WO2017133845A1 - Method for reconstructing a test object in an x-ray ct method in an x-ray ct system without a manipulator - Google Patents

Abstract

The invention relates to a method for reconstructing a test object in an x-ray CT method in an x-ray CT system, which has an x-ray source having a focus and has an x-ray detector but does not have a manipulator, wherein, in advance, the positions of at least three markers, which are fixedly connected to an object, are precisely determined. Then, the test object is fastened to the object in a stationary manner by means of suitable measures, wherein this fastening occurs for the duration of the creation of recordings of the test object, then the object is moved together with the test object on an arbitrary path between the x-ray source and the x-ray detector by the x-ray CT system and recordings are created during this movement, the projection geometry of each individual recording is calculated from the known positions of the markers in relation to each other in the recording, and then a CT reconstruction of the test object is performed by means of a suitable algorithm on the basis of the association of the individual recordings with the respective projection geometries.

Description

 Method for reconstructing a test object in an X-ray CT method in an X-ray CT system without a manipulator

The invention relates to a method for the reconstruction of a

Test object in an X-ray CT method in an X-ray CT system, which has an X-ray source with focus, an X-ray detector but no manipulator.

Method for the three-dimensional reconstruction of X-ray transparent

Test objects taking account of a certain projection geometry from projections are known from the field of computed tomography. For such a reconstruction of a test object from projections of different projection directions in an X-ray CT apparatus, an exact knowledge of the projection geometry of each individual projection is required. In terms of X-ray imaging, the projection geometry describes the relative position of the test object to the focus of the X-ray source and the X-ray detector. to

Verification of the actual projection geometry in each projection must be different points of the test object in projections

Projection directions are exactly recovered. In transparent

These objects can be superimposed on test objects by unknown object structures, so that in the worst case the points are not recognized, or their position can not be determined without error.

Markers are special objects with known properties. In the context of this application, a marker is understood to mean an object that has at least one image feature that is unique in each projection performed

is locatable. To determine the projection geometry from a projection, at least three image features must be able to be determined. For this purpose, the markers must be determined in at least three positions within a projection, which are not in a trajectory belonging plane. These

Image features may be present either within a single object (e.g., the corners of a triangle) or formed on multiple objects be (for example, three balls). Even small errors in the position determination can lead to large errors in the reconstructed

Lead projection geometry. Typically, the projections used lie on a circle around the test object. The range of the focal point of the

X-ray source is here as a trajectory or more precisely as

 Circle trajectory called. But there are also other trajectories possible, such as helix trajectories.

The projections must be made as above - from different

Projections directions are included. Depending on how completely the test object has been imaged along the so-called trajectory in the sense of the Tuy-Smith theorem (all planes through a test object have to intersect the trajectory of the x-ray source in at least one point), the more detailed the image of the reconstructed volume.

Therefore, according to the state of the art for reconstruction, as precise a manipulator as possible is used in order to drive off a specific trajectory. The problem here is that the more precise such a manipulator system is, the more

more expensive is the production, where the increase can be exponentially recognized. Classically, a manipulator consists of industrial

 Computed tomography from an array of orthogonal linear axes and usually at least one axis of rotation to rotate the test object can. A standard circular trajectory and a detector of 1,000 pixels wide typically requires more than 1,000 projections to reconstruct the volume. However, the circle trajectory fulfills the Tuy-Smith theorem only for the middle layer, so more complex trajectories are preferable. The more different details from as few

Projection directions are mapped, the fewer projections are required to represent a volume of high quality. In order to be able to use a projection for the reconstruction of the test object, it must be possible due to the

Projection geometry the test object can be mapped within the projection surface. The projection geometry must be at the time of reconstruction be known to allow an unambiguous assignment of the images to the projection geometry under which it was recorded, and then from a known reconstruction algorithm

Three-dimensional rendering of the test object to calculate.

The object of the invention is to provide a method with which a

Reconstruction of a test part, at least a portion of the same, can be achieved with the least possible expenditure on equipment. The object is achieved by a method having the features of patent claim 1. The method according to the invention describes an X-ray CT method for the three-dimensional reconstruction of a test part in an X-ray CT system. In contrast to the methods known from the prior art, which require a precise manipulator as possible, must in the inventive

Method the projection geometry can not be prepared with the help of such a manipulator beforehand. Therefore, it is also possible to have a test object too

reconstruct, located on a free path between X - ray tube and

X-ray detector moves. For example, the test object can be in free fall, for example on a helical path (analogous to a helix trajectory) or on a trajectory or roll down a ramp. In X-ray imaging, X-ray source and X-ray detector must be arranged fixedly in space. The projection geometry to everyone

Projection is calculated by an external observer from the respective X-ray image based on the known relative positions of the markers to one another. The test object or the projection geometry are still unknown at the time of the individual recording. This problem is solved in that a fixed arrangement of markers are fixedly connected during the preparation of the recordings with the test object. By using the markers, the exact position of the test object in three-dimensional space is subsequently determined. The position of the markers can be chosen randomly. Only the distances to each other must not be variable. This means that the markers are also in fixed positions to the test object during the measurement. According to the invention, it is necessary to determine in advance the positions of the markers on the object precisely. Preferably, the object with the markers for it is measured by means of a coordinate measuring system. The positions of each projected marker on the screen are recorded in each shot

(Projection) determined. With the help of a suitable optimization calculation including the projection model, the position of the markers in three-dimensional space can be determined. Not all markers need to be visible in each projection, but all the markers must have identified enough positions in the projection plane. It must be ensured that the markers are recognized correctly for different projection directions. This can be facilitated by a random arrangement of the markers, as it is easier with symmetrical position distribution of the marker depending on the perspective

Overlays and can lead to confusion. By the

Using the markers, the exact position of the test object in three-dimensional space (projection geometry) can be subsequently determined. Then, by means of an algorithm known from the prior art (for example in the context of the use of the software "CERA Xplorer" of the company Siemens Healthcare GmbH), a three-dimensional reconstruction of the test object is carried out With the method according to the invention, it is no longer necessary to use a manipulator in X-ray -CT method to use and the cost factor manipulator is eliminated.

An advantageous development of the invention provides that the test object is fixed in a vessel, in particular a mesh basket or a hollow body, which in particular has the shape of a cylinder, and the markers are attached to the vessel. The vessel must be suitable for fixing the test object. The fixation of the test object can be done for example by positive or negative pressure or by appropriate filler or

conventional fastening devices such as grippers, clamps, straps or tensioning devices. For a circular trajectory, a cylinder is an advantageous shape; this can be designed as a barrel-shaped vessel. A further advantageous embodiment of the invention provides that the vessel made of a material with the lowest possible

 Mass attenuation coefficient exists. As a result, the smallest possible superimposition of the image of the test object through the vessel is achieved.

A further advantageous development of the invention provides that the material of the marker has a mass attenuation coefficient which is significantly different from that of the test object, in particular deviates from this by at least 20%. This will ensure that you are in the

The fluoroscopic image recognizes a clear difference between the image of the marker (s) and the structures of the test object.

A further advantageous development of the invention provides that all markers are made of the same material and / or balls. If all markers are made of the same material, the segmentation of the markers within the projection can be simplified, as only a certain gray scale range has to be searched for. The spherical shape is an extremely suitable shape for a marker, because the center of gravity of a sphere due to the point symmetry of all

Projection directions can be accurately determined and balls can be machined very precisely. It is therefore particularly preferable if all markers are spheres. By incorporating the properties of the markers as spherical, this can be exploited as prior knowledge, because due to the symmetry of the center of gravity of the sphere is always mapped to the same point in space in the projection.

A further advantageous embodiment of the invention provides that the arbitrary web is a free fall, a screw path, a ballistic web, in particular a trajectory, or a roll on a ramp. A further advantageous cost-saving development of the invention provides that the X-ray source and the X-ray detector are fixed to each other in space and their alignment with each other not with the highest precision must be done because the projection geometry can be calculated from the arrangement of the marker.

A further advantageous development of the invention provides that the markers have as far as possible a distance from each other but during the

Creation of the recordings so in the beam path of the X-ray tube are that they are still detected in the X-ray detector. Positioning the markers becomes more robust the farther individual markers are from each other. The larger the distance between two markers, the more accurate the distance of the markers on the projection can be measured. Ideally, therefore, the markers are imaged near the edges of the projection, thus using the relative distances of remotely located markers.

A further advantageous development of the invention provides that the motion blurring directly dependent on the exposure time of the X-ray detector is determined in each case by the evaluation of the markers. Through a generally iterative approach, the motion blur can be modeled and eliminated and thus the detail resolution can be increased. All specified in the dependent claims features of advantageous

Further developments are each individually and in any combination of the invention.

In the following, an advantageous embodiment of a

inventive method explained.

The method according to the invention runs in a conventional X-ray CT system (for example the model "Y.MU56TB" from YXLON International GmbH), the manipulator used hitherto being omitted, for example the model "Y.TU-225- D04 "of the company YXLON International GmbH and as X-ray detector, for example, the model" XRD 0822 AP18 IND "from PerkinElmer used. The inventive method is based on the use of a three-dimensional array of markers, which are visible in the X-ray projection and attached to a cylindrical vessel. Balls are used as markers, because the center of gravity of a sphere can be determined exactly from all projection directions due to the point symmetry, and balls can be made very precisely mechanically. The markers and the vessel are made of materials whose mass attenuation coefficients are significantly different from the mass attenuation coefficient of the test object, for example, the mass attenuation coefficients differ by more than 50% from each other.

As a result, the fluoroscopic image shows a clear difference between the image of the marker or the vessel and the structures of the test object. In addition to the material of the markers, their size and number must also be taken into account. In particular, the choice of a suitable size and the material are highly dependent on each other. The number of markers should be based on the size, shape and magnification of the volume to be reconstructed. Since there is only a limited space available for the markers, their size and number can also compete with each other.

In addition, an unfavorable choice of material may entail unwanted side effects, such as X-ray scattering, and must therefore also be matched to the application. A general rule for the correct choice of the three factors can not be formulated. Instead, the choice should be made individually for each application.

The arrangement of the markers must ensure that a sufficient number of markers are imaged in each projection image of a CT scan. In order to be able to ensure this requirement independently of the test object, the arrangement of the markers, as stated above, takes place around the test object on the vessel. The distribution of the markers can be done completely arbitrarily, as long as the previously this condition is not violated. Depending on the application, it may be useful not to fall below a minimum distance between adjacent markers in order to avoid overlays several markers in the projection image as possible. The number of markers may not be less than a minimum of three markers, but may be any size, provided that the

previous criteria. In projection pictures with

The test object can be superimposed on its structures with the marker projections. Should the result of the reconstruction be adversely affected, this can be reduced by a clever arrangement of the markers.

The relative position of the markers relative to one another is determined by means of a known, highly accurate surveying method, for example by means of a coordinate measuring device, prior to the taking of the images. In the vessel, a test object is set, so that the test object during the entire recording of the same in the X-ray CT system is fixed to the vessel and thus to the markers. The fixation of the test object can be done for example by positive or negative pressure or by

corresponding filling material or conventional fastening devices, such as grippers, clamps, straps or tensioning devices.

In the projection images, the markers must be found and identified.

The structure resulting from the imaging of a marker in the X-ray image is referred to below as marker projection. To determine the

It is the image coordinate of such a two-dimensional marker projection

necessary that a unique anchor point can be defined for them.

This is made possible by the use of spherical markers. The

perspective projection of a spherical structure always results in a circular figure. This applies regardless of the position or orientation of the spherical marker. As an anchor point for the marker projection is therefore their focus. Determining the center of gravity of a circular Surface also provides robust results on digital image data, as long as the marker projection covers a sufficient number of pixels. The actual resolution of the X-ray detector and thus of the entire X-ray CT system then plays a minor role. Thanks to the sub-pixel-accurate determination of the center of gravity of a marker projection, deviations are on the order of magnitude that are extremely small, virtually negligible.

To be able to draw conclusions about the position and position of the test object in the X-ray CT system - and thus the projection geometry - from the position of the markers, it must be ensured that the relative position of the markers to the test object does not change. This is done by the above-mentioned fixation of the test object in the vessel.

Only if it is known where a marker can be seen in the projection image, its position can be determined in space and later used to assign the projection geometry to the respective recording. The challenge in extracting the marker projections is the distinction between these and other structures depicted in the projection image. Thus, parts of the vessel are also within the projected on the X-ray detector volume. The extraction of the marker projections takes place in two steps. First of all, all structures shown in the picture are recorded separately. Then it is determined on the basis of suitable criteria which of these structures corresponds to a marker projection and which not. The algorithm for determining the visible markers in a projection image can be divided into three logical steps. First, the visible structures are separated from the background. The result is a binary image of the same size, which classifies each pixel as foreground or background. Due to the spherical shape of the markers, some assumptions about the mapping by this generated structure can be used in the projection image for identification. So it can be assumed that a minimal enclosing rectangle, which is spanned along the axes in the image coordinate system around a marker structure, is almost square. Accordingly, its aspect ratio is almost equal to one. In addition, the entire structure must be within a maximum radius of the center of the structure, with no background pixels within that radius. The area of the structures can also be used to evaluate whether it can be a marker structure. It should be noted, however, that the markers are located at different positions in the room and are therefore displayed differently magnified. So it makes sense to define the permissible area in an interval, as long as their position in the room can be limited accordingly.

These criteria can be used to check whether a structure has a

Marker projection or not. All other structures are not relevant for the further course of the position determination and can be discarded.

Then, the focus is determined for each structure that represents a marker. The information extracted from the projection image then corresponds to a set of image coordinates, which respectively represent the position of a marker projection in the X-ray image. This is followed by an optimization calculation in order to determine the position of the markers in three-dimensional space. Thus, the projection geometry is known for each projection image and thus also the exact position and orientation of the test object in the X-ray CT system. The test object moves in the X-ray CT system on a completely freely selectable - possibly determined by chance - web. Thus, the object - together with the vessel in which it is fixed - can be thrown on a trajectory through the X-ray CT system. At the same time, it can rotate about different rotation axes at the same time - which corresponds to a rotational axis changing its spatial direction - (perform a quasi-tumbling motion, as it were). A ramp could also be installed between the x-ray tube and the x-ray detector, on which the cylindrical vessel rolls. It is particularly advantageous if the vessel by means of suitable guides and the gravity following a helical path. By means of the method according to the invention, a reconstruction of the test object which moves on any desired trajectory through the X-ray CT system is possible.

Since the test object is constantly in motion during recording, there is also some blurring with respect to the actual projection geometry. This is not unique because it changes due to movement during the recording time, which has a certain amount of time. This blur is reduced the shorter the exposure time of the X-ray detector. However, a correction can also be made by evaluating the markers by determining this blurring and by a generally iterative approach, the motion blur with respect to the test object is modeled and eliminated and thus the detail resolution is increased.

Finally, based on the known projection geometries in conjunction with the associated projection images by applying methods and algorithms known from the prior art, a very good three-dimensional reconstruction of the test object can be carried out - for example by means of the software "CERA Xplorer" from Siemens Healthcare GmbH.

Claims

claims Method for reconstructing a test object in an X-ray CT Method in an X-ray CT system having an X-ray source with focus, an X-ray detector but no manipulator, being preceded by the positions of at least three markers fixed with one Object are connected, can be precisely determined Thereafter, the test object by means of suitable measures is fixed on Object set, with this setting for the duration of the creation of Recordings of the test object take place, Thereafter, the object is moved with the test object on any path between X-ray source and X-ray detector through the X-ray CT system and created during this movement recordings, the projection geometry of each shot is from the calculated relative positions of the markers to each other in the respective recording, Thereafter, based on the assignment of the individual images to the respective projection geometry, a CT reconstruction of the test object is carried out by means of a suitable algorithm. Method according to claim 1, wherein the object is a vessel, in particular a mesh basket or a hollow body, which in particular has the shape of a cylinder. Method according to claim 2, wherein the vessel consists of a material with the lowest possible mass attenuation coefficient. Method according to one of the preceding claims, wherein the material of the marker has a mass attenuation coefficient which is significantly different from that of the test object, in particular by at least 20% thereof. 5. The method according to any one of the preceding claims, wherein the markers are all made of the same material and / or balls. 6. The method according to any one of the preceding claims, wherein the  any track is a free fall, a helical track, a ballistic track, in particular a runway, or a roll on a ramp. 7. The method according to any one of the preceding claims, wherein the  X-ray source and the X-ray detector fix each other in space  are arranged and their alignment with each other is not done with the highest precision. 8. The method according to any one of the preceding claims, wherein the  Markers have the widest possible distance from each other but during the creation of the recordings in the beam path of the  X-ray tube are that they are still detected in the X-ray detector. 9. The method according to any one of the preceding claims, wherein the directly dependent on the exposure time of the X-ray detector  Motion blur in the shots is determined by the evaluation of each marker.