DE202019003376U1 - X-ray system for iteratively determining an optimal coordinate transformation between overlapping volumes reconstructed from volume data sets of discretely scanned object areas - Google Patents

Abstract

X-ray system, for the iterative determination of an optimal coordinate transformation between a first N-dimensional partial volume V1 and a second N-dimensional partial volume V2 of an extended object, these partial volumes partially overlapping in the overlap volumes OV1 and UOV2, and wherein the X-ray system is suitable
a) to record partial volumes V1 and V2 by discrete scanning of the expanded object,
b) applying a band-pass filter to the overlap volumes OV1 and UOV2 within each iteration step i which removes from the overlap volume OV1 (i) those frequencies which are not included in the overlap volume UOV2 (i) and which is from the overlap volume UOV2 (i) removes those frequencies which are not contained in the overlap volume OV1 (i), whereby filtered overlap volumes FOV1 (i) and FUOV2 (i) are obtained,
c) To determine a coordinate transformation by iteratively optimizing a similarity measure between the filtered overlap volumes FOV1 (i) and FUOV2 (i).

Description

X-ray system for the iterative determination of an optimal coordinate transformation between overlapping volumes reconstructed from volume data sets of discretely scanned object areas.

Known image recording devices for imaging or measuring objects are always limited by a measuring method applied to a measuring or image field within which they scan an object and possibly represent. Outside of such a measurement field, there is by definition no information about the object. If the measuring field is too small for a selected application, it is known to extend this by making several shots. The totality of the measuring fields preferably form the object in such a way that their edges are at least adjacent to each other, in general even overlapping. It is known that a combination of error-and artifact-poor combination of individual measuring fields to an entire measuring field on precise knowledge of the geometries and measurement methods of the image recording devices is made possible by coordinate transformations between the measuring fields are determined. Furthermore, it is known that in the absence of such knowledge, the coordinate transformations can be obtained from the comparison of corresponding image regions in the overlapping regions of the individual measurement fields. In this case, an initially assumed coordinate transformation is preferably improved in an iterative optimization method until the method has reached a sufficient congruence of the two measuring fields in its overlapping region in accordance with a termination criterion. The degree of congruence is described by means of a quantitatively formulated similarity measure or measure of dissimilarity, which is correspondingly maximized or minimized, ie optimized, by the optimization method. Such methods are known in particular as inelastic or elastic registration. Since it is possible at any time to transform a similarity measure to be maximized into a measure of dissimilarity to be minimized, for example by multiplication by "-1", the term similarity measure will be used in the following to represent both optimization strategies.

In the general case, the measurement fields thereby correspond to volume data sets of a spatial region which has a continuum of physical values which is scanned discretely and possibly incompletely within the volume with the measuring apparatus, the volume being described, for example, by voxels at periodic halftone dots. The space can, but does not have to correspond to a physical space and thus can have one, two, three or more dimensions. Corresponding higher-dimensional volumes can be achieved, for example, by three spatial dimensions plus a time coordinate ( 4D) be characterized or, for example, by three spatial dimensions, which are detected in multiple spectral regions.

The sampling of the volume can be done directly or calculated from initially recorded intermediate steps. The latter is used, for example, in X-ray computed tomography, where from the acquisition of two-dimensional X-ray projections, for example by means of the filtered back-projection method, a three-dimensional representation of the absorption properties of the examined object, for example a patient, is calculated.

Scanned volumes always have a representation in Fourier space, which can generally be calculated by a discrete Fourier transform. However, even without their explicit calculation, there is a one-to-one correspondence between the representations in the space space and in the Fourier space. In particular, measurement errors or artefacts which manifest themselves in the detected volume, the so-called spatial space, can be attributed a direct correspondence in Fourier space without directly calculating the Fourier transformation. The coordinate axes in the Fourier space are represented by wavenumbers or frequencies along the corresponding spatial axes, and thus a coordinate in the Fourier space corresponds to a tuple at wavenumbers or frequencies along these axes (referred to below as frequencies). In general, such a Fourier coordinate has a complex number containing the amplitude and phase of a (space) frequency. If the corresponding frequency was not detected by the measuring method, this number is ideally equal to zero at the corresponding coordinate in the Fourier space. However, due to limited numerical precision, or downstream image processing, the actual complex number may be nonzero, but generally has a comparatively small absolute amount. In the following, in this case, a "missing" or "not sampled", "non-existent" or "unrecorded" frequency shall be used. Likewise, the "removal" of frequencies is said to be the case if the complex numbers collected on them are set to zero or a very small number.

It is known that artifacts can occur in the reconstruction of the resulting volume due to the scanning. A class of such artefacts manifests itself by the fact that in the Fourier space, ie a Fourier transform of the Scanned volume, one or more contiguous areas exist that are approximately zero and thus missing, but this is not associated with an object property. In such a case one speaks of artifacts by an incomplete scan. The expression of these artifacts in the image or spatial space generally depends on their location in the volume. Thus, when registering individual volumes, the situation that different image areas differ in their artifacts occurs. The application of a similarity measure is thus error-prone, since identical object areas in the discrete scan are similar only to a certain extent and possibly insufficient. As a consequence, the registration has errors that are reflected in the accuracy of the determined coordinate transformation. Thus, significant artifacts are expected after the individual volumes are joined together.

In known cone beam computer tomographs, an X-ray image recording unit rotates from an X-ray emitter and a detector, preferably a flat detector, taking a series of X-ray projection images in a scan plane around a scanning center. Such cone beam computer tomographs can be designed as gantry-based systems with the possibility of continuous rotation, but in particular include stationary or mobile C-arms, as well as additional systems for position detection in radiotherapy or micro-computed tomography in preclinical research or in material science. From the recorded X-ray projections of a scan, an X-ray volume can be reconstructed. If the region of interest (ROI) in an object is small compared to the reconstructed volume, the diagnostic volume information is obtained with a single scan. For certain diagnostic tasks, however, a first volume reconstructed from a single scan to produce a 3D data set is too small. In order to increase a reconstructed volume, a 3D data set is generated by means of at least one further scan, the reconstruction of which results in a second volume that differs from the first volume, wherein edge regions of the first and second volumes overlap. Such a method can be extended by attaching additional volumes to the respective last reconstructed volume to generate even larger volumes. If the ROI is large compared to a reconstructed volume, it is known to record projection data of the entire ROI and to reconstruct a composite volume by recording further scans with offset scan lines.

Examples of use are the generation of the volume of an entire spinal column, or the bolus tracking over the length of an entire leg, or the acquisition of a whole body volume. If the object or its storage and the cone-beam computed tomography are referenced to each other, the coordinate transformation between the two scan planes is known for displacement of the scan plane for the purpose of recording a further X-ray projection data set. This is usually the case with stationary CT devices.

However, when using a mobile cone beam computer tomograph, such as a mobile C-arm, after positioning the C-arm with respect to the object on a bearing, referencing within a calibration run is necessary, and the referencing becomes invalid if the C- Sheet is relocated non-reproducibly.

If a coordinate transformation between the coordinate system of the cone beam computer tomograph and the coordinate system of the object and, if the object during the X-ray examination in a fixed spatial relationship to the storage, for example, a patient bed, positioned to the coordinate system of the storage is not known, then a coordinate transformation in be determined in the above manner such that the greatest possible agreement of the absorption values of the voxels in the corresponding overlap areas.

It is known that cone-beam computer tomographs generate so-called cone beam artifacts based on an incompletely sampled Fourier space. The strength of these artifacts is greater, the greater the distance of a volume region from the central layer or the central scan plane. In the Fourier space, a structure similar to a double cone with an opening angle remains, within which no measured values, in particular frequencies, are detected. Furthermore, it is known that a computed tomography, which can not be completely traversed for example due to imminent collisions with the storage, also does not scan certain areas in the Fourier space due to lack of angular coverage. The known algorithms for registering image volumes do not take into account the presence of such artifacts, therefore their application leads to registration errors and the total volume composed of sub-volumes has visible transitions between the two individual volumes.

The novel X-ray system requires the provision of two overlapping ones Volumes and an estimated coordinate transformation between the position of the two volumes.

A requirement in the image recording of the sub-volumes may moreover be that the overlapping area of the first and second volumes is as small as possible, for example to keep a radiation dose applied to the patient as low as possible, or the time required for recording the total volume to be as small as possible hold; On the other hand, the overlap must not be too small in order to be able to calculate the relative position and orientation of the first to the second volume during the registration, with the aim of being able to join the individual volumes as artifact-free as possible.

If the volumes of the examination object and of the image recording device are not registered with one another, overlapping individual volumes recorded one after the other can be registered one after the other in the described and known manner, and then combined by means of a software algorithm to form a total volume, which, however, contains the errors mentioned in the form may have visible artifacts.

Apparatus and methods for obtaining composite volumes from individual volumes are known in the art.

From the document DE10014668A1 is a method for minimizing Cone-Beam artifacts by supplementing missing frequencies in Fourier space known.

From the document DE102007025862A1 For example, a system and method are known for geometry-based registration of sub-volumes and for forming a composite volume with the goal of avoiding artifacts.

From the document DE102007034218B4 For example, there is known an apparatus and method for composing an image consisting of at least two frames, a tagged object, using optical images for registration.

From the document DE102008062030A1 For example, an arrangement and a method for positioning objects to be imaged and 3D stitching are known, wherein a common control of C-arm and patient couch is provided by using position sensors.

From the document EP520778B1 is a method for tomographic imaging by means of oblique rays known to be artifacts are eliminated, which are caused by stitching of two sets of scans, wherein the plane of a scan is not perpendicular to the z-axis.

From the document Seungeon Kim, Yongjin Chang, and Jong Beom Ra: "Reduction of Cone Angle Effect in Cardiac Motion Correction Based on Partia / Angle Reconstructed Images in CT", 4th International Conference on Image Formation in X-Ray Computed Tomography, Bamberg, Germany 18.-22. July 2016, Proceedings of the 4th CT Meeting, page 439 ff., Is known to approximate Cone Beam artifacts for motion compensation in cardiac CT.

From the document US20140267267A1 For example, an apparatus and a method for 3D stitching of overlapping individual volumes are known, wherein the stitching is based on the voxel values in the overlap volumes.

From the document US20150371420A1 For example, a method is known for registering partial volumes and assembling the partial volumes into a composite volume using artifact positions in the X-ray volume.

From the document US6711293B1 For example, a method for composing subvolumes is known in which extreme pixel values of subtraction images are used for registration.

From the document US9129389B2 For example, a CT apparatus and a method are known in which the contrast differences between temporally successive recorded CT disc volumes are corrected.

From the document WO11039672A1 a method for 3D stitching of partial volumes is known in which image data of the individual volumes are used for the registration.

From the document WO12048070A1 For example, a method for 3D stitching of overlapping partial volumes is known in which an alpha blending / morphing method is provided in the overlap volumes and voxels in the individual volumes prior to stitching are weighted, the weight being the distance of the voxel from the edge of the single volume depends.

From the document WO12139031A is a method for generating a composite volume using Cone-Beam-CT known in the Cone-beam artifacts are used as a measure of inaccurate Volumenstitching.

From the document WO16097174A1 For example, a method for 3D stitching using the isocenter positions of the individual volumes is known.

There is a need for a method of obtaining an improved composite volume of images whose Fourier spaces are incompletely scanned. There continues to be a need for an improved markerless registration of, in particular, low-contrast computed tomography scans to a total 3D volume as well as an X-ray diagnostic system for carrying out markerless registration.

The object is to provide a device which provides an algorithm for determining a coordinate transformation, for example, with which a second of two overlapping volumes is registered to a first volume and with which a second volume is attached in the correct position to a first volume, wherein the Disadvantages of incompletely sampled Fourier spaces eliminated and the accuracy of the registration of individual volumes is increased.

Furthermore, it is the object of the innovation to provide an X-ray diagnostic system of a controllable cone beam computer tomography and a controllable motor-adjustable storage, which allows the recording of overlapping volume data sets.

The object is solved by the features of the independent claims. Further, particularly advantageous embodiments and developments emerge from the dependent claims.

With regard to the device, the object is achieved by the features of claim 1.

The object is achieved by iteratively determining, as part of the registration, starting from an initial, estimated coordinate transformation, a coordinate transformation in which the degree of correspondence of modified or filtered voxel values of the first and second overlap volumes reaches a maximum value.

There will be larger scan volumes, especially along the long axis of storage, e.g. along the spine or lower limb of a patient. For example, scolioses and their correction within a surgical procedure can be better represented in this way.

Preferably, the steps for registering and assembling individual volumes into a total volume are implemented by software. A largely software-based implementation has the advantage that also previously used volume reconstructions can be retrofitted with a composition of individual volumes to a total volume for image acquisition systems in a simple way by a software update. In this respect, the object is also achieved by a corresponding computer program product with a computer program that can be loaded directly into a memory device of an image recording system, for example a cone beam computer tomograph, with program sections in order to execute all steps of the coordinate transformation when the computer program is executed in the control device. Such a computer program product, in addition to the computer program optionally additional components such as documentation and / or additional components, including hardware components for using the software include.

For transport to the control device and / or for storage on or in the control device, a computer-readable medium, for example a memory stick, a hard disk or another portable or permanently installed data carrier can serve, on which the program sections of the computer program which can be read and executed by a computer unit of the control device are stored are. The transport can also be used to connect to a hospital information system connected to a network, to a radiology information system or to a global network, in which systems the program sections of the computer program that are readable and executable by a computer unit of the control device are stored. The computer unit may e.g. for this purpose have one or more cooperating microprocessors or the like. Here please remove the forced page break or whatever is present

• • die Uberlappungsvolumina in den Fourierraum transformiert,
• • in den Fourierräumen alle Frequenzen, die in der Fourierrepräsentation des jeweils anderen Überlappungsvolumens nicht auftreten, aus den jeweiligen Fourierräumen entfernt, beispielsweise durch Multiplikation mit 0, und so zwei gefilterte Fourierräume der Überlappungsvolumina erzeugt,

oder
b)

• • im Ortsraum auf die Überlappungsvolumina wirkt,
• • der in seiner Wirkung ebenfalls alle Frequenzen, die in den entsprechenden Fourierräumen des jeweils anderen Überlappungsvolumens nicht auftreten aus den jeweiligen entsprechenden Fourierräumen entfernt und so zwei gefilterte Versionen der Überlappungsvolumina erzeugt,

wobei ein Ähnlichkeitsmaß der gefilterten Überlappungsvolumina bestimmt wird und wobei die Koordinatentransformation, die das optimale Ähnlichkeitsmaß der gefilterten Überlappungsvolumina erzeugt, als die optimale Koordinatentransformation für die Registrierung verwendet wird. Die nicht erfassten Frequenzen können dabei direkt im Fourierraum dadurch identifiziert werden, dass der an ihnen vorliegende Absolutbetrag kleiner als ein zu wählender Schwellwert ist. Alternativ können die unabgetasteten Bereiche bei Vorliegen entsprechender Informationen aus der Kenntnis der Aufnahmegeometrie bzw. Abtastart/Abtasttrajektorie und der speziellen Ausgestaltung eines Messverfahrens errechnet werden.The object of the innovation is achieved, in particular, by receiving two overlapping volumes of an object area together with an estimated registration information in the form of a coordinate transformation and, on the basis of the initial estimated coordinate transformation, by iterative change of the coordinate transformation, resulting overlapping volumes are calculated, which are filtered by an algorithm. which adjusts the artifacts in the overlap volumes and which includes a bandpass filter which
α)

• Transformed the overlap volumes into the Fourier space,
• • in the Fourier spaces all frequencies that do not occur in the Fourier representation of the other overlap volume, respectively, from the respective Fourier spaces removed, for example by multiplication by 0, and so produces two filtered Fourier spaces of the overlap volumes,

or
b)

• Acts in space on the overlap volumes,
• • which also removes in its effect all the frequencies that do not occur in the corresponding Fourier spaces of the other overlapping volume from the respective corresponding Fourier spaces and thus produces two filtered versions of the overlap volumes,

wherein a similarity measure of the filtered overlap volumes is determined, and wherein the coordinate transformation that produces the optimal similarity measure of the filtered overlap volumes is used as the optimal coordinate transformation for the registration. The unrecorded frequencies can be identified directly in the Fourier space in that the absolute value present at them is smaller than a threshold value to be selected. Alternatively, the unsampled areas can be calculated from the knowledge of the recording geometry or scanning type / scanning trajectory and the specific configuration of a measuring method if appropriate information is available.

The overlap volumes of the two individual volumes are thus each subjected to filtering for the approximation of artifacts, the artifacts corresponding to an incompletely scanned Fourier space of at least one subregion in at least one subvolume. For example, in the case of computed tomography, these may originate from the group of cone-beam artifacts or artifacts from incomplete angular coverage (scan angle range less than, for example, 180 degrees plus cone angle).

The result of the iterative determination of a coordinate transformation with an estimate of the coordinate transformation for the first iteration step is a comparatively improved coordinate transformation, which is used for the referencing of the individual volumes in the spatial space. The individual volumes are shifted with the determined optimal coordinate transformation and rotated (inelastic registration, ie globally valid coordinate transformation), or additionally deformed (elastic registration, ie local or voxelweise coordinate transformation), for example, a first recorded volume data set is connected to an intrinsic coordinate system, and the second or all further volume data sets are registered with respect to the intrinsic coordinate system. After registration of the two volumes, the voxel values of the overlap regions are joined together in a known manner to avoid the generation of artifacts at the transitions between the two volumes. At this time, the central or residual volumes outside the overlap volumes in the voxel's absorbance values remain unchanged, except for the effects of interpolation steps required to image both volumes on the same discrete grid. When joining the individual volumes, however, redundant image information in the overlapping area is used in such a way that the highest possible image quality of the reconstructed total volume is achieved in order, for example, to optimally use an applied radiation dose.

The algorithm for evaluating the overlap areas of a first and a second volume has as a result, to determine the relative position and orientation of the two volumes to each other. The algorithm uses information about an assumed or estimated relative position and orientation of the individual volumes. These assumed values for the position and orientation of the individual volumes are obtained, for example, from the travel path and the travel direction of the cone beam computer tomograph, preferably a C-arm X-ray device, or the storage of the object, for example a patient couch, in space.

It is envisaged to use the determined optimal registration information to establish registration between the cone-beam computed tomography, preferably a C-arm, and the coordinate system of storage. Such a registration enables navigation in the volume composed of the individual volumes.

In order to establish a registration or referencing between the coordinate system of a cone-beam computed tomography and the coordinate system of storage, an extended phantom with radiopaque structures may preferably be used as the object instead of a patient, and the registration step may take place prior to an X-ray examination. If the bearing and the cone-beam computed tomography are referenced to one another, this referencing remains valid until the cone beam computer tomograph and / or the bearing of the object are displaced in a non-reproducible manner. As long as the registration is valid, for example, storage in a well-defined location in the Coordinate system of the cone beam computer tomograph procedure.

It is also envisaged that the reference coordinate system will not pass through the first volume VI but through the second volume V2 set. Alternatively, it is provided to use a third, for example, external or space-fixed coordinate system, to which the two volumes V1 and V2 be mapped accordingly. The above steps change accordingly.

The combination of the two individual volumes V1 and V2 takes place in the location space using the one with the optimum similarity measure S (i) connected coordinate transformation T12 (i) ,

To avoid artifacts in the spatial space is provided at the transitions between in both volumes V1 and V2 present overlap areas and only in a single volume V1 or V2 identify the contributions of each individual volume taking into account a transitional function that avoids hard transitions; eg by means of a sigmoide instead of a rectangle.

It is optional to increase the sub-pixel accuracy by means of interpolation, ie by scaling up the volumes, subsequent registration and subsequent scaling down of the image resolution.

To generate a total volume composed of two individual volumes, it is necessary to suitably combine those image regions which are present in both individual volumes. It is intended to arithmetically average the original gray values of the voxels, thereby preserving the cone beam artifacts.

It is therefore to be regarded as advantageous to use in each case the voxel from that individual volume, which each has less cone beam artifacts. It is further provided to weight the voxels from both volumes in a weighted manner, wherein the weighting factors are determined from the respectively existing artifact level, for example corresponding to the cone beam angle associated with a voxel or a cone angle averaged over the scan area in a non-circular focus trajectory a scan. It is further provided to determine the weighting factors about the measurement error of the discrete sample on a voxel. For example, the recording situation may have led to one of the two sub-volumes having a higher noise in subregions, so that a corresponding voxel is weighted less heavily in the combination.

It is further provided after the registration, but before the combination of the sub-volumes, first separately for each overlap volume to fill the missing areas or frequencies in one of the Fourrierräume with the corresponding areas or frequencies from the respective other Fourier space, if and insofar as they exist there are to minimize the artifacts in the total volume to be combined as best as possible.

It is intended, with a slight movement-related displacement of parts of an object between the individual recordings of the scans of the 3D data records, to carry out an elastic registration in addition to a rigid registration of the two volumes.

It is envisaged to use the acquired registration information for calibration between cone-beam computed tomography and storage. The calibration / registration remains valid as long as the displacements of the cone beam computer tomographs and / or storage are reproducible. The obtained calibration data can also be used, for example, for collision avoidance.

It is envisaged to adjust the cone-beam computer tomograph, in particular a mobile C-arm, by adjusting a so-called Wigwag angle about a pivot axis which is perpendicular in space and contains the scan plane of the first scan. In this case, the angle orientation of the receiving unit is changed in a stationary drive stand. This is done by means of a motor drive or manually by the user, for example by means of a reference marker, or an indication of a position encoder reading this angle.

If the measuring field is changed in this way, this corresponds to a rotation instead of a translation of the recording unit. Around the center of the measuring field, e.g. to move back to the spine can be used on the cone beam computer tomography, especially on a C-arm existing linear axis. This requires an additionally available travel path on the linear axis, since it may also be required during recording to maintain a virtual isocenter.

A change in the orientation of the receiving unit may exacerbate collision situations. Performing additional collision checks is therefore particularly important in this case. As above, these should, if possible, and especially in the presence of a corresponding motorization of the deflection, be carried out completely before the start of the first scan, for example to shorten the time, during which the breathing of a patient is stopped.

It is envisaged that a user manually triggers the individual steps via an input interface, for example by pressing a button or an input area on a touchscreen / GUI.

It is envisaged to move the entire cone-beam computer tomograph under motor control, for example in the case of ceiling mounting of a cone beam computer tomograph, in particular a C-arm.

In a mobile chassis, it is intended that the user place the wheels in a 90 degree position prior to the start of a scan, or the cone beam computer tomograph will do so on its own. Preferably, this is verified via a position sensor for the alignment of the wheels of the device software. The user ensures that sufficient travel is available for all positions. The initial position of the cone beam computer tomograph must be selected so that all other positions can be reached. The display of the C-arm then gives instructions as to how far the C-arm, and possibly in which direction it is to be moved. For example, for a single volume size of 16 cm and an overlap of 2 cm, a motion of 14 cm would be requested.

In the case of motorized mobile chassis, the cone beam computer tomograph may preferentially perform the required movement using an existing position sensor. Preferably, this motorized movement is executed only when the user touches, for example, a button or an area on a touch screen / GUI in order to be able to immediately stop the process in the event of an imminent collision.

In this case, for the iterative determination of the coordinate transformation in the manner described above, the overlapping areas assumed differently within each iteration step are filtered in such a way that only the areas present in both Fourier spaces of the two subvolumes are used to determine the similarity measure.

A similarity operator is then applied to the two filtered overlap volumes in spatial space or Fourier space by iteratively changing the coordinate transformation and a similarity measure at each iteration step i S (i) is determined. The coordinate transformation in which the value of S (i) is optimal, as optimal coordinate transformation to register the volumes V1 and V2 used.

It is expedient to store the determined coordinate transformation as referencing information between the cone beam computer tomograph and the object storage. The storage of the coordinate transformation, or referencing information, can be used in this embodiment to directly combine partial volumes recorded in the future without further optimization.

In the case of the use of a cone-beam computed tomography, for example a C-arm, the relative position and / or orientation change of the cone-beam computer tomograph relative to the ROI takes place optionally by manual and / or motor-driven displacement of the cone-beam computer tomograph and / or storage, preferably a patient bed, with the examination object. The change in position and / or orientation between the recording of the first 3D data record and the recording of the second 3D data record is carried out such that a required overlap of the partial volumes is given.

In 1 is a C-arm X-ray system 1 shown schematically, which is intended for the recording of projection shots of a scan.

A C-arm 2 carries at its one end the X-ray tube 3 , the collimator 5 and at least parts of the high voltage generator 4 and at the other end and the X-ray tubes 3 opposite to an X-ray image detector 6 , in particular a flat panel detector FPD. The C-arm 2 is motor-controlled adjustable in several axes in space, the axes have sensors for detecting the degree of adjustment. In particular, the C-arm 2 along its circumference in a holder guided by a motor. This movement is called orbital motion / orbital axis. The C-arm holder, not shown, is motor-displaceable in at least one horizontal and one vertical axis. A C-arm X-ray system equipped with these three motorized axles 1 is set up a scan, with a scan plane 51 with a series of 2D X-ray projection images for a subsequent reconstruction of an X-ray volume with plane trajectories of the focal point 8th around a scan center 20 to realize.

An x-ray tube 3 generates an X-ray focus and is preferably designed as a rotary anode X-ray tubes.

An oil filled high voltage generator 4 represents all for the operation of the X-ray tube 3 required parameters such as accelerating voltage, tube current, cathode heating current, rotating anode motor current, pulse frequency and pulse width available.

Preferably, there is between the x-ray focus and the object 21 a controllable collimator 5 provided for insertion of the outgoing from the focus X-ray beam. It is intended to change the aperture scan angle dependent.

At the arcuate C-arm 2 is an X-ray image detector 6 , in particular a flat panel detector "FPD" at a certain distance and in a specific orientation to the focus and the C-arm 2 arranged. It is envisaged the distance and the orientation of the X-ray image detector 6 reproducibly and controllably variable.

From the read image data of the X-ray image detector 6 in a projection image is in a projection image generation unit 7 generates a raw image of a projection of the object and an image processing and storage unit 106 fed to.

A scan center 20 represents the center of the region of interest (ROI) to be reconstructed 50 inside the object 21 represents.

As a central beam 22 becomes the connecting line between the X-ray focus and the center of the input window of the X-ray image detector 6 or another X-ray receiver. It lies in the orbital plane of the C-arm perpendicular to the entrance window of the X-ray image detector 6 ,

An object to be examined 21 is on a storage 23 , preferably a patient bed, placed. Warehousing 23 may be designed to be after initial alignment with the C-arm X-ray system 1 , during the scanning of the object 21 , remains rigid. It is further provided the storage 23 equipped with motorized, adjustable axles and the movement of storage 23 by means of a storage engine control unit 25 synchronous with the movement of the C-arm X-ray system 1 to control, or a specific position and orientation of storage 23 adjust. It is provided the motion control unit 103 to record a series of scans to produce a compound volume so that the object 21 concerning the C-arm X-ray system 1 is initially positioned so that the scan center and the scan plane according to the task set, for example, in the center of a first ROI is positioned. In this orientation, a scan is recorded, for example, with a series of X-ray projection images and a volume is reconstructed from the projection data set whose coordinate system is that of the C-arm X-ray system 1 equivalent. For example, for a next one of a series of scans, the object becomes 21 along the scan axis passing through the first scan center and perpendicular to the first scan plane, traversing a distance equal to, for example, three quarters of the extent of the scan scan ROI in the direction of the scan axis. The C-arm X-ray system 1 remains spatially unchanged with its scanning center and its orientation. By sensors on the adjustment axes of storage 23 can through the motion control unit 103 the position and orientation of the object 21 be approached reproducible. In the second scan position, another data set is now stored in the fixed-space coordinate system of the C-arm X-ray system 1 recorded and reconstructed the volume. In the event that C-arm X-ray system 1 and storage 23 are referenced to each other, the measured values of the displacement of the storage 23 be used between the first and the second scan to a regulation for the displacement of the second volume V2 in the coordinate system of the fixed C-arm X-ray system 1 in the form of a coordinate transformation T to investigate. By applying the coordinate transformation to the second volume V2 If this is so shifted and rotated and possibly deformed that corresponding object areas in the two overlapping areas of the two volumes ( V1 . V2 ) are represented by the same voxels. The voxels of the overlap areas are eventually combined and averaged in a known manner to achieve a smooth transition between the two volumes.

In the event that the C-arm X-ray system 1 and storage 23 are not referenced to each other, the two volumes can be shifted and / or tilted when applying the coordinate transformation against each other, so that corresponding object areas in the overlapping areas come to rest at different voxel positions.

To the two volumes ( V1 . V2 ) in the correct orientation to one another, an optimal match of the corresponding object areas in the overlapping areas can be calculated in the manner described. The coordinate transformation determined in this way is then applied to the second volume and the volumes then oriented in the correct position are then connected to one another in a known manner, for which the terms "sewn" or "gesticht" are also used synonymously.

For a displacement of the scan plane and the scan center by means of the motor movement of the bearing 23 is through the motion control unit 103 a communication between C-arm X-ray system 1 and storage 23 produced. Before the start of the first scan, the C-arm X-ray system 1 and storage 23 brought into predetermined starting positions.

If the storage 23 is ready for the acquisition of a first projection data set, the C-arm X-ray system 1 precisely aligned. Thereafter, optional collision checks are carried out, if necessary for each position of the storage to be approached 23 , To this process, especially when using a C-arm X-ray system 1 To accelerate, the collision checks can be performed in both directions, eg for a first single volume in the forward direction and for a second volume in the reverse direction.

In the case of application to a patient, before starting a scan, if necessary, the patient's respiration is stopped to avoid the occurrence of motion artifacts. Preferably, then the recording protocol for all individual volumes without interruption to perform. So a first scan is taken, then storage 23 then start recording the second scan.

In the event that neither the C-arm X-ray system 1 still the storage 23 are motor-adjustable, but have position sensor in the adjustment axes, a user interface / GUI / MMI is provided by means of which the user is asked to perform the mechanical adjustment steps, the correct execution is monitored (for example, whether the storage 23 in the correct position, taking into account a predetermined tolerance). In the additional absence of an interface eliminates the need to verify the correct execution by the C-arm X-ray system 1 ,

In the presence of a motorization, but absence of a position sensor, the C-arm X-ray system can 1 be calibrated on the or the motors of the patient bed.

The C-arm X-ray system 1 has a volume comparison unit for performing the similarity determination 102 in which the in a Fourier transform and filtering unit 118 generated frequency representations of the two volumes are compared with the goal to determine an optimal registration information or coordinate transformation between the two volumes.

The volume comparison unit 102 is configured to associate a first volume of a portion of an object with a spatially fixed first coordinate system and a second volume of another portion of the object that overlaps the first portion of the object with a second coordinate system and an estimated coordinate transformation between the first and second coordinate systems received, wherein the second coordinate system approximately by the estimated coordinate transformation T12 in the first coordinate system K1 is transferred. In the case of the X-ray system, the two volumes represent the filtered X-ray volumes of the overlap volumes determined on the basis of the estimated coordinate transformation. For this purpose, the reconstructed X-ray volumes of the Fourier transformation and filtering unit 118 provided, each of which performs a Fourier transform and a filtering of the data in the frequency space. The Fourier transformed and optionally filtered data are called Fourier volumes below.

The optimal coordinate transformation determined in the iteration process is provided by the volume comparison unit 102 output. The output coordinate transformation T12 is from the reconstruction unit 117 receive the second single volume V2 opposite to the volume V1 relocate such that a combination volume results, that of the first volume V1 and the second volume V2 exists, and wherein the two volumes are combined with each other in the correct position.

A C-arm X-ray system engine control unit 24 controls all motor-adjustable axes of the C-arm X-ray system 1 and represents the motion control unit 103 Measured values of position sensors of non-motorized adjustable axes are available.

A storage engine control unit 25 and a C-arm X-ray system engine control unit 24 be from the motion control unit 103 of the C-arm X-ray system 1 controlled synchronously.

A higher-level system control unit 100 preferably has a bus system with a plurality of control, arithmetic and memory components.

An X-ray source control unit 101 controls the high voltage generator 4 , the x-ray tube 3 and the collimator 5 , It has a dose rate control or Automatic Exposure Control AEC, which in an X-ray image recording the parameters of the X-ray tube 3 such controls that with the X-ray detector 6 an X-ray image of a given quality, for example, with a predetermined signal / noise ratio is generated.

C-arm X-ray system 1 has a motion control unit 103 in particular, when executed as a C-arm in the scan parameter generating device 107 received focus trajectory and the location of the scanning center with respect to the focus trajectory receives.

An image capture control unit 104 triggers the recording of a projection recording, unless a prevention signal prevents the triggering of the recording. A prevention signal is generated, in particular in the case of a C-arm, for example, by a collision detected immediately before or by an emergency stop of the X-ray system initiated by an operator. After triggering a projection recording, the control of the tube current, the tube voltage and / or the exposure time in an exposure control of the X-ray source control unit 101 applied a dose in the X-ray detector 6 produces a projection image of a selected quality. As a measure of the image quality, for example, the signal-to-noise ratio can be specified.

An X-ray detector control unit 105 controls and controls all functions of the X-ray detector 6 such as image acquisition, image reading, dark current and gain correction and removal of residual charges after the end of the exposure time.

In an image processing and storage unit 106 The raw image data are extracted from the projection image generation unit 7 corrected, filtered or edited according to given image processing algorithms and stored. They then stand the reconstruction unit 117 to perform a volume reconstruction.

In a scan parameter generation unit 107 All parameters are generated for a scan. The scan parameter generation unit 107 receives, for example, the position of the scanning center 20 in the coordinate system of the C-arm X-ray system 1 , a focus and a detector trajectory, a total number of intended X-ray projection images of the scan and a proposed tube voltage for the scan. The generated scan parameters are in the motion control unit 103 , the X-ray control unit 101 and the image pickup control unit 104 and other units, such as a Fourier transform and filtering unit 118 made available.

The scan parameter generation unit 107 is set up to the position of a scanning center 20 in the coordinate system of the C-arm X-ray system 1 to recieve. The scan center can be adjusted, for example, by manual alignment of the central beam 22 to the scan center 20 determined in two independent spatial directions and entered by an operator / operator in the system control unit.

The scan parameter generation unit 107 is further set up trajectories of focus 8th and the X-ray detector 6 in the coordinate system of the C-arm X-ray system 1 to recieve. Especially when using a C-arm X-ray system 1 become the trajectories of an organ program database 115 after being selected by an operator and possibly adapted to the given examination situation, for example to the patient's dimensions.

It is at least a display 111 intended for the output of image, control and patient data. Via an input unit 112 become instructions of an operator / operator 113 in the system control unit 100 entered. It is envisaged that at least one display 111 as graphical user interface to design GUI.

An input device unit 112 is used to enter operating commands of the operator / an operator 113 and the import of system programs into the system, if they do not have a network 140 are importable.

The system control 100 has a mass storage 116 and an output unit 114 on. Under an output unit 114 each device is understood by means of the examination results and / or system parameters on any information carriers can be output.

The system control 100 has an organ program database 115 on, in the X-ray and scan parameters for different diagnostic tasks are stored retrievable. The individual organ programs are by an operator 113 adaptable to the existing circumstances, for example the patient dimensions and to the current diagnostic task. An organ program may, for example, represent the recording of a spine or a leg from the groin to the foot.

The system control unit 100 has a reconstruction unit 117 on, in which the processed and stored projection data from the image processing and storage unit 106 along with the corrected data for the Projection geometries of the respective projection shots from the motion control unit 103 are received and a 3D volume is reconstructed.

The system controller 100 preferably has a DICOM interface 130 to a network 140 on, by means of inputs and outputs between the C-arm X-ray system 1 and, for example, a hospital information system HIS 142 , a radiology information system RIS 141 and a global network. As DICOM input, the transfer of patient data and patient-specific scan parameters from previous examinations is provided. Has a cone beam computer tomograph, especially a C-arm X-ray system 1 In addition to the three controllable adjustment axes described above, another with an adjustment component perpendicular to the orbital plane, it is possible to realize non-planar focus trajectories.

The relationship between the position values of the drive motors and the pair of points on the detector and focus trajectories is theoretically based on the knowledge of the kinematics of the C-arm 2 calculable, but is known to be checked by means of a calibration run. Deviations from the theoretical kinematics are corrected as part of a calibration run and stored as calibrated kinematics, preferably in LUTs in a memory of the C-arm X-ray diagnostic system retrievable and taken into account in a volume reconstruction. For a volume reconstruction, the projection geometries of the calibrated kinematics are thus used for each projection image.

In 2 the design of the novel X-ray system for the iterative determination of an optimal registration information between two overlapping volumes is shown schematically.

First, a first spatial region of a total volume is discretely with a measuring device, for example a C-arm X-ray system 1 sampled and from the obtained volume data set a first volume V1 reconstructed. Subsequently, the measuring device is displaced relative to the total volume and a second space region overlapping the first spatial region is scanned discretely and a second volume is obtained from the obtained volume data set V2 reconstructed. The volume V1 is in a coordinate system K1 and the volume V2 in a coordinate system K2 shown. Through a coordinate transformation to be determined T12 becomes the coordinate system K2 in the coordinate system K1 transformed. Alternatively, by means of the inverse coordinate transformation T21 the coordinate system K1 in the coordinate system K2 transformed. To accelerate the convergence of the iteration in determining the optimal coordinate transformation T12 respectively. T21 an estimated coordinate transformation is specified. The estimate can be obtained on the basis of the approximate knowledge of the displacement of the measuring device; it is also intended to provide an estimate of the coordinate transformation from a first evaluation of any existing high-contrast landmarks in the two volumes V1 and V2 to win.

In the step S2 becomes the first volume V1 with the coordinate system K1 received and in step S3 the second volume V2 with the coordinate system K2 and an initial estimated coordinate transformation T12 receive. The initial coordinate transformation is used as start value T12 used for the first iteration step.

The reconstruction of individual volumes V1 and V2 is preferably carried out immediately after recording the respective volume data set. If more than two individual volumes are to be combined to form a total volume, the recording of further volume data records and the reconstruction of further sub-volumes can already take place during the iteration process.

If a total volume is to be composed of three or more sub-volumes, first of all, a first total volume is formed from two sub-volumes, to which the third sub-volume is attached. The first total volume in the iteration is like the first volume V1 and the third subvolume as the second volume V2 treated. It may be advantageous if as the first volume V1 a partial volume is selected in the interior of the total volume and the attachment of additional partial volumes takes place successively in two spatial directions. In general, when using a parallel computer, the determination of the coordinate transformations between the first volume V1 and take place parallel to this adjacent sub-volumes.

In step S4 in the iteration step i, the overlap volume OV1 (i) of the first volume V1 assuming the coordinate transformation T12 (i) in the coordinate system K1 determined.

In step S5 becomes the overlap volume OV2 (i) of the second volume V2 assuming the coordinate transformation T12 (i) in the coordinate system K1 determined.

The calculated overlap volumes OV1 (i) and OV2 (i) generally have a different raster structure of the voxels. The voxel patterns of the overlap volumes can be shifted by fractions of the raster period and the raster axes tilted relative to one another.

In step S6 become the voxels of the overlap volume OV2 (i) on voxels in the raster of the voxels of the overlap volume OV1 (i) converted by means of known methods, such as interpolation, between nearest neighbor voxels or by volume resampling. The halftone dots of the voxels of the converted overlap volume UOV2 (i) lie on the halftone dots of the voxels of the overlap volume OV1 (i) in the coordinate system K1 , At each grid point in the area of the overlap volumes, the result is a voxel value from the first volume V1 and a converted voxel value from the second volume V2 in front.

In step S7 becomes a bandpass on the overlap volume OV1 (i) and in the step S8 a bandpass on the overlap volume UOV2 (i) applied. On the overlap volume OV1 (i) and on the converted overlap volume UOV2 (i) In each case in the manner described, a bandpass is applied, which adjusts the Fourier spaces of the two overlap volumes temporarily or reversibly so that both contain only those frequencies that are also included in each other.

It is envisaged that the bandpass will apply spectral analysis to the Fourier representations of the two overlap volumes and filter the Fourier spaces such that all frequencies which are not present in the Fourier spectrum of the other overlap volume are removed from the Fourier spectrum. This gives two filtered Fourier spaces of the overlap volumes. These are transformed back into the space and you get corrected voxel values of the volumes FOV1 (i) and FOV2K1 (i) in the coordinate system K1 , In these volumes, the artifacts are opposite to the overlap volumes OV1 (i) and UOV2 (i) partially reinforced because they were matched in appearance.

Alternatively, it is provided that the bandpass acts directly in the position space, but in its effect equalizes the two Fourier spaces in the same manner, without explicitly performing a Fourier transformation. Again, one obtains corrected voxel values of the volumes FOV1 (i) and FOV2K1 (i) in the coordinate system K1 , In these volumes, the artifacts are opposite to the overlap volumes OV1 (i) and UOV2 (i) again partially reinforced, as they were matched in appearance.

In step S9 becomes a similarity measure to the filtered overlap volumes FOV1 (i) and FUOV2 (i) applied. The similarity measure S (i) and the associated coordinate transformation T12 (i) are stored in a memory. When a maximum number of iteration steps is reached, or when a maximum value of the similarity measure S (i) is achieved in which no improvement in the similarity measure S (i) would be expected in further iterations, the iteration is aborted and to step S10 jumped. If this is not the case, a jump is made to step S3 increasing the value of i to i + 1 and a new coordinate transformation T12 (i + 1) , This new coordinate transformation is determined in the most efficient way by the optimization method used. It is also provided that an optimization method, eg a method from the family of genetic optimization, per iteration step i multiple coordinate transformation T12 (i, j) for example, in parallel execution checks and from it a new set of coordinate transformations T12 (i + 1, j) created in the next iteration step i + 1.

• • Methoden aus der konvexen oder nicht-konvexen Optimierung wie Gradientenabstieg, (L-)BFGS(-B), Nelder-Mead,
• • genetische Optimierung, simuliertes Annealing,
• • Methoden aus dem Feld der Phasenkorrelierung, um direkt im FourierRaum arbeiten zu können.

The automatic determination of the selection of coordinate transformations T12 (i) For example, but not limited to the examples below, an optimization algorithm is provided from the set of algorithms mentioned below:

• • Methods from convex or non-convex optimization such as gradient descent, (L-) BFGS (-B), Nelder-Mead,
• • genetic optimization, simulated annealing,
• • Methods from the Phase Correlation field to work directly in Fourier space.

Such an optimization algorithm is arranged to automatically select the coordinate transformations so that the similarity measure gradually approaches the optimum value.

In step S10 becomes the optimal similarity measure S (i) determined and the associated coordinate transformation T12 (i) provided as the optimal registration information for further use.

The optimal coordinate transformation T12 (i) from step S10 is used to assemble the first volume V1 with the second volume V2 used according to a method known from the prior art. This will be the entire volume V2 in the coordinate system K1 and then the voxel values of the entire volume are transformed V2 on voxels in the raster of the first volume voxels V1 converted by means of known methods, such as interpolation between nearest neighbor voxels or by volume resampling. The halftone dots of the voxels of the converted overlap volume UV2 lie on the halftone dots of the voxels of the volume V1 in the coordinate system K1 , At each grid point, in the area of the overlap volumes, the result is a voxel value from the first volume V1 and a converted voxel value from the second converted volume UV2 in front.

In 3a and 3b schematically show projections based on differently selected radiation source geometries.

3a shows a parallel beam geometry 31 , The parallel rays 31 then form a three-dimensional object 32 as a two-dimensional structure 33 from. When using a parallel beam geometry 31 becomes the three-dimensional object 32 so pictured that all edges 34 and 35 the two-dimensional structure 33 can be displayed completely and sharply.

3b shows a cone beam geometry 41 , The cone beams 41 then form a three-dimensional object 42 as a two-dimensional structure 43 from. When using cone beams 41 becomes the three-dimensional object 42 so imaged that not all edges of the two-dimensional structure 43 be presented sharply. The edges which are completely scanned by the cone beams become sharp edges 45 displayed. The edges which are not completely (incompletely) scanned by the cone beams are called blurred edges 44 displayed. These as blurred edges 44 Imaged structures correspond to an incompletely scanned space.

Further, particularly advantageous embodiments and refinements of the invention emerge from the dependent claims and the description, wherein the independent claims of a claim category can also be developed analogously to the dependent claims of another claim category and in particular also individual features of different embodiments or variants to new embodiments or variants can be combined.

• 1: Schematische Darstellung eines Kegelstrahl-Computertomographen am Beispiel eines C-Bogens
• 2: Schematische Darstellung des Ablaufs der Koordinatentransformation
• 3a: Schematische Darstellung einer Projektion eines Objektes mittels einer Parallelstrahlgeometrie
• 3b: Schematische Darstellung einer Projektion eines Objektes mittels einer Kegelstrahlgeometrie

The innovation is explained in more detail below with reference to the figures.

• 1 Schematic representation of a cone beam computer tomograph using the example of a C-arm
• 2 : Schematic representation of the course of the coordinate transformation
• 3a : Schematic representation of a projection of an object by means of a parallel beam geometry
• 3b : Schematic representation of a projection of an object by means of a cone beam geometry

LIST OF REFERENCE NUMBERS

1

C-arm X-ray system

2

arc-shaped C-arm

3

X-ray tube

4

High voltage generator

5

collimator

6

Rontgenbilddetektor

7

Projection image generation unit

20, 20 ', 20 "

scan center

21

object

22

central beam

23

patient support

24

Cone beam computed tomography scanner motor control unit

25

Storage engine control unit

50

ROI

51, 51 ', 51 "

scan plane

100

Control unit

101

X-ray control unit

102

Volume comparison unit

103

Motion control unit

104

Image capture control unit

105

X-ray detector control unit

106

Image processing and storage unit

107

Scan parameter generation unit

108

Overlap volume determining unit

111

display

112

Input device unit

113

Operator / Operator

114

output unit

115

Organ program database

116

mass storage

117

reconstruction unit

118

Fourier transformation and filtering unit

130

DICOM Interface

140

network

141

RIS radiology information system

142

HIS Hospital Information System

31

Parallel beam geometry

32

Three-dimensional object

33

Two-dimensional structure

34

fully imaged edge

35

fully imaged edge

41

Cone-beam geometry

42

Three-dimensional object

43

Two-dimensional structure

44

fully imaged edge

45

incompletely imaged edge

V1

first volume

V2

second volume

K1

Coordinate system of the first volume

K2

Coordinate system of the second volume

T12 (i)

Coordinate transformation of the ith iteration step between the coordinate system K1 and the coordinate system K2

OV1 (i)

Overlap volume of the first volume V1 in the coordinate system K1 for the iteration step i

OV2 (i)

Overlap volume of the second volume V2 in the coordinate system K1 for the iteration step i

UOV2 (i)

Converted overlap volume in the coordinate system K1 in which the voxels of the overlap area OV2 (i) on voxels in the raster of the overlap area OV1 (i) are converted.

UV2

Converted second volume in the coordinate system K1 , where the voxels of the with the optimal coordinate transformation T12 transformed second volume V2 in the coordinate system K1 on voxels in the raster of the first volume voxels V1 are converted.

F

bandpass

FOV1 (i)

Overlap volume with changed voxel values after applying the bandpass F on the volume OV1 (i)

FUOV2 (i)

Overlap volume with changed voxel values after applying the bandpass F on the volume UOV2 (i)

S (i)

Similarity measure of the iteration step

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• DE 10014668 A1 [0016]
• DE 102007025862 A1 [0017]
• DE 102007034218 B4 [0018]
• DE 102008062030 A1 [0019]
• EP 520778 B1 [0020]
• US 20140267267 A1 [0022]
• US 20150371420 A1 [0023]
• US Pat. No. 6711293 B1 [0024]
• US 9129389 B2 [0025]
• WO 11039672 A1 [0026]
• WO 12048070 A1 [0027]
• WO 12139031 A [0028]
• WO 16097174 A1 [0029]

Claims (24)

X-ray system, for the iterative determination of an optimal coordinate transformation between a first N-dimensional partial volume V1 and a second N-dimensional partial volume V2 of an extended object, these partial volumes partially overlapping in the overlap volumes OV1 and UOV2, and wherein the X-ray system is suitable a) to record partial volumes V1 and V2 by discrete scanning of the expanded object, b) applying a bandpass filter to the overlap volumes OV1 and UOV2 within each iteration step i which removes from the overlap volume OV1 (i) those frequencies which are not included in the overlap volume UOV2 (i) and which is from the overlap volume UOV2 (i) removes those frequencies which are not contained in the overlap volume OV1 (i), whereby filtered overlap volumes FOV1 (i) and FUOV2 (i) are obtained, c) To determine a coordinate transformation by iteratively optimizing a similarity measure between the filtered overlap volumes FOV1 (i) and FUOV2 (i). X-ray system after Claim 1 characterized in that the bandpass filter determines the frequencies to be removed from information about the type or geometry of the scan. X-ray system after Claim 1 characterized in that the bandpass filter to be removed frequencies in the Fourier space of the overlap volumes OV1 (i) and UOV2 (i) determines that the absolute amount applied to them falls below a predetermined threshold. X-ray system according to one of the preceding claims, characterized in that the frequencies to be removed by means of the bandpass filter in the spatial space temporarily and for each assumption of the overlap volumes OV1 (i) and UOV2 (i) separately before calculation of the similarity measure from the two overlap volumes are removed. X-ray system according to one of Claims 1 - 3 characterized in that the frequencies to be removed by means of the bandpass filter in the Fourier space are removed temporarily and separately for each iteration step i before calculating the similarity measure under the current assumption of a coordinate transformation from the Fourier space of the two overlap volumes OV1 (i) and UOV2 (i). X-ray system according to one of the preceding claims, characterized in that the scanning is carried out by means of a cone-beam computer tomograph. X-ray system after Claim 6 characterized in that the cone beam computer tomograph is a mobile C-arm. X-ray system after Claim 1 characterized in that the overlap of the two partial volumes V1 and V2 is achieved by adjusting a storage of the object. X-ray system after Claim 8 characterized in that the storage is adjustable by motor. X-ray system after the Claims 9 and 6 or 9 and 7 characterized in that the overlap is achieved in that the storage is adjusted by the cone beam computer tomography motor. X-ray system according to one of Claims 6 or 7 or 8th or 9 characterized in that the cone beam computer tomograph before starting the scan asks an operator to make a manual adjustment of storage and the operator confirms the manual adjustment. X-ray system after Claim 11 characterized in that the storage has a position sensor and the cone beam computer tomograph by monitoring this position encoder ensures the correct adjustment. X-ray system after Claim 6 or 7 characterized in that the overlap is achieved by moving the cone beam computer tomograph relative to the object in space. X-ray system after Claim 13 characterized in that the cone beam computer tomograph is moved by a motor. X-ray system after Claim 13 or 14 characterized in that the cone beam computer tomograph has a position sensor and ensures the correct overlap of the subvolumes V1 and V2 by evaluating the position sensor. X-ray system according to one of the preceding claims, characterized in that the partial volumes V1 and V2 are joined together using the determined coordinate transformation to a combined volume. X-ray system after Claim 16 characterized in that the combination of the partial volumes V1 and V2 takes place in that the combined volume in the overlapping area of the Subvolumes V1 and V2 after applying the coordinate transformation at each coordinate by weighted averaging of the measured values present in the subvolumes V1 and V2 at this coordinate. X-ray system after the Claims 6 or 17 characterized in that the weights are a function of the cone beam angle present at a coordinate. X-ray system according to one of Claims 16 . 17 or 18 characterized in that the combined volume in the overlap region of the sub-volumes is generated so that initially the frequencies contained in the Fourierraum an overlap volume are transmitted at such locations in the Fourier space of the other, where this itself contains no frequencies. X-ray system according to one of the preceding claims, characterized in that after the combination of the partial volumes V1 and V2, additional partial volumes are added to the resulting total volume in the same way. X-ray system after one of the previous ones Claims 6 . 7 . 8th or 9 characterized in that the determined coordinate transformation between the cone beam computer tomography and the object storage is stored. X-ray system after Claim 21 characterized in that the stored coordinate transformation is used to accommodate future sub-volumes without re-optimization directly after one of Claims 16 - 19 to combine. Computer program product with a computer program, which is directly in a memory unit of a control unit of a cone beam computer tomography, especially a C-arm, loadable, with program sections to all steps according to one of Claims 1 to 22 when the computer program is executed in the control unit of the cone-beam computer tomograph. Computer-readable medium on which are stored by a computer unit readable and executable program sections to all steps according to one of Claims 1 to 22 execute when the program sections are executed by the computer unit.

Images