JPS585649A - Method of processing ultrasonic echo signal - 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 The present invention relates to ultrasonic echo signals from different directions of incidence, such as in ultrasonic image processing in the field of material inspection and histology, where ultrasonic echo signals from different scanning directions are combined. Concerning methods of processing signals.

Ultrasonic transducers are generally used in the field of ultrasound echotomography, which simulates zero objects with a finite ultrasound antennae, similar to known X-ray tomography, in a direction related to a predetermined ultrasound antennae width. The action is disturbed (for example, G, W
Aco-ustic Echo C by adelika
computer tomography”, Acous
tical Imaging, Vol, Q, A, F
, Metherell (ed.), Plenum 1
980).

Since the object to be simulated must be completely covered with ultrasound waves, known tomography methods based on ultrasound principles have hitherto used wide flat piezoelectric transducers (see above-mentioned document). However, this type of method or device has drawbacks. For example, if the object is near the vibrator,
This usually results in amplification and extinction of the echo signals due to interference, which leads to a deterioration of the object simulation. Therefore, the pulse waveform of this type of vibrator is significantly worse than that of a vibrator with a small area. On the other hand, a small resonator with a wavelength Q of about 514 has a drawback of low sensitivity. Therefore, despite the fact that they have sufficiently wide ultrasonic antennae,
T! ! Very unsuitable: German Patent Application No. 3,038,052 already proposes scanning an object and imaging it by means of ultrasound echotomography. This is aimed at using ultrasonic transducers with narrow ultrasonic antennae. Such ultrasonic transducers produce short, intensely sounding pulses and are therefore much more suitable for ultrasonic echography applications than the previously mentioned ultrasonic transducers.

The method proposed in the above patent application publication takes the following measures to achieve the problem.

That is, the object to be reproduced is scanned from a number of scanning positions one after the other in a stepwise manner (i.e. according to a scanning process) by at least one ultrasonic transducer with a narrow ultrasonic antenna from different directions of incidence in each case; The switching of direction is oriented transversely to the ultrasound antennae, such that the echoes from each of the positions of the scanning process are summed, thereby achieving the effect of a wide ultrasound antennae. It is.

The 7'L E image displayed by the known B-scan method is
1 in the lateral direction, with a dark "smudge" constrained by a finite ultrasound antennae width. This "stain" can be reduced to a large extent with an appropriate filter, but
However, when a sensitive filter is applied, artifactual image artifacts appear, which impede diagnostic evaluation. Such artificial image disturbances are mainly caused by oscillation in the filter, for example.

Pure scanning equipment or equipment that performs similar copying methods is commercially available. In these, efforts are made to narrow the effective antenna width by using analog or digital aiming. Due to the complex nature of ultrasound propagation in biological tissue, and the details of which are not yet known,
This type of method is limited to a narrow range.

It has already been proposed to improve the resolution in the so-called B-scan method. Approximately 50 depending on the appropriate action
0 This operation is based on the idea of detecting the filter nucleus from the transmitter's ultrasonic antenna and the ultrasonic signal receiver. 〇 The theoretical resolution can be improved by appropriate treatment of the filter nucleus. does increase clearly, but in extreme cases an artificial disturbance occurs that appears stronger than the original real signal.

It is an object of the present invention to prevent artificial distortions in a method of processing ultrasound signals from various incident directions, such as ultrasound image processing in the field of materials inspection and histology, to the extent that image diagnosis is significantly impaired. The aim is to obtain a significantly improved resolution of the object without causing any damage.

This object is achieved according to the invention by the features defined in claim 1.

The invention is based on the recognition that artifacts are mostly stationary properties and are significantly influenced by local randomness when filtering image signals.

Therefore, the present invention is based on the fact that by averaging a plurality of independent images of this type, artificial artifacts can be suppressed and the real signal can be strengthened. It has the advantage that it is possible to obtain image quality that is significantly improved and facilitates diagnosis.

DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in detail with reference to embodiments with reference to the drawings.

Figure 1 shows three different diagrams a, b, and c for a certain temporarily fixed object cross-section, respectively showing the filter frequency characteristics used for filtering the partial image, the changed filter kernel frequency characteristics, and the frequency characteristics of the corresponding echo signal. The filtered contours of the two needles are shown. The horizontal axes of the top diagram each represent the position frequency.

The horizontal axes of the second diagram each represent the position coordinates of the overlapping filters. The horizontal axis of the bottom diagram each represents the horizontal coordinate of a B-scan transverse layer. The vertical axes of the top diagram each represent the relative amplitude of the filter kernel in Fourier space. The vertical axes of the second diagram each represent relative amplitude in position space. The vertical axes of the bottom diagram each represent relative signal amplitude.

FIG. 2 diagrammatically illustrates the method steps for carrying out the method of the invention as an embodiment of the invention, namely the generation of a number of initial images, the generation of intermediate images and the generation of a combined image.

As generally stated, FIG. 1 is diagram a.

Frequency characteristics of the filter used to filter the partial images in the form b, c (top diagram), modulated filter nuclear frequency transfer characteristics (middle diagram), two needles in the corresponding echo signal The filtered contour of the shaped part (lower diagram) is shown. From the triple diagram a, it can be seen that the resulting contour (bottom diagram) has very sharp edges on both needles, but its base area is strongly affected by image noise. becomes. Triple diagram b shows the slightly sharpened needle of #1 in the echo signal. Of course, the image noise level is so low that the depressions present between the two needles can be clearly distinguished. The remaining triple diagram C shows that the image noise level is even significantly lower. but,
At the same time, the sharpness of the needle representation is lost, which corresponds to the lowest resolution compared to both other triple diagrams a, b.

Therefore, it is clear that the second case of triple diagram b is the optimal form from the diagnostician's standpoint. This has a compromise between sufficiently large resolution of the object and sufficient suppression of image noise.

The number "160" at the beginning of each triple diagram.

'120' and 11971' respectively represent the filter core/cutoff frequency k, which indicates a quantity related to the filter characteristics.The optimum value =120 was determined for the biopsy based on theoretical considerations and actual tests.

According to the present invention, the image quality is further improved as follows. That is, in order to suppress noise that mainly appears in a steady distribution during high-resolution image generation, such as oscillation phenomena, and artificial data that occurs relatively thickly in object reproduction, a large number of initial images V with different scanning parameters are used.
□, v2...v.

is generated and is first filtered to take into account the 1 smear caused by the given ultrasonic antenna, resulting in an intermediate image 2.
．． 22...Many initial images v1゜■ to be made into Zn
・・・・・・V are combined, and at that time intermediate image Z□
.

2 n Z...The majority of the artificial data 2 n that becomes visible in Z is canceled out by the resulting averaging of the artificial data, so that a high-resolution combined image K results. It is to be.

As already explained, FIG. A large number of partial images T1. T2...
... Demonstrates the generation of Tn and the generation of a high-resolution, low-noise combined image.

The different scanning parameters are, according to the invention, variations in the scanning direction in one plane or in two planes, variations in the scanning distance only, variations in the scanning distance and scanning direction, ultrasound antennae and/or ultrasound pulses. or any static change.

According to the present invention, the initial image can be obtained using a known B-scan method, a known fan-shaped scan method to a known composite scan method, or an "antennary stain" is observed mostly in the tl main direction, and various parts of the initial image are They can be generated in any manner such that these principal directions in the range are different with respect to other principal directions.

According to the invention, the combination can be effected additively, multiplicatively or by directly or indirectly mixed additive/multiplicative superposition. Preferably, the combination is carried out with different weights for the individual scanning directions depending on the respective image structure.

In the main scanning direction, it is desirable to distribute over a complete circle or a partial circle at equal distances for initial image generation. However, the main scanning direction can also be given arbitrarily. For applications in the medical field, it is preferred that the main scanning direction covers only a small angular range, so as to be free, for example in the case of examinations between two ribs. In that case, the main scanning direction can cover two or more sectors.

According to the invention, each initial image V with a nucleus calculated from the antennal contour is laterally filtered and converted into an intermediate image Z. In this case, it is preferable that the filter nucleus is so strongly resolved as to cause secondary antennae in the unlikely event that it occurs.

However, the secondary antennae are at least partially re-averaged when superimposing multiple signals.

Both linear and non-linear filters can be used for filtering the initial image V. The partial image T, i.e. the intermediate image Z resulting from the initial image V, is preferably rectified before and/or after the filter; furthermore, the partial image T is smoothed at one or more optional points of the method. It is good.

Preferably, the filter also acts in the depth direction of the partial image. In this case, the pulse waveform of the ultrasonic signal can also be taken into consideration depending on the case.

According to an advantageous embodiment of the method according to the invention, the partial image T
, that is, the initial image V and/or the intermediate image Z1 and/
Alternatively, in the combined image K the residual image noise is reduced by further filters or by a global image processing scheme.

The filter used may be at least partially a pure convolution filter in position space.

Its characteristic quantities can be locally and/or globally dependent on the image content to be filtered.

Preferably, the partial image T is processed in one or more iterative steps depending on their mutual relationship into a combined image of the respective preceding iterative step. The filter or filters used in one or more iteration steps are determined depending on the combined image K of the previous iteration steps.

According to the invention, several or all individual method steps can be computer-controlled.

[Brief explanation of the drawing]

Figure 1 shows three different diagrams a, 'b, and c for a hypothetical target cross-section, showing the frequency characteristics of the filter used to filter the partial image, the modified filter kernel-frequency characteristics, and the corresponding FIG. 2 shows the filtered contours of two needles in the echo signal, and FIG.・Vo...Initial image I 2 Z Z...Z...Intermediate image 1.12n. TT ・・・・・・To ・・・・・・・・・Partial image I 2

Claims (1)

[Claims] l) In a method for processing ultrasound echo signals from different directions of incidence, such as ultrasound image processing in the field of material inspection and histology, where ultrasound echo signals from different scanning directions are combined. , in order to avoid artificial disturbances that occur to a relatively high degree with respect to disturbances that appear mainly in a stationary distribution when generating high-resolution images, a large number of initial images with different scanning parameters are generated, and these large number of initial The images are first filtered into an intermediate image and then combined, taking into account the "smudge" caused by the ultrasound antennae, and most of the artifact artifacts that become visible in the intermediate image are then combined. A method for processing an ultrasound echo signal, characterized in that the components of intra-artificial disturbance are canceled out by averaging to produce a high-resolution combined image with relatively little disturbance. 2) Processing method according to claim 1, characterized in that the different scanning parameters are produced by changing the scanning direction within one plane. 3) A processing method according to claim 1, characterized in that the different scanning parameters are produced by changing the scanning direction in three dimensions. 4) Processing method according to claim 1, characterized in that the different scanning parameters are produced by varying the scanning distance. 5) Processing method according to claim 1, characterized in that the different scanning parameters are produced by varying the scanning direction and the scanning distance. 6) Processing method according to claim 1, characterized in that the different scanning parameters are produced by varying the ultrasound antennae and/or the ultrasound pulses. 7) A processing method according to claim 1, characterized in that the different scanning parameters ζ parameters are produced by arbitrary steady-state changes. 8) The processing method according to claim 1, characterized in that the initial image is generated by a known B-scan method. 9) The initial image is generated by a known fan-shaped scan method. A processing method according to claim 1, wherein: 10) The processing method according to claim 1, wherein the initial image is generated by a known composite scanning method. 11) The initial image can be formed in any manner in which 6 antennal blotches "f!!" are recognized in the main directions, and these main directions in the various partial images of the initial image are different with respect to the other main directions. 12) The processing method according to claim 1, characterized in that the combination is performed by superposition by addition ◎ 13 ) The processing method according to claim 1, characterized in that the combination is performed by superposition by multiplication. 14) The combination is performed by adding/mixing indirectly or directly.
15) The processing method according to claim 1, characterized in that the combination is carried out by superimposition by multiplication. A processing method according to any one of claims 12 to 14. 16) The processing method according to any one of claims 1 to 15, wherein the main scanning direction for generating the initial image is distributed equidistantly over a complete circle or a partial circle. 17) Four scan directions are arbitrarily given for initial image generation. The processing method according to any one of claims 1 to 15, characterized in that: 0 18) The main scanning direction covers only a small sector to free the space between two ribs during medical examinations, for example. The processing method according to claim 17, characterized in that: J9) The processing method according to claim 18, wherein the main scanning direction covers two or more fan shapes. 20) Processing method according to any one of claims 1 to 19, characterized in that each initial image is laterally filtered into an intermediate image with a kernel calculated from the antennal contour. 21) A processing method according to claim 20, characterized in that the filter kernels are so strongly resolved as to produce secondary antennae that are at least partially re-smoothed upon superimposition. 22) A processing method according to claim 20 or 21, characterized in that a linear filter is used. 23) The processing method according to claim 20 or 21, characterized in that a nonlinear filter is used. 24) Claim 2, characterized in that the partial images are rectified before or after filtering, or before and after filtering.
The processing method according to item 2 or item 23. 25) Processing method according to any one of claims 20 to 23, characterized in that the partial image is smoothed at one or more arbitrary points in the course of the method. 26) The filter is a partial image 27) Part of the processing method according to any one of claims 1 to 25, characterized in that the processing method acts in the depth direction of images, i.e. initial images and/or intermediate images;
and/or in the combined image, the residual image noise is reduced by further filters or by a global image processing method.
The treatment method described in any of the above. 28) Processing method according to any one of claims 1 to 27, characterized in that the filter used is at least partially a pure convolution filter in position space. 29) The characteristic quantities of the filter or filters used depend locally and/or globally on the image content to be filtered. Processing method according to claim 1, characterized in that the partial images are processed in one or more iterative steps depending on their mutual relationship in each case into a combined image of the preceding iterative steps. The processing method according to any one of Items 29 to 29. 31) One or more filters used in one or more iteration steps are determined depending on the combined image of the preceding iteration steps. Processing method described in . 32) Claim 1, characterized in that each or all of the method steps are computer-controlled.
32. The processing method according to any one of Items 31 to 31.