Test pattern and method for calibrating an X-ray imaging device

1. A calibrating test pattern comprising:

a volume support equipped with markers with a radiological absorbance providing contrast with respect to the volume support, the markers being distributed in a three-dimensional pattern and in subsets distributed in substantially parallel straight lines, such that sequences of cross-ratios are constructed from the subsets of markers, each sequence of cross-ratios comprising a single cross-ratio for each quadruplet of markers,

wherein quadruplets of the markers are in an order depending on the rank number of the respective markers along the straight line on which they are aligned in a first direction, said order being common to all of the cross-ratios, and when a subset of markers comprises at least five markers, the order of the cross-ratios in the respective sequences of cross-ratios is defined by a predefined rule common to all of the sequences of cross-ratios, each of the sequences of cross-ratios being different.

2. The calibrating test pattern according to claim 1 , wherein all of the markers have substantially the same size and substantially the same shape.

3. The calibrating test pattern according to claim 1 , wherein the straight lines are observable generatrices of a cylinder.

4. The calibrating test pattern according to claim 1 , wherein for all of the projection conditions under which images are acquired during calibration, no marker projection overlaps another marker projection and/or when the marker projections issued from two subsets of markers do overlap, the projections of the markers of the other subsets of markers do not overlap.

5. The calibrating test pattern according to claim 1 , wherein the subsets of markers are distributed in respective observable straight-line segments parallel to a z-axis, the straight-line segments being of the same length and having the same coordinates along said z-axis, each straight-line segment accommodating a first positive integer m of sites capable of being occupied by a marker, any two consecutive sites considered along said observable segment being spaced apart by a pitch, each site respectively being assigned a first value or a second value depending on whether the site is occupied by a marker or not, the markers being distributed so that the values attributed to a second positive integer n, at most equal to m, of any consecutive sites considered in a given direction along respective straight-line segments form respective binary codes composed of n bits, each binary code composed of n bits formed in said direction being unique.

6. The calibrating test pattern according to claim 5 , wherein the first positive integer m is higher than the second positive integer n.

7. The calibrating test pattern according to claim 5 , wherein the markers are distributed over the test pattern, such that, for a known number of straight-line segments and of sites per straight-line segments, and for known degrees of occupation of the sites of straight-line segments, a difference between the binary codes formed by the values taken by the m consecutive sites accommodated by respective straight-line segments in the direction is maximal, the binary codes being sections of a series obtained by means of a LFSR of n bits, m being lower than or equal to n.

8. The calibrating test pattern according to claim 1 , wherein the order of the cross-ratios in each cross-ratio sequence constructed from a subset of markers aligned on a straight line is defined in the following way:

for markers denoted A g having rank numbers g = 1 to N along the straight

line in a

second direction,

 if i is from 1 to N−3 then:

(-if j is from i+1 to N−2 then:

 (-if k is from j+1 to N−1 then:

(-if l is from k+1 to N then:

 (-insert the following cross-ratio in the sequence, the

 following cross-ratio being a cross-ratio calculated

 with the markers A g=i , A g=j , A g=k , A g=l ,

 − l=l+1)

− k=k+1)

− j=j+1),

 −i=i+1),

where N is a positive integer.

9. The calibrating test pattern according to claim 1 , wherein the order of the markers in each quadruplet of markers is the order of the markers along the straight line on which they are aligned in the first direction.

10. A method for determining geometric characteristics of an x-ray imaging device for producing three-dimensional projections of an object by reconstruction based on two-dimensional projections of said object, said method using a calibrating test pattern according to claim 1 , the method comprising the following steps:

placing the calibrating test pattern in a projection zone between an x-ray source and an x-ray detector;

acquiring at least one projection of the calibrating test pattern in at least one imaging-device geometry defined by the positions of the source, the test pattern, the detector, and relative orientations;

and, for each projection of the test pattern:

detecting the projections of markers in the projection;

determining the positions of the projections of markers in the projection;

detecting alignments of projections of the markers in respective image straight lines; and

for each alignment of projections of markers:

forming an image sequence of cross-ratios from the projections of markers forming said alignment, the image sequence comprising a single cross-ratio per quadruplet of projections of markers in which the projections of markers are ordered in the order depending on the rank number of respective projections of markers along the corresponding image straight line in the predefined first direction, and, when a set of projections of markers comprises at least five marker projections, the order of the cross-ratios in the image sequences is defined by the predefined rule along the corresponding image straight line; and

for each marker projection forming said detected alignment, identifying the generating marker by attributing a cross-ratio sequence formed from the marker projections of said alignment to a portion of a cross-ratio sequence formed by markers of the test pattern of the same number of cross-ratios as the image sequence; and

determining the geometric characteristics that the imaging device had during the acquisition of the projection of the test pattern from the positions of those projections of markers which were detected and the positions of the respective corresponding markers.

11. The calibrating method according to the claim 10 , wherein the step of detecting alignments is carried out by applying a Hough transform to the positions of the projections of markers that were detected in the step of detecting the projections of the markers.

12. A process for manufacturing a calibrating test pattern according to claim 5 , comprising a step of distributing the markers such that the binary codes are sections of a series obtained by means of an LFSR of n bits.