Measuring linear separations in digital radiographs
First Claim
1. A method for increasing measurement precision in digital radiography, comprising:
- (a) receiving digital radiographic data for one or more objects in an actual condition;
(b) creating a first data profile based on the received data, the first data profile corresponding to a selected region of the one or more objects and being associated with the actual condition;
(c) deriving a second data profile from the received data, the second data profile representing an expected data profile for the one or more objects when in a reference condition;
(d) calculating a difference between the first data profile created in step (b) and the second data profile derived in step (c); and
(e) determining, based upon the calculated difference, a degree by which the actual condition varies from the reference condition in the selected region.
6 Assignments
0 Petitions
Accused Products
Abstract
Digital pixel data is obtained from radiographic imaging of one or more objects, and corresponds to an imaged area containing a feature to be measured. A data profile for a region around the measured feature is created from the digital pixel data. A reference profile is then created from the data profile. The reference profile represents an expected data profile for a reference condition of the objects, and accounts for the point spread function of the imager. The difference between the data profile and the reference profile is calculated. Based on that difference, the degree by which the actual condition of the objects varies from the reference condition is determined. The calculated difference can be compared to a lookup table mapping previously calculated differences to degrees of variation from the reference condition. The calculated difference can also be used as an input to an experimentally derived formula.
50 Citations
29 Claims
-
1. A method for increasing measurement precision in digital radiography, comprising:
-
(a) receiving digital radiographic data for one or more objects in an actual condition;
(b) creating a first data profile based on the received data, the first data profile corresponding to a selected region of the one or more objects and being associated with the actual condition;
(c) deriving a second data profile from the received data, the second data profile representing an expected data profile for the one or more objects when in a reference condition;
(d) calculating a difference between the first data profile created in step (b) and the second data profile derived in step (c); and
(e) determining, based upon the calculated difference, a degree by which the actual condition varies from the reference condition in the selected region. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14)
-
-
15. A computer-readable medium having stored thereon data representing sequences of instructions which, when executed by a processor, cause the processor to perform steps comprising:
-
(a) receiving digital radiographic data for one or more objects in an actual condition;
(b) creating a first data profile based on the received data, the first data profile corresponding to a selected region of the one or more objects and being associated with the actual condition;
(c) deriving a second data profile from the received data, the second data profile representing an expected data profile for the one or more objects when in a reference condition;
(d) calculating a difference between the first data profile created in step (b) and the second data profile derived in step (c); and
(e) determining, based upon the calculated difference, a degree by which the actual condition varies from the reference condition in the selected region. - View Dependent Claims (16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28)
-
-
29. A method for determining a linear separation distance between two objects having substantially dissimilar densities, comprising:
-
(a) receiving digital image data for the objects, the data created using an imager having a Gaussian point spread function with a standard deviation of the same order of magnitude as the separation distance;
(b) creating an initial data profile from a first set of pixel data values from the image data received in step (a), the first set of pixel data values corresponding to a first region of an area imaged by the imager;
(c) creating a normalizing data profile from a second set of pixel data values from the image data received in step (a), the second set of pixel data values corresponding to a second region of the image area;
(d) normalizing the initial data profile by the normalizing data profile to yield a normalized data profile;
(e) deriving a reference data profile from the normalized data profile, the reference data profile representing an expected data profile for the first region when the objects have a reference separation, the reference data profile having a first section approximated by
wherein;
P(x,M,σ
) is the Gaussian point spread function for the imager,x is a location along an axis of the separation between the objects, M is the location of a boundary of one of the objects, σ
is the standard deviation of the imager point spread function,x0 and xfinal are linear locations along the separation axis and located on opposite sides of the boundary, and H is a constant;
(f) estimating M based on the maximum of the first derivative of the normalized data profile;
(g) calculating differences between the reference data profile with varying values for M and the normalized data profile;
(h) choosing, based on the differences calculated in step (g), an optimized value for M;
(i) deriving functions for lines corresponding to second and third sections of the reference data profile on opposite sides of the first section along the separation axis;
(j) fitting the first section to the second and third sections; and
(k) determining the degree of variance from the reference separation using the formula s=[Δ
(s)−
a0*R′
(s)MAX−
b0]/[a1*R′
(s)MAX+b1], wherein;
s is the object separation, Δ
(s) is a difference between the normalized data profile and the reference data profile,R′
(s)MAX is the maximum of the first derivative of the normalized data profile, anda0, a1, b0 and b1 are constants derived from measurements of test samples having known separations.
-
Specification