a clock counting circuit (not shown), and the clock count value is set in the register set by the control circuit 70. When it becomes the same as the stored value, a transmission start trigger pulse is output, and the pulse becomes an input of a transmission amplifier (not shown). The excitation pulse signal output of this transmission amplifier outputs to the electroacoustic transducer a voltage that generates a sound pressure sufficient to obtain a reflection echo from the inside of the subject. Made via.

The transmission / reception separation circuit 11 is a circuit (not shown) utilizing the non-linearity of the diode, and is composed of a circuit for shunting so that a high voltage signal current does not directly flow into the wave rectification circuit 13. The excitation pulse signal of sufficient voltage that has passed through the transmission / reception separation circuit 11 drives the electroacoustic transducer. In this way, the transmission circuit 12 of FIG. 1 outputs an excitation pulse to the electroacoustic transducer group 10 according to the set delay time, and this pulse is converted by the electroacoustic transducer group 10 into a wave in the subject.

The reflected pulse reflected from the subject is converted again into an electric signal by the electroacoustic transducer group 10, and becomes an input of the reception delay circuit 13 via the transmission / reception separation circuit 11.
The reception delay circuit 13 aligns the phases of the reception signal group with the delay time set according to the command signal CNT2 of the control circuit 70. The reception delay circuit 13 amplifies the input reception signal by an amplifier (not shown) provided inside. The output of the amplifier is delayed by a tapped delay element (not shown) provided inside. The delay element gives a different delay to the received signal depending on the tap position for extracting the signal, and a selection circuit (not shown) for selecting one from the tap outputs is connected to the output terminal group.

The selection circuit includes a register, and the content of the register is set according to the command signal CNT1 of the control circuit 70. The delay time group of the reception delay circuit 13 is
It has been converted into tap selection data suitable for control.
It is stored in the program memory circuit 74. This data is transferred by the central processing unit 7 before the start of the wave transmission operation.
1 indicates the data storage address of the program storage circuit 74 and the address corresponding to a register (not shown) in the reception delay circuit 13 as the address data M for the address decoder 72.
It is realized by outputting as ADR. At this time, the address decoder 72 outputs control commands CHADR, S3, REN4, ADR according to the address data MADR. CHADR designates which register in the reception delay circuit 13, and S3 indicates that the selector 73 is the data bus D.
Specify to select B. CHADR and the data bus DB are command signals CNT for the reception delay circuit 13 of FIG.
It is 1.

The respective reception signals phased by the reception delay circuit 13 become the respective inputs of the selector group 14. The selector group 14 is controlled by the command signal CNT3 of the control circuit 70. Selector group 14
Includes a register, and according to the contents of the register, each reception input is connected to the adder 16 in a state where the reception signal collection for correction is not performed. The setting of this built-in register is performed by the central processing unit 71 by setting the address corresponding to the register as the address data MADR to the address decoder 72 and setting the set value in the data bus DB before the transmission operation is started.
It is realized by outputting to. At this time, the address decoder 72 outputs the control commands CSEL, S3 according to the address data MADR. CSEL designates a register in the selector group 14 and selects each gate by the data bus D.
The data on B specifies. S3 specifies that the selector 73 selects the data bus DB. CSEL and data bus DB in FIG. 7 are command signals C for the selector group 14 in FIG.
It is NT3. The inputs of the adder 16 are added in parallel and become the input of the display circuit 18.

The display circuit 18 includes a detection circuit, a logarithmic compression circuit, a coordinate conversion circuit, a display device and the like which are not shown,
A tomographic image is displayed by converting the phasing addition result into a luminance signal of the display device and sequentially scanning the bright lines on the display device.

Next, the correction operation to which the present invention is applied will be described.

First, one of the output destinations of the selector group 14 is used as an input of the quadrature detection circuit 15 as a signal group Si. Ultrasonic transmission / reception is performed by the same operation as that described above, and the output of the reception delay circuit 13 is simultaneously input by the selector group 14 to the selectors 150 in the quadrature detection circuit 15 of FIG. 7. The selector 150 selects a target reception signal from the signal group Si by the control signal CH. The control signal CH is the output of the address decoder 72, and the address data M of the central processing unit 71.
It is designated by ADR and data on the data bus DB.

The output of the selector 150 is branched and becomes the input of the frequency mixers 153 and 154. Local oscillator 151,
152 outputs a sine wave reference signal whose frequency is almost equal to that of the received signal carrier and whose phase differs by 90 degrees. The frequency mixers 153 and 154 perform frequency mixing of the reception signal and the reference signal, and the reception signal becomes a sum signal of the baseband signal and the harmonic signal. The outputs of the frequency mixers 153 and 154 are filtered by the low-frequency pass filters 155 and 156 so that only the baseband signal remains and form the quadrature detection outputs I and Q of the quadrature detection circuit 15. The quadrature detection outputs I and Q are analog-digital converters 171 in the time gate circuit 17,
It becomes the input of 172.

Analog-digital converters 171, 17
Reference numeral 2 converts each of the quadrature detection signals into a digital signal and outputs the digital signal to the data buses 191D and 192D in the correlation calculation circuit 19. The sampling operation of the analog-digital converters 171 and 172 is controlled by the counter 173. Sampling is performed in synchronization with a clock signal (not shown), and permission of sampling and output operation to the data buses 191D and 192D is determined by the output of the counter 173. The counter 173 is controlled to start counting by the command signal GTCTL. Simultaneously with the start of counting of the clock signal according to the command signal GTCTL, sampling of the analog-digital converters 171 and 172 and output operation to the data buses 191D and 192D are permitted, and sampling is performed when the count value reaches a predetermined value. And the output to the data buses 191D and 192D is completed. The data sampled in synchronization with these operations are stored in the storage circuit 191,
Sequentially stored in 192.

The storage circuits 191 and 192 are controlled by the write permission commands WEN1 and WEN2 and the address command ADR. Further, the selector 192 does not connect the inputs of the data buses 191D and 192D to any part by the command signal S1 from the address decoder 72. The operation of capturing the received signal by the time gate is performed while keeping the setting conditions of the transmission circuit 12 and the reception delay circuit 13 of FIG. 1 under the same delay time condition, and the selector 15 is repeated each time the transmission and reception of ultrasonic waves are repeated.
0 changes the received signal selected. Also, the phase of the reference wave output of the local oscillator is synchronized with the transmission start time and is constant under the same delay time condition. Every time the reception signal to be selected is changed, the address command ADR of the storage circuits 191 and 192
The value of is changed and the received signal value is sequentially stored in different storage spaces.

After the time gate signals obtained from the received signals necessary for use in the correlation calculation are stored in the storage circuits 191, 192, the operation for calculating the cross-correlation or autocorrelation function in the frequency domain is started. The selector 193 receives the selection output result of the data buses 191D and 192D by the command signal S1 as an input of the selector 197. The selector 197 outputs the output of the selector 193 to the shift register 19 according to the command signal S2.
Input 4 The shift register 194 inputs a time-series signal from the selector 193 by the shift pulse signal SP1, and outputs the contents of the register in parallel to the fast Fourier calculation circuit 195 according to the output permission command OEN. At this time,
The shift register 194 stores the complex signal values read from both the storage circuits 191 and 192.

The fast Fourier calculation circuit 195 outputs the result of the Fourier transform of the contents of the shift register 194 according to a clock (not shown) to the shift register 196. The shift register 196 takes in the outputs of the fast Fourier calculation circuit 195 in parallel according to the input permission command IEN,
The shift pulse signal SP2 converts parallel signals to serial signals and outputs the serial signals to the selector 73. The selector 73 connects the output of the shift register 196 to the data input / output of the central processing unit 71 by the command signal S3.

The central processing unit 71 is the shift register 1
The address data assigned to the 96 outputs is output as MADR, and the address decoder 72 sets the command signal S3 based on it. The data read into the register in the central processing circuit 71 is sequentially stored in the memory circuit 75 by the cooperative operation of the address decoder 72 and the selector 73. The memory circuit 75 uses the central processing unit 71 according to the control signals WED3 and ADR output from the address decoder 72.
The data output by is input from the data bus and stored.
By storing all the contents of the shift register 196 in the memory circuit 75, the Fourier transform result of one received signal in the memory circuits 191 and 192 cut out by the time gate is stored in the memory circuit 75. . This operation is repeated for all the received signals stored in the storage circuits 191, 192.

When calculating the autocorrelation function, the central processing circuit 71, the memory circuit 75, and the address decoder 72 are used.
Between the data transfer between the two and the calculation in the central processing unit 71, a pair of complex conjugates and the same ones are obtained for the Fourier transform result of the same received signal, and the complex multiplication is performed for each corresponding frequency component. The complex conjugate data of is output to the selector 197 via the data bus DB. Selector 1
97 sequentially stores the inputs in the shift register 194 according to the command signal S2. Here, the fast Fourier calculation circuit 195 performs the Fourier transform similarly to the original Fourier transform of the time gate signal, but the inverse Fourier transform is achieved because the complex conjugate data is input. The obtained result is stored in the memory circuit 75 again.

In the case of the calculation of the cross-correlation function, one pair of the complex conjugate of one and the other one of the Fourier transform results of different reception signals are obtained, and the complex multiplication is performed for each corresponding frequency component. The complex conjugate data of is output to the selector 197 via the data bus DB. Subsequent inverse Fourier transform and storage in the storage circuit 75 are similar to the calculation of the autocorrelation function.

The output values of the autocorrelation calculation of all the received signals stored in the storage circuit 75 are stored in the storage circuit 75 after the absolute value is calculated by the central processing circuit 71 and standardized by each maximum value. The storage result is evaluated by an evaluation function having the threshold value Rt as a parameter.

Here, some embodiments of how to obtain the evaluation function will be described.

First, as shown in FIG. 6, an autocorrelation coefficient (function) curve obtained by the above-described configuration and calculation is set to 60 for a certain received signal cut out by a time gate. The autocorrelation coefficient is equivalent to calculating the overlap integral while sequentially increasing the time difference, and increases or decreases each time there is a similar portion in its own waveform. If the time difference is 0 (horizontal axis origin), it is 1, and at the time difference corresponding to the position that gives the maximum such as between 62 and 63 in the figure, similar parts must exist again in their own waveforms. Means When the threshold value Rt is set in the range of 0 <Rt <1, it is possible to find at least two or more points that cross the threshold level, if there is one or more similar parts in the own waveform. These points are 6 in the figure.
1 to 63. If the threshold value Rt is close to 1, the number of crossings becomes 1 in most cases, and in other cases, there is a possibility of finding many crossing points due to fluctuations in the correlation coefficient.

In this way, the degree of the similar portion in the own waveform can be evaluated by using the crossing number of the threshold value with respect to the autocorrelation coefficient curve as the evaluation function. In this evaluation calculation, the difference between the autocorrelation coefficient of each received signal stored in the storage circuit 75 and the threshold value Rt is calculated in order from the coefficient value with a time difference of 0, and is counted every time the sign of the calculation result is inverted. It can be realized by doing. In this evaluation function, the evaluation result is rejected when the number of crossings is, for example, 2 or more.

Next, an evaluation function as another embodiment first takes the sum of the section lengths exceeding a predetermined threshold value Rt. Then, the section lengths s1 and s2 shown in FIG. 6 are respectively added. The fact that the section length exceeding the threshold value Rt is large indicates that a large number of similar portions appear in the waveform of itself or there is a portion where the change in the similarity is very small with respect to the time difference change. This evaluation operation is performed by the memory circuit 7
The autocorrelation coefficient of each received signal stored in
This can be realized by calculating the difference from the threshold value Rt in order from the coefficient value of, and counting the sections where the sign is positive. In this evaluation function, the evaluation result is rejected when the number of crossings is, for example, 1/4 or more of the correlation coefficient calculation length D in FIG.

Furthermore, the evaluation function as another embodiment is
First, the sum of squared sums of correlation coefficients in a predetermined time difference section is compared with a predetermined threshold value. This evaluation operation is performed by the memory circuit 7
When calculating the absolute value of the autocorrelation coefficient of each received signal stored in 5, the sum can be calculated without performing the square root operation and the magnitude can be compared with the threshold value. In this evaluation function, the evaluation result is rejected when the total sum exceeds the threshold value.

In the ultrasonic diagnostic apparatus shown in this embodiment, all of the above three evaluation functions are prepared, and the operator selects the threshold value like Rt by an input device (not shown).
Or you can enter it.

The above evaluation functions are stored in the storage circuits 191, 192.
It is applied to all or some of the received signals held therein, and the number of rejected received signals is counted. When the number of rejects reaches a predetermined number, weighting is performed by the window function so as to shorten the time length of the waveform of the signal value before the correlation calculation.

Here, the window function is simply a rectangular window whose value is 0 for a predetermined time width from both ends of the waveform or Hanning.
A window or the like can be used.

Next, the optimization of the time length when the rectangular window is used as the window function and the number of curves crossing the threshold value Rt is selected as the evaluation function will be described in detail with reference to FIGS. 4 and 7. . First, the time gate circuit 17 of FIG.
The waveforms stored in the storage circuits 191 and 192 are designated as 410 and 411 in FIG. The waveform stored here is that after quadrature detection, but the waveform in the same time window portion before quadrature detection is shown as 40. The waveform 410 is the output I of the quadrature detection circuit 15 cut out by the time gate circuit 17, and is stored in the storage circuit 191. The waveform 411 is obtained from the output Q of the quadrature detection circuit 15 and stored in the storage circuit 192. In both cases, the time length is T1.

The result of calculating the autocorrelation coefficient (function) of this signal by the operation of the correlation calculation circuit 19 and the control circuit 70 is the curve 4
Twelve. The data of this curve is stored in the memory circuit 75 by the method using the above-mentioned fast Fourier transform and inverse transform. The central processing unit 71 sequentially determines the difference between each value on the autocorrelation coefficient curve stored as array data in the memory circuit 75 and the threshold value Rt according to the procedure of the program in the program memory circuit 74 from the position of time difference 0. The number of crossings is obtained by calculating and counting the number of inversions when the sign of the result is inverted. In the case of the curve 412, a correlation coefficient greater than or equal to the threshold value Rt is recognized near the time difference t1 except near the time difference 0. In this case, the number of crossings is 3 which is 2 or more, and therefore the evaluation result is rejected.

If even one of the autocorrelation coefficient curves of all the received signals rejects this evaluation, the subsequent operation of shortening the effective length of the time window is repeated. The operation of re-calculating the autocorrelation coefficient curve by shortening the effective length of the time window is performed by the quadrature detection waveforms held in the storage circuits 191 and 192 in FIG. 7 to the shift register 194 via the selectors 193 and 197. It is realized in the process of transferring. The output of the selector 197 can be set to 0 by the command signal S2. The quadrature detection waveforms held in the storage circuits 191 and 192 are synchronized with the shift pulse SP1 and the shift register 1
Transfer to 94. During a period in which a portion corresponding to a predetermined length of the leading portion or the trailing portion of the waveform is being transferred, the output of 0 by the command signal S2 of the selector 197 is stored in the shift register 194. This stored data is used to calculate the autocorrelation function by the method using the fast Fourier transform already described. The quadrature detection waveforms thus shortened to the time length T2 are designated as 420 and 421 in FIG. 4, and the calculation result of the autocorrelation function thereof is designated as 422. Since the value of the number of crossings obtained here is 3 again, the evaluation result is rejected. As a result of performing the time window adjustment again in the process of transferring to the shift register 194, the quadrature detection waveform shortened to the time length T3 is designated as 430 and 431 in FIG. 4, and the calculation result of the autocorrelation function is designated as 432. Here, since the value of the number of crossings obtained again is 1, the evaluation result is acceptable.

This inspection is performed for all the received signals, and if all the received signals are acceptable, the time length for performing subsequent received signal cross-correlation is determined as T3. However, in this case, depending on the characteristics of the received signal, even if the effective time window width is shortened to a predetermined minimum time width or less, if there is a received waveform that rejects the evaluation result, do not reduce it further. To do.

Next, the cross-correlation coefficient between the received signals with the effective time width optimized is sequentially obtained for each received signal pair of the electroacoustic transducers that are adjacent to each other, and the time difference that maximizes the correlation is obtained. Here, the cross-correlation coefficient can be similarly obtained by the configuration and operation already described, only by setting one of the two mutually conjugate waveforms used in the calculation of the auto-correlation coefficient to be another received signal. Search for the cross-correlation coefficient stored in the memory circuit 75,
The time difference between the positions giving the maximum value is the corrected time value. It should be noted that the frequency domain data before the inverse Fourier calculation is performed in the process of obtaining the cross-correlation coefficient, 0 is added in reverse to increase the data length by the same operation as when the effective time length is changed. When the inverse Fourier calculation is performed, the accuracy of the obtained correction time value can be improved.

Using the time difference thus obtained, the delay time group for the transmission circuit 12 and the reception delay circuit 13 is corrected. In this case, the central processing unit 71 corrects the values stored in the internal registers of the wave transmission circuit 12 and the reception delay circuit 13 in accordance with the delay time calculation program held in the program storage circuit 74, so that accurate phasing is performed. Will be done.

Although the rectangular window is used as the window function in the above configuration, the Hanning window or the like may be used as described above. In this case, it can be realized only by a configuration in which a multiplier is newly added immediately before the input to the shift register 194.

The procedure for detecting the correction value is shown in FIG. 5 as a flow chart. When the correction is started (step 500), one of the electroacoustic transducers is selected and the reception signal is selected (step 510). After the quadrature detection of the received signal (step 52), the waveform is cut out by the time gate (step 53). A sufficient time width is secured for this cutout time width so that it can be dealt with by the maximum effective length of the window function later.

Next, the window function effective length is set (step 540). If the window function is a rectangular window, this effective length is
The maximum width of the coefficient portion where the coefficient value is not 0, and the cycle is the cycle in the case of the Hanning window. Here, in the present embodiment, the center of the window function matches the center of the signal section where the waveform is cut out by the time gate with the center of the window function.

Next, after confirming that the effective length is equal to or greater than the preset minimum value (step 541), if the result is equal to or greater than the minimum value, a window function is applied (step 5).
5). The autocorrelation coefficient calculation is performed with the result obtained by multiplying each value of the corresponding received waveform by each coefficient of the window function (step 56). This autocorrelation coefficient operation is a fast Fourier transform
(FFT) The complex conjugate of the calculation result is taken, and the product of the original result and the inverse FFT calculation can be used to obtain the product.

Next, the autocorrelation coefficient curve obtained by obtaining the absolute value of the inverse FFT calculation result is standardized by the maximum value (step 57). The standardized autocorrelation coefficient curve is inspected by a preset evaluation function (step 58). If the result of the inspection is negative, the process returns to step 540, and a window function having a shorter effective length in time is newly set. If the inspection result is acceptable, after the window function waveform or the effective length is stored in the storage circuit 75 of FIG. 7, it is confirmed that the processing of steps 510 to 58 has been completed for all the received signals. Confirm (step 511). If not completed, return to the point of selecting another received signal,
Steps 510 to 58 are repeated. If it is completed, the window function waveform or effective length is read from the memory circuit 75, and the corresponding received signals are multiplied, and the cross-correlation calculation 59 between the received signals with window function weights corresponding to the adjacent electroacoustic transducers is performed. Is done. Based on the result of the cross-correlation calculation, the delay time error in transmission and reception is corrected.

In the ultrasonic diagnostic apparatus shown in such an embodiment, first, after the time delay processing performed under the assumption that the sound velocity is constant, the reception waveform corresponding to each reception element cut out by the time gate is extracted. By calculating the autocorrelation function of the signal, it is possible to detect a similar waveform portion in the received signal caused by the diffraction of the sound field.

In this case, since the autocorrelation function does not have the maximum for each carrier wave period by performing the quadrature detection processing by the carrier wave frequency, the similar portion can be easily detected. Then, if the result of the autocorrelation calculation shows a value equal to or greater than a predetermined threshold value at a time position away from the position where the time difference is 0, it is determined that a significant similar waveform may overlap the time difference portion. , The effective time length of the window function by which the received waveform before the correlation calculation is multiplied is shortened by a predetermined width.

By optimizing the time length of the window function performed in this way, the waveform for which the cross-correlation calculation is performed is determined for each received signal, and the time difference to be corrected is obtained from the correlation calculation values. That is, the time difference to be corrected is appropriately obtained by the cross-correlation calculation between the reception signals with the window function weights.

Therefore, it is possible to eliminate the influence of the similar waveform portion (false image) caused by the diffraction of the sound field on the correlation calculation value when the cross correlation calculation is performed by such processing,
Extremely accurate phasing can be obtained.

Although FFT is used for the correlation coefficient calculation in the above-mentioned embodiment, it is not limited to this and other calculation methods such as simple convolution calculation may be used.

If necessary, the cross-correlation calculation between the received signals is performed only by the waveform cut out by the time gate before applying the window function, and after the tentative correction obtained, the waveform is cut out by the time gate again. Of course, the window function effective length may be set by the autocorrelation and the delay time error may be corrected by the final cross-correlation.

Further, the quadrature detection circuit 15, the time gate circuit 17, and the correlation calculation circuit 19 are configured such that a single one is sequentially assigned to different received signals or signal pairs by the signal selection means, but the present invention is not limited to this. Of course, a plurality of units may be provided to perform parallel processing.

Further, in the present embodiment, the phasing is performed with the received signal as an analog signal,
It is needless to say that the present invention is not limited to this, and may be performed after the digital calculation.

Further, in the present embodiment, the received signal for correction is acquired during repetition of transmission / reception under the same phasing condition, but the present invention is not limited to this, and in order to achieve more accurate correction. Needless to say, a configuration may be used in which correction is performed on the stored information by providing means for acquiring the received signals at the same time.

[0077]

As is apparent from the above description, according to the ultrasonic diagnostic apparatus of the present invention, extremely accurate phasing can be obtained.

[Brief description of drawings]

FIG. 1 is a schematic configuration diagram illustrating an embodiment of an ultrasonic diagnostic apparatus according to the present invention.

FIG. 2 is a conceptual diagram illustrating the principle of electronic focusing in a conventional ultrasonic diagnostic apparatus.

FIG. 3 is a conceptual diagram illustrating a conventional phasing error due to diffraction on a non-uniform sound velocity boundary surface.

FIG. 4 is a diagram illustrating optimization of a window function by an autocorrelation coefficient in the ultrasonic diagnostic apparatus of the present invention.

FIG. 5 is a flowchart illustrating an optimization procedure of a window function based on an autocorrelation coefficient in the ultrasonic diagnostic apparatus of the present invention.

FIG. 6 is a diagram illustrating an evaluation function of an autocorrelation coefficient in the ultrasonic diagnostic apparatus of the present invention.

FIG. 7 is a diagram for explaining an example of a configuration for realizing an optimizing operation of a window function in the ultrasonic diagnostic apparatus of the present invention.

[Explanation of symbols]

10 ... Electroacoustic transducer group, 11 ... Transmission / reception separating circuit, 12 ...
Transmission circuit, 13 ... Reception delay circuit, 131 ... Delay control circuit, 14 ... Selector group, 15 ... Quadrature detection circuit, 16 ... Adder, 17 ... Time gate circuit, 18 ... Display circuit, 19 ... Waveform memory, 191 ... Correlation calculation circuit, 20 ... Reflection source, 20
DESCRIPTION OF SYMBOLS 1 ... Reflected wave, 21 ... Delay amount, 22 ... Reflected signal, 30 ... Boundary of layers with different sound speeds, 31 ... Layers with different sound speeds, 320 ...
Main group of corrugations, 33 ... Additional waves due to some boundary shape.

Claims (5)

[Claims] 1. An ultrasonic diagnostic apparatus comprising means for correcting a delay time given to each signal of a transmitted wave or a received wave based on a time difference obtained by a cross-correlation calculation between the received signals. Means for obtaining an autocorrelation coefficient by multiplying the window function of the time length, means for optimizing the effective time length of the window function based on the calculation result of a predetermined evaluation function for this autocorrelation coefficient value, and each received signal An ultrasonic diagnostic apparatus comprising: a means for performing the cross-correlation calculation using a window function optimized for each and a result obtained by multiplying a received signal. 2. The ultrasonic diagnostic apparatus according to claim 1, wherein the received signal multiplied by the window function is subjected to quadrature detection processing with a carrier frequency. 3. The ultrasonic diagnostic apparatus according to claim 1, wherein the calculation result of the evaluation function is based on the number of crosses of the autocorrelation coefficient curve for a predetermined threshold value. 4. The ultrasonic diagnostic apparatus according to claim 1, wherein the calculation result of the evaluation function is based on the sum of the section lengths over which the autocorrelation coefficient curve exceeds a predetermined threshold value. 5. The ultrasonic diagnostic apparatus according to claim 1, wherein the calculation result of the evaluation function is based on a magnitude comparison between a square sum of the autocorrelation coefficient in a predetermined time difference section and a predetermined value.