JPS6274346A - Ultrasonic diagnostic apparatus - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] [Technical Field of the Invention] The present invention relates to an ultrasonic diagnostic apparatus that improves the azimuth resolution in the scanning direction according to distance changes in the distance direction in sector scanning, and particularly requires two-dimensional Doppler scanning. The present invention relates to improving the image quality of an apparatus for obtaining B-mode images of an ultrasonic blood flow imaging apparatus.

[Technical background of the invention and its problems]

Typically, real-time B-mode displays move an ultrasound beam across a subject by electronically scanning each probe in a group of probes, and with each pulsed emission, the internal tissue of the subject is The probe receives an ultrasonic echo signal from the transceiver, processes it in the transmitting/receiving section, converts it into a digital signal, and sequentially stores it in a line buffer memory group. Raster data is cyclically generated from this line buffer memory group. The data is read out, added by an adder, stored in a frame memory, subjected to image processing, and displayed on a TV monitor.

However, in this type of device, there is a problem in that a lateral flow occurs in a back B-mode image in a blood flow imaging image. That is, the wire phantom image during sector quantification shown in FIG. 17 (conditions: 4-rate repeating scan, scan angle 9
As shown in 64 0° rasters and a depth of field of 15 cm, phantoms with wire diameters of 11 or less are also displayed as elongated in the scanning direction due to the fact that the ultrasound beam is not sharp. When scanning the same raster multiple times in succession, the only way to display the viewing angle and frame period without deteriorating is to reduce the number of scanning lines, and for this reason, it is necessary to consider the characteristics of the scanning direction. be.

Next, to further explain the above problem using a schematic explanatory diagram showing the processing from the scanning beam transmitter/receiver output to frame memory writing shown in FIG. 18, FIG. 18(a) is captured as one point in the tomographic image This is the response function estimated from the scanning beam transmitter/receiver output of the power wire image. This shows the wire phantom image at a depth of 80 fins (focus point), and in subsequent simulations, it was approximated to a Gaussian function in the scanning direction. Figure 18 (b
) is a process in which the scanning beam transmitter/receiver output (line data) as shown in Fig. 18(a) is digitally converted into raster data and written into the frame memory. There is an operation to coordinate coordinate transform and write it into the frame memory, but this has the problem that the sampling distance in the scanning direction changes depending on the distance in the distance direction, and secondly, the spatial There is a problem in that it can be regarded as a superposition of the above response function. In FIG. 18('C), when the coordinates are converted and written into a digital square called a frame memory, it is necessary to pixelize the line data according to the magnification in the distance direction according to the display magnification. moreover,
Interpolated data is generated between these written rasques. A display image is constructed by such a route.

From the above, if we consider the wire phantom as a star-like point source treated optically and assume that the transmission path of the ultrasonic device does not change at a certain distance and always has a linear response, then the image on the frame memory can be considered to have systematic errors in the transmission path of the ultrasonic diagnostic device.

To further explain the systematic error in the scanning direction of the above problem using FIG. On the other hand, the fact that a point reflection source such as a wire phantom is missed by the transmission path of the ultrasound device is considered to be a systematic error in the transmission path of the ultrasound device, and this is expressed as a function of the scanning direction called B (x). Then, when actually writing to the frame memory, convolution integration is performed as C(x) = A (x) *B (x), and this C(X) is further digitized. In the frame memory, the gradation data becomes dull in the scanning direction and has discontinuity due to digitization. Based on this idea, the 18th
By inserting a correction circuit into the signal processing path shown in the figure, it becomes possible to return to the original wire phantom image, and blurring in the scanning direction is eliminated.

[Purpose of the invention]

The present invention has been made in view of the above circumstances, and it is an object of the present invention to provide an ultrasonic diagnostic apparatus equipped with a correction means capable of returning to the original wire phantom image.

[Summary of the invention]

The outline of the present invention for achieving the above object is to sample an ultrasonic echo signal obtained in response to an ultrasonic pulse, convert it into a digital signal, and sequentially store it in a line buffer memory group. In an ultrasonic diagnostic apparatus that performs image processing while cyclically reading out raster data from the line panzer memory, convolution calculation is performed on the raster data at the same depth read out from the line panzer memory with a function that changes depending on the depth. It is in.

[Embodiments of the invention]

The principle of the present invention and its embodiments will be described below with reference to the accompanying drawings.

First, the principle of the present invention will be explained.

In Fig. 18, C(x) has systematic errors as mentioned above, so corrections are made for this.The correction methods include correction in the spatial frequency domain and correction in the distance domain (scanning direction). This can be considered as a correction. Therefore, in the following embodiment, as shown in FIG.
), we decided to obtain the correction data A'' (X) by convolving D (X), which is the inverse function (B (x) *D (x) - δ(x) delta function), with C (x). This relational expression is expressed as follows using the explanatory diagram of FIG.

However, in the ultrasonic lasque on the CJ+1' scanning side,
Reflected echo data D4 for the th raster distance j. . ~Dj+11' Complementary distance number series A"4 at distance j
．． , - INT (n/2): processing data of scanning force direction i - I NT (n/2)-th raster distance j in the ultrasonic raster on the display side (frame memory), In this case, on the display side, INT ( n/2)≦i≦N −I NT(n/2).

Here, the correction function series 5. , is a function that varies depending on the raster distance j, that is, the depth. For example, in sector scanning, the beam spreads in a fan shape, so the sampling pitch in the arc direction differs between shallow and deep areas. In other words, the number of data for an arc of the same length is small in the deep part (
, many in shallow parts. Therefore, in order to have the same correction function series for arcs of the same length, this series is 4. ,
is a wide even function in the shallow part, and a narrow even function in the deep part.

The program diagram to which the correction means using the above relational expression is applied and the principle of basic timing are explained with reference to Fig. 3 (a) and (b). It consists of a scanning beam transmitter/receiver that performs scanning, a frame memory that sequentially writes data obtained by scanning with a probe onto a two-dimensional surface and reads it out asynchronously at the cycle of the TV monitor, and a data processing unit that is a correction means.
Among reflected echo images of ultrasonic waves, the B-mode image in particular is scanned in sectors and linearly, and the image is constructed in a frame memory and displayed on a TV monitor. At that time, when the scanning beam transceiver output is written to one frame memory,
One-dimensional line data (A mode signal) in which the scanning direction is fixed and reflected echo data changes in the depth direction over time.
will be transferred as During this transfer, the data processing section, which is the correction means, performs correction processing to improve the azimuth resolution in the scanning direction. This correction processing involves passing raster data of several bones in the scanning direction through the scanning system. A correction function series for correcting the obtained systematic error in the scanning direction is convolved, and the correction function changes depending on the distance direction and the scanning angle and focus conditions such as multi-step focusing. In FIG. 3(a), numeral 1 denotes a line buffer memory, which stores n lines of raster data in the scanning direction and transfers it to the right every raster cycle. Reference numeral 2 denotes a weight ROM multiplier whose weight series changes according to the distance direction, and the data read from the line buffer memory 1 is combined with the raster data outputted at the same distance direction clock as the scanning beam transmission/reception signal according to the distance direction. It is multiplied by the output signal from the weight ROM multiplier circuit 2 whose weight series changes. 3 is an adder which adds the respective multiplied signals and writes the result into the frame memory 4.

In this case, the scanning method scans the same raster four times, and the blood flow data (displayed in color) requires one minute for four rasters for sampling and one minute for 2 rasters for calculation, so it is actually scanned. It is output with a delay of one period of 6 rasters from the start, and the output is also once every period of 4 rasters.

In contrast, echo data (black and white) from the scanning beam transceiver can be captured without any time lag with respect to the scanning raster.

However, in this case, if a delay can be applied to the echo data processing, the echo data and blood flow data can be taken into the frame memory 4 at the same timing, so that the hardware can be made smaller. Therefore, by making effective use of free time, input echo data is stored in line buffer memory 1, and the data up to two rasters ago and the currently scanned data are stored in line buffer memory 1.
The echo outputs of the same scan scan are added by an adder 3 using a weight ROM multiplication circuit 2 whose weight sequence changes depending on the distance. In this way, the correction function is convoluted and outputted to the write interpolation block, which is the frame memory 4, at the same timing as the blood flow data.

Figure 4 shows the spatial frequency characteristics of systematic errors in the scanning direction at a distance of 15 cm (however, the device used has an electronic focus point at a distance of 3 am, and 1 pixel corresponds to approximately 0.5 fins). ), and the systematic error (function giving the particle shape) in the scanning direction of the echo data is given on the spatial frequency axis.

This has a unique function depending on the ultrasound diagnostic equipment, and the distance (
Depth). Therefore, the systematic error indicates the spatial frequency characteristic of the systematic error in the scanning direction in units of one pixel of the frame memory when the distance is fixed at 15 cm based on the measurement results of an actual device. Furthermore,
For ease of later simulation, the systematic error is approximated by a Gaussian function, and the 40 dB echo gain width at a distance of 15 cm is set to 4σ, and j (x) = −exp (−x ” / δ2)ζ
``Approximated by a function of δ.

G, (ω) = exp () becomes.

However, based on actual measurements, 13mm=25pix is set.

1 pix is the length of one pixel in the frame memory.

Figure 5 shows the spatial frequency characteristics in the scanning direction by sampling, and to give the unit 1 pixel of the frame memory.
The unit is . Further, the sampling interval in the scanning direction is α (pixel).

In the case of linear scanning, this characteristic does not change depending on the distance, but in the case of sector scanning, the sampling interval α in the scanning direction increases as the distance increases, and the cutoff frequency of the spatial frequency increases due to interpolation. decreases (the spatial frequency of 1/2α is called the cutoff frequency)
. Furthermore, since the spatial frequency in the scanning direction becomes a zero-order holder, its transfer function Gz(s) is as follows, and the spatial frequency characteristic is as follows: Gz(ω)=17. When the cutoff frequency ωα that is I″″i″″α is calculated, □=β...■ is obtained. Also, since the distance and the sampling distance in the scanning direction are proportional, α-r-Depth ...■ However, γ: proportionality constant Depth: distance ■, ■ From the formula, 2・γ ω α = □ α・Depth, and the Depth is The larger the cutoff frequency ω
It can be seen that α becomes smaller.

FIG. 6 shows the spatial frequency characteristics of the correction circuit when the correction weight series for each of the three rasters of scanning lines is set to 20''-0,5,a,=2.0.a2=-0,5.

Note that these correction weight series change depending on the distance and the spatial frequency characteristics of the device in the scanning direction. Furthermore, since processing is performed in units of raster, in the case of sector scanning, the sampling distance α in the scanning direction changes depending on the distance, similar to the sampling characteristics shown in FIG. As shown in FIG. 6, it can be seen that it acts as a high-frequency emphasis filter in a spatial frequency region lower than the sampling distance twice α. This is effective when correcting high-frequency components of 1/2α or less that are lost due to interpolation processing or the like. Furthermore, Cz+(z)=aoz2+a+z +az/z", and the spatial frequency characteristic is G3(ω) = SQRT (Re(G3+4w))
"+Im(G3LJw+)2).Additionally, by adding the characteristics of linear interpolation in n pixel interval, G, (z) =--(z'-'+z'-"+ -=
+1)/z'-'-Gz(z) Furthermore, since it is considered to be given by the sampling characteristic Gz(z) and the moving average characteristic, it is given by G, (ω)=.

Figures 7 to 9 outline the first method;
A pair of line buffer memories a and lb store n lines of raster data in the scanning direction, and are transferred to the right every raster cycle. In this case, line buffer memories la and lb are paired as line buffer memory a and lb are RD (read) and WR (write).
This is because the operations cannot be performed at the same time, so the operations are switched alternately every raster period. Its operation is as shown in FIG. 8, and each line buffer memory la
, lb pair, the RD data is assumed to come out with the same distance direction clock as the signal from the scanning beam transceiver 5,
Weight generator 2a whose weight sequence changes depending on the distance direction
~■ and the respective raster data are multiplied by the multiplier 2b, and these multiplied data are added by the adder 3, written into the frame memory 4, image processed, and displayed on the TV monitor 6. In this way, the number raster in the scanning direction and the weight series whose weight changes depending on the distance in the distance direction are combined into composite j1. - Frame memory 4
This is what you write to.

FIG. 9 shows the relationship between the scanning raster of the probe 7 and the scanning raster of the frame memory 4 in Method 1, and the coordinates and length units of the scanning side of the ultrasound beam and the display side of the TV monitor 6 are 1. It corresponds to 1:1. 1
.about.N is a raster on the scanning side, which is scanned in the order shown in FIG. 8 during one ultrasonic screen configuration period. On the other hand, when the display side convolves a correction function that is symmetrical in the scanning direction, it starts from half the value of the n series of correction functions,
N-n pieces of corrected raster data are generated up to the value obtained by subtracting half the value from N.

Figures 10 to 12 outline the second method, which differs from the first method described above by scanning the same raster m times on the scanning side and then moving to the next raster. This means that only one line period is written in the data obtained for m947 minutes by scanning m times, and in the remaining periods, the written data is repeatedly read out until the writing period of the next scan arrives. In addition, in order to cause the right side to write what is read by the left line buffer memory 1 during transfer between the line buffer memories 1, a maximum of m weight sequences are required. The subsequent processing, like the first method described above, is to convolve line data for several rasters in the scanning direction with a weight series of a correction function series that changes depending on the distance, as shown in Figure 11. RD and WR
It operates at the timing of FIG. 12 shows the relationship between the scanning side raster and the display side raster as in the first method, and the scanning side is 1, 2, . ... Scanning is performed for m pieces up to 1M, but the display side generates raster data that has been subjected to correction processing of M - m + 1 lines from α, which is half of m, to β, which is half of m from M. It represents that.

Figures 13 and 14 show raster data in the scanning direction [Yes] to ■ and the weight series of the azimuth resolution correction function in the scanning direction ■ to ■.
It shows convolution processing with RATE
1 refers to the above-mentioned 1. This is the Lascou periodic signal of the second method, and PIXCK 1 is a clock signal that gives the unit in the distance direction.
It is 7ri. The reflected echo signal of the ultrasound is RATE 1
Since the transducer of the probe 7 is excited and received at the Z timing of RATE, the time from Z of RATE is proportional to the distance, assuming that the speed of sound is constant. This RATE 1 clears the counter, PIXCK 1 increments the counter, gives distance data to the weight generator 2a, and generates weights according to the display field of view, scanning angle, number of scanning lines, focus point, etc. It is configured to generate a series of correction series data from the devices ① to ◯, and perform convolution with the data in the line buffer memory 1 of ① to ◯, respectively.

Since the present invention is made as described above, as is clear from the path of the scanning direction signal to the display plane of the frame memory shown in FIG.
(x), there is a relationship in the distance domain (referring to the inverse domain of spatial frequency) in the scanning direction, and in the spatial frequency domain, H(ω)=G'(ω)・F(ω) Become a relationship.

In addition, the zero-order holder characteristic CZ (ω) and the characteristic G, (ω) of linear interpolation, which operates at a shorter period than the holder period such as interpolation, are added, and the characteristic L in the frame memory display is
(x) is expressed in the frequency domain L(ω) as L(ω)=
H(ω)・CZ(ω)・Ca(ω). on the other hand,
Adding the aforementioned correction characteristic G3(ω) to this, L'(ω)=H(ω)・02(ω)・G, j(ω)・G
4(ω), indicating that correction is possible.

An example of the effect of the correction process is shown in FIG. 16, which shows the spatial frequency characteristics in the scanning direction at a distance of 10 cm when a magnification is given so that IPix #0.5 as is displayed on the frame memory. Figure 16(a) shows L(ω), which is the characteristic of the conventional device, and Figure 16(b) shows the weight sequence -0.
, 5, 2, 0, -0, 5. FIG. 16(b) shows the characteristics of the device of the present invention that has undergone correction processing of . ing. still,
Although the present invention has been exemplified by sector scanning in the above embodiment, it goes without saying that it can also be applied to linear scanning, etc. in which the beam width is changed depending on the depth.

〔Effect of the invention〕

As described above, the present invention is configured to convolve the correction function series D4.3 for correcting systematic errors in the scanning direction that occur through the scanning system into raster data for several lines in the scanning direction, so that the high frequency region characteristics are corrected. This eliminates the lateral flow phenomenon in the B-mode image.

[Brief explanation of drawings]

1 and 2 are schematic diagrams illustrating the principle of the present invention, 3(a) and 3(b) are block diagrams and data processing timing charts illustrating the principle of the present invention, and 4-
and Figure 1' Figure J is a graph showing the frequency characteristics of the present invention, Figure 7.
9 to 9 show an embodiment of the first method, respectively. FIG. 7 is a block diagram, FIG. 8 is a time chart of data processing, and FIG. 9 is a schematic explanatory diagram showing the relationship between scanning rasters. It is. FIGS. 10 to 12 show an embodiment of the second method, with FIG. 10 being a block diagram, FIG. 11 being a time chart of data processing, and FIG. 12 being a schematic explanatory diagram of scanning Lasque. Figures 13 and 14 show the convolution process of raster data and correction function weight series.
FIG. 3 is a block diagram, and FIG. 14 is a time chart of data processing. FIG. 15 is a schematic explanatory diagram showing the path of the scanning direction signal to the display plane of the frame memory, and FIG. 16 (
a) is a graph showing conventional frequency characteristics, Fig. 16(b)
is a graph showing the frequency characteristics due to the correction processing of the present invention, FIG. 17 is an explanatory diagram showing a wire phantom image by sector scanning, FIG. 18 is a schematic explanatory diagram showing the processing from the scanning beam transmitter/receiver output to frame memory writing, FIG. 19 is a schematic explanatory diagram showing systematic errors in the scanning direction of the ultrasonic diagnostic apparatus. 1, la, lb...line ha sofa memory, 2...
Weight ROM multiplication circuit, 3... Adder, 4... Frame memory. Agent Patent attorney Noriyuki Ken Yudo Ogo Norio 1st autopsy Rough 11th original forgery 7- K(x) L(x) Figure 16 (b) Aiji Toru Figure 17

Claims (1)

[Claims] Ultrasonic echo signals obtained in response to ultrasonic pulses are sampled, converted into digital signals, and sequentially stored in a line buffer memory group, and image processing is performed while cyclically reading raster data from this line buffer memory group. An ultrasonic diagnostic apparatus characterized in that the ultrasonic diagnostic apparatus is configured to perform a convolution calculation on raster data of the same depth read from the line buffer memory with a function that changes depending on the depth.