WO2002073232A1

WO2002073232A1 - System for image reproduction for ultrasound computer tomographs

Info

Publication number: WO2002073232A1
Application number: PCT/EP2002/002637
Authority: WO
Inventors: Miroslaw Wrobel; Frank Heister; Janet Grassmann
Original assignee: Sonem Gmbh
Priority date: 2001-03-14
Filing date: 2002-03-09
Publication date: 2002-09-19
Also published as: DE10112034A1

Abstract

The invention relates to a system for image reproduction for ultrasound computer tomographs. The invention is based on a development of image reproduction using ultrasound with optional modulation function. The image reproduction system for ultrasound computer tomographs comprises at least two transmitters and at least two receivers that are arranged around the object in known relation to one another so that the object is exposed to the ultrasonic waves and the response signals are received. The transmitter signals have different modulation functions that do not mutually influence one another. The system further comprises a correlation unit in which every received signal is correlated with every transmitted signal, a calculation unit for the parameters of the reproduction, provided with a calculation unit for the envelopes, a calculation unit for the index matrix and a downstream integration unit for the index matrix, as well as a display unit.

Description

Arrangement for image reproduction for computer tomographs with ultrasound

The invention relates to an arrangement for image reproduction for computer tomographs with ultrasound. The ner driving is based on a further development of the image reproduction when using ultrasound signals with any modulation function.

The idea of quickly acquiring nolume data using ultrasound signals is based on the principle of imaging systems in computer tomography (computed tomography). In the following, the well-known principle of computer tomography will be briefly presented in order to motivate the measuring arrangement and the reconstruction method for the 3D ultrasound system.

3 shows the measuring principle of a computer tomograph. An object 3 is arranged in a fixed coordinate system (x, y) and thus represents a location-dependent brightness distribution f (x, y). The actual measuring device is arranged in a rotatable coordinate system (p, θ). The fulcrum of the coordinate system (p, θ) and that of the fixed system (x, y) are identical. The projections P (p, θ) 14 now denote the attenuations of an X-ray beam of the intensity I ₀ along the radiation paths shown in FIG. 3. If the projections P (p, θ) 14 are carried out for all normal distances from origin p and all angles from 0 ° to 180 °, the so-called Radon transform of the brightness distribution f (x, y) is obtained via object 3.

4 shows an original image A, its radon transform B and the result of the inverse radon transformation C. The radon transform B contains all the information necessary for the reconstruction.

A computer tomograph was also known, which uses microwaves or

Ultrasonic waves are working. Ultrasound is preferably used here in the reflection process. The image display methods used correspond to those of Radar technology. While both narrowband and broadband signals can be used in the transmission method, only broadband signals, for example in the form of short pulses, are used in the reflection method. The energy received after passing through the object or the arrival time or transit time of the signal are used for evaluation. A B-image is generated, in which the brightness is controlled by the intensity of the ultrasound echoes and the position of the illuminated point by the sound beam direction and the echo time on a screen. The ultrasound head can also generate a plane wave instead of a spherical wave, ie the ultrasound head can consist of a linear transducer.

DE 199 15 583.6 discloses another device and a driving method that is used to evaluate the response signals of ultrasound signals that are transmitted to an unknown structure. When driving ner, an unknown object is irradiated with any modulation function by at least one transmission signal. The response signals are then picked up by at least three receivers known in their position relative to the transmitter and correlated with the transmission signal. The paths of the transmission signal to the individual receivers are calculated in a downstream calculation unit via the reflective structure. In a second calculation unit, the spatial coordinates of the reflective structure are then determined from these signals. The basis is that the locations of the same distance between the transmitter and the receiver lie on an ellipse. The signals are therefore no longer evaluated only on the basis of the transit times, but rather the signal paths.

In the ner driving described above, the intersection of the possible reflection locations is determined to determine the spatial coordinates of a reflector. This intersection is unambiguous, in the time series recorded on all receivers, assigned a value which is determined by the transit time of the signal. The signal values determined with regard to this spatial point now determine the brightness of the assigned pixel in the reconstructed image.

The described driving works without errors assuming the existence of individual, point-like reflectors. The disadvantage of the method described above however, consists in the forced reconstruction of artifacts in the case of a certain number and constellation of the reflectors.

5 shows a scenario with a transmitter, two receivers and three object points. The reflectors are each on their associated elliptical arches. The intersection points of the ellipses now determine the spatial coordinates of the reflection points to which the values of the reflected energy corresponding to the transit times are assigned. In the constellation shown in Fig. 5, the designated artifact must be reconstructed. The assignment of the reflected energy between the receivers can no longer be made unambiguously from a certain number of reflectors. The minimum number of reflectors is three for two receivers. Adding an additional receiver would solve the problem for this particular constellation. The minimum constellation at which phantom points 10 would have to be reconstructed would then contain four object points 3. With a number of N possible object points, (N-1) receivers are necessary in order not to have to reconstruct phantom points. A corresponding consideration applies to the number of transmitters.

The object of the present invention is therefore to provide an improved measuring system which, like the computer tomograph, supplies the radon transform of the object to be imaged as a result of the measurement. The novel device is intended to eliminate the disadvantages described above and to correctly reconstruct a set of N object points with a smaller number of transmitter and receiver elements.

The object is solved by the appended claims. Arrangement for image reproduction for computer tomographs with ultrasound consists of any number but at least two transmitters and any number but at least two receivers, which are arranged in known positions relative to one another around the object, so that the object is completely sonicated and all response signals are received. The transmission signals of the individual transmitters have different modulation functions that do not influence one another. The received signals are digitized. In a correlator unit, every received signal is correlated with every transmitted signal. The correlator unit consists of a matrix of correlators with the size of the product of the number of transmitters and the number of receivers, each transmitter being connected to a correlator of each receiver and each receiver being connected to a correlator of each transmitter. The calculation unit consists of a calculation part for the parameters of the image, a calculation part for the envelope, a subsequent calculation part for the index matrix and at least one discrete integration unit. A display device is coupled to the calculation unit. After the calculation part for the envelope curve, a windowing in the frequency range can be provided.

Since the calculation unit processes digital signals, the received analog signals are converted into digital signals. Corresponding A / D converters can be connected upstream or downstream of the individual correlators.

The calculation unit for the parameters of the figure has first storage matrices for storing the individual correlation results between the individual received signals and the respective transmission signal and second storage matrices for storing the sums of the individual correlation results from the first storage matrix.

The arrangement of the ultrasound components provides angle-dependent projections of the measurement object. Regardless of the starting point of an ultrasonic transmitter, a signal is emitted in all spatial directions. This can be both a frequency-modulated signal and a short pulse. Under the

A prerequisite for the principle of supeposition for mechanical waves is that all potential reflection locations that make a constructive contribution to the reflected energy at a certain point in time have the same distance from the ultrasound transmitter to the ultrasound receiver. The points thus determined are arranged on an ellipse arc around E ₁ and Si (FIG. 5). Since the distance between the transmitter and the receiver is known, the extent of the ellipse arc depends only on the distance between the transmitter and the reflector and therefore only on the corresponding signal propagation time. The time series that is recorded at the receiver E ₁ corresponds to the projection P with respect to an angle θ (FIG. 3). The angle θ is defined as the angle between the normal to the distance between the transmitter and receiver and the x-axis of the fixed coordinate system. The angle θ is varied by the combination of transmitters and receivers, even though the ultrasound elements are in a static arrangement. By varying θ in the range 0 ° to 180 °, the analogue to the radon transform is finally obtained. Simulations have shown that the arrangement of the transmitting and receiving elements on a semicircle (hemisphere) is not absolutely necessary for the basic functioning of the reconstruction process.

The transmission and reception characteristics of the ultrasound elements do not necessarily have to be 180 °. It is only necessary that the object area of interest is illuminated uniformly. This must be taken into account in the design of an ultrasound head by aligning the ultrasound elements depending on their placement within the ultrasound head.

For reasons of better clarity, only the two-dimensional case is considered below. Since a uniform spherical propagation of the ultrasound takes place starting from the sound source, instead of the ellipses, rotationally symmetrical ellipsoids have to be considered for the real three-dimensional case. With the measurement method described, it is possible to obtain only volume data without a special arrangement of the reflection locations. The reconstruction process can be expanded to three dimensions without having to make any fundamental changes. All representations and methods of the two-dimensional case can be applied analogously to three dimensions.

It is crucial for the arrangement presented that the principle is not based on the measurement of the attenuation along a irradiated path, but on a measure of the reflected energy along a curved path. The underlying signal processing is explained below and corresponding references to the reconstruction method are given.

In principle, an ultrasound transmitter is combined with an ultrasound receiver. The signal (frequency-modulated signal or pulse) propagates from the transmitter in all spatial directions. In parallel, all receivers are recorded, so that all possible transmitter-receiver combinations are available for the reconstruction of the active transmitter. This procedure is repeated in sequence for all transmitters. In this way, all possible combinations of transmitters and receivers are available for reconstruction. The signal processing path is shown below on the basis of the signal recorded at the receiver.

After the measurement, the correlation with the associated FM code (code of frequency modulation) takes place. The broadcast signal. The envelope curve is calculated by bandpass filtering, squaring the signal and then lowpass filtering with appropriately designed digital zero-phase filters. Windowing in the frequency domain is carried out with one of the following window functions: 'Hamming', 'Hanning', 'Shepp-Logan', and 'Cosinus'.

For the reconstruction, the mathematical background should also be for the two-dimensional case, i.e. the reconstruction algorithm. As already mentioned, the process can be expanded to three dimensions without any problems. As already mentioned above, the reflection sites whose backscattered energy are added at the same times in the time series recorded on the receiver are on ellipses with transmitter and receiver in one focal point each. The parameters of the rotated ellipse are clearly determined by the distance and the time that the sound takes to travel from the transmitter to a point of the ellipse to the receiver.

The described mapping of the time series on the corresponding elliptical arcs gives a so-called stripe image with respect to the transmitter and the associated one

Consignee. The number of transmitter-receiver combinations used results in the number of resulting stripe images. An index matrix must be calculated for each transmitter-receiver combination. The intersection of the index matrix (FIG. 9) with the plane parallel to the object plane always results in an ellipse. For the three-dimensional case, a three-dimensional matrix is obtained and, instead of the ellipses, rotationally symmetrical ellipsoids.

The construction of the index matrices is shown as follows:

Basic equation of the ellipse:

4b ² + d ²

Rotation and translation:

D; Rotation matrix, t = translation vector

j -2J

(G1.1)

Starting from the basic equation of the ellipse in parameter form, the equation of the rotated and shifted ellipse is derived as described above. As already mentioned above, the parameters d, φ and the vectors t, y of the rotated ellipse are known or can be derived from existing quantities. The only unknowns in Eq. 1 are xi and the half parameter b of the ellipse, which is time-dependent on the signal propagation time. Equation 1 can be solved for these two unknowns and you get exactly one real solution for the parameter of interest b. Because of this solution, the signal propagation time from the transmitter can be calculated to any point in the object area. The result is the index matrix described above.

The complete reconstruction of a grayscale image is obtained by adding the individual stripe images. The gray value of each pixel thus corresponds to the integral of the angles which are reflected in the energy.

The invention will be explained below using an exemplary embodiment. The same reference numerals designate the same or similar parts in the individual drawings.

Fig. 1 shows the basic structure of an arrangement for image reproduction for

Computer tomographs with ultrasound; 2 shows an illustration for explaining the reconstruction of images;

3 shows an illustration of the measurement principle of computer tomography;

4A-4C show an image, its radon transform and the result of the inverse

Radon transform;

5 shows a diagram for explaining the forced reconstruction of an artifact (object);

6 shows a two-dimensional representation of the measurement principle for data acquisition;

Fig. 7 shows the way of data processing;

8 shows an illustration for explaining the calculation of idex matrices on the basis of the ellipse equation; and Fig. 9 shows an index matrix for constructing a stripe image.

1 shows the basic structure of an arrangement for image reproduction for computed tomography with ultrasound. The arrangement for image reproduction for computer tomographs with ultrasound consists of any number of transmitters 1- 1 to 1-s and any number of receivers 2-1 to 2-e, which are known in the art

Positions to each other are arranged around the object 3, so that the object 3 is completely sonicated and all response signals are received, the transmission signals of the individual transmitters 1-1 to 1-s have different modulation functions that do not influence each other (FM code 6-1 to 6-s). The individual transmitters 1-1 to 1-s are coupled to signal generators 4-1 to 4-s, which generate transmit signals with the desired modulation functions that do not influence one another. In this exemplary embodiment, the individual receivers 2-1 to 2-e are A / D converters 5-1 to 5-e. A correlation unit 17 follows with the size of the product of the number of transmitters 1 and the number of receivers 2, each transmitter 1 being connected to a correlator of each receiver 2 and each receiver 2 being connected to a correlator of each transmitter 1. The individual results of the correlation unit 17 are stored individually in first memories 7-1-1 to 7-es. In Fig. 7 the way is

Data processing shown. The time series and the locations of the reflections are determined in the following calculation unit 18 for the parameters of the illustration of the envelope). These results are stored in second memories 9-1-1 to 9-e-s. The results stored in the second memories 9-1-1 to 9-e-s represent the values of the individual stripe images 12, which sum up to give the gray image 13, which can be displayed on a visualization device (not shown). In this exemplary embodiment, the correlator unit 17 is preceded by an A / D converter 5-1 to 5-e. However, it is also possible to connect the A / D converters 5-1 to 5-e downstream of the correlator unit 17. All functions of the arrangement for image reproduction are controlled by a control unit 11.

2 shows an illustration for explaining the reconstruction of gray images 13. In order to construct stripe images 12 from the values contained in memories 9-1-1 to 9-es, the values contained in memories 9-1-1 to 9-es become mapped onto a memory array by the mapping rule derived from Eq. The non-zero values are clearly mapped onto an array of ellipses, depending on the position of the sensors belonging to this recording. The spatial orientation of the ellipse family is dependent on the known position of the transmitter 1 and the receiver 2 relative to one another. This mapping is done for each transmitter-receiver combination under consideration. The gray-scale image 13 is then obtained by summing the individual stripe images 12. The two-dimensional representation of the measurement principle for data acquisition by means of ultrasound is shown in FIG. 6. The arrangement of ultrasound components shown in FIG. 6 provides angle-dependent projections of the measurement object analogous to the measurement system described in FIG. 3. As has already been described, an ultrasound transmitter emits a signal in all directions without direction. It can be both a frequency-modulated signal and a pulse. Assuming the principle of supeposition for mechanical waves, all potential object points 3, which make a constructive contribution to the reflected energy at a certain point in time, have the same distance from the ultrasound transmitter and from the ultrasound receiver. The object points 3 thus determined are arranged on an ellipse arch around Ei and S _t . Since the distance between transmitter 1 and receiver 2 is known, the extent of the ellipse arc depends only on the distance between the transmitter-reflector and receiver and thus only on the corresponding signal propagation time. The time series recorded on the receiver Ei corresponds to the projection P with respect to an angle θ, where θ is the angle between the normal to the distance between the transmitter and receiver and the x-axis of the fixed coordinate system. The angle θ is varied by the combination of transmitters 1 and receivers 2, although the ultrasound elements are in a static arrangement. The variation of θ in the range 0 ° to 180 ° finally gives the analogue to the radon transform in FIG. 4 B. Simulations have shown that the arrangement of the transmitters 1 and receivers 2 on a semicircle (hemisphere) for the basic functioning of the reconstruction process is not absolutely necessary.

The underlying signal processing is explained below and corresponding references to the reconstruction method are given. In the principle illustrated in FIG. 6, an ultrasound transmitter is combined with an ultrasound receiver. The signal (FM signal, or pulse) propagates from the transmitter in all spatial directions. In parallel, all receivers 2 are recorded, so that all possible transmitter-receiver combinations for the

Reconstruction algorithm with respect to the active transmitter 1 are available. This procedure is repeated in sequence for all transmitters 1. In this way, are available Reconstruction of all possible combinations of transmitters 1 and 2 receivers available. Based on the signal recorded at the receiver, the signal processing path is shown below with reference to FIG. 7. The procedure can be applied to all measured time series, including the three-dimensional case.

7 shows the way of signal processing. After the data acquisition 16 takes place in the correlator unit 17 with the associated FM codes 6. The envelope 18 is calculated by bandpass filtering, squaring the signal and subsequent lowpass filtering with appropriately designed digital zero phase filters. The windowing in frequency range 19 already described is carried out with one of the following window functions: 'Hamming', 'Hanning', 'Shepp-Logan', and 'Cosinus'.

For the reconstruction, the mathematical background for the two-dimensional case, i.e. the reconstruction algorithm. As already mentioned, the algorithm can easily be extended to three dimensions. The illustration shown in FIG. 8 is intended to serve as an explanation. 8 shows a world coordinate system with respect to which the positions of all ultrasound components are defined. The area in which object 3 is located is marked with a rectangle. It defines the subset of the world coordinate system in the bordered area, which represents the result of the reconstruction as a gray scale image 13. As already mentioned above, the reflection locations whose backscattered energy add up at the same times in the time series recorded on the receivers are on ellipses with the corresponding transmitters and receivers in a focal point in each case. The parameters of the rotated ellipse shown in FIG. 8 are uniquely determined by the distance d and the time it takes for the sound to travel from S to a point of the ellipse to E.

The described mapping of the time series on the corresponding ellipse arcs gives a so-called stripe image 12 with respect to the individual transmitters 1-1 to 1-s and the respective receiver 2-1 to 2-e. The number of transmitter-receiver combinations used results in the number of resulting stripe images 12. 9 shows an index matrix for mapping the time series mentioned onto a strip image 12. An index matrix must be calculated for each transmitter-receiver combination. The base area of the index matrix shown in FIG. 9 is identical to the elements (pixels / voxels) of the object area. This assigns an index value that refers to the entry in the time series to be displayed. The intersection of the index matrix shown above with the plane parallel to the object plane always results in an ellipse. For the three-dimensional case, a three-dimensional matrix is obtained and, instead of the ellipses, rotationally symmetrical ellipsoids.

Starting from the basic equation of the ellipse in parameter form, the equation of the rotated and shifted ellipse is derived as described above. On the basis of this solution, the signal propagation time can be calculated from each transmitter 1 to every point of the object area 3. The result is the index matrix described above.

The complete reconstruction of a gray scale image 13 is obtained by adding the individual stripe images 12. The gray scale value of each pixel thus corresponds to the integral of the angles reflected in the energy and different angles.

Claims

claims

1. Arrangement for image reproduction for computer tomographs with ultrasound consisting of any number but at least two transmitters (1) and any number but at least two receivers (2) known in

Positions to one another are arranged around the object, so that the object (3) is sonicated and the response signals are received, the transmission signals of the individual transmitters having different modulation functions which do not influence one another, characterized in that the receivers in succession a) have an A / D Converter, b) a correlator unit (17) in which each received signal is correlated with each transmitted signal; c) a calculation unit (18) for the parameters of the figure, with a calculation part for the envelope, a subsequent calculation part for the

Index matrix (12) and at least one discrete integration unit (13); and d) a display device is connected downstream.

2. Arrangement for image reproduction for computer tomographs with ultrasound according to claim 1, characterized in that after the calculation part for the envelope a window in the frequency range (19) is provided.