US 10,969,771 B2Granted

Computed tomography for non-destructive evaluation of manufactured parts

OwnerLuke Mohr, Yu Hsuan Lee +2inventors

InventorsLuke Mohr, Yu Hsuan Lee, Nozomu Okuda, Alex Kitt

Life of this patent

Estimating the term…

Loading patent…

Abstract

A method for using computed tomography for non-destructive evaluation of a part, comprising: providing a computed tomography system; loading predetermined computed tomography system setup information; customizing settings with predetermined algorithmic parameters and functions depending on part features; processing projection counts; processing prior knowledge data relevant to the part; creating masking data from the customized settings; pre-processing reconstruction results based on the customized settings; performing algorithmic adjustments based on prior knowledge data and part geometry; performing algorithm adjustments during iterations; and post-processing reconstruction results.

Claims16

Dependent claims

A method for using computed tomography for non-destructive evaluation of a part by reconstructing a volume from X-ray projections gathered from a tomographic system using limited information, the method comprising:

(a) collecting measured projection data from a tomographic system using system hardware information including geometric configuration and scanning parameters;
(b) processing projection images and count data;
(c) processing and aligning prior knowledge data;
(d) generating a sinogram mask of the projection data from the prior knowledge data;
(e) generating a reconstruction mask in three-dimensional volume grid space from the prior knowledge;
(f) generating an initial three-dimensional relaxation matrix from the prior knowledge data;
(g) cropping projection data, prior knowledge data, sinogram and reconstruction mask data, and relaxation matrix data based on part and scanning geometry for reconstruction computation speed improvement;
(h) generating an estimate reconstruction based on the input masked prior knowledge data and populating a masked reconstruction space;
(i) forward projecting the estimate reconstruction and calculating the difference between the measured projection data and the estimated projection data;
(j) back projecting the differences over the reconstruction volume;
(k) weighting the back projected differences with the relaxation matrix and then combining the information with the estimate reconstruction and updating the estimate reconstruction;
(l) combining a set number of previous back projected differences to function as a heatmap for features not aligned or able to be aligned by the prior knowledge data;
(m) updating the relaxation matrix as a function of an accelerating or decelerating parameter based on the current iteration of the process and the calculated heatmap;
(n) repeating steps (i)-(m) until a convergence criterion is met;
(o) post-processing the reconstruction results by applying a smaller reconstruction mask than that generated in step (e);
and
(p) outputting the reconstructed tomographic image.

14 more elements

(a) collecting measured projection data from a tomographic system using system hardware information including geometric configuration and scanning parameters;
(b) processing projection images and count data based on part geometry information including dimensions and material composition;
(c) processing and aligning prior knowledge data, wherein the prior knowledge data includes three-dimensional CAD data, other reconstruction data derived from the same part, and part geometry information including dimensions and material composition;
(d) generating a sinogram mask of the projection data from the prior knowledge data;
(e) generating a reconstruction mask in three-dimensional volume grid space from the prior knowledge;
(f) generating a three-dimensional relaxation matrix from the prior knowledge data;
(g) cropping projection data, prior knowledge data, sinogram and reconstruction mask data, and relaxation matrix data based on part and scanning geometry for reconstruction computation speed improvement;
(h) generating an estimate reconstruction based on the input masked prior knowledge data and populating a masked reconstruction space;
(i) forward projecting the estimate reconstruction and calculating the difference between the measured projection data and the estimated projection data;
(j) back-projecting the differences over the reconstruction volume according to an SIRT-like iterative process;
(k) weighting the back projected differences with the relaxation matrix and then combining the information with the estimate reconstruction, and updating the estimate reconstruction;
(l) combining a set number of previous back projected differences to function as a heatmap for features not aligned or able to be aligned by the prior knowledge data;
(m) updating the relaxation matrix as a function of an accelerating or decelerating parameter based on the current iteration of the process and the calculated heatmap;
(n) repeating steps (i)-(m) until a convergence criterion is met;
(o) post-processing the reconstruction results by applying a smaller reconstruction mask than that generated in step (e);
and
(p) outputting the reconstructed tomographic image.

14 more elements

Dependencies are inferred from the claim text.

Facts

Application: 16/438,619Filed 12 Jun 2019
Published: US 2020/0393823 A117 Dec 2020
Granted: 6 Apr 2021
Term ends: Not estimatedThe source records don’t support an estimate.

Classification10 codes

G05B 19/41875
Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible…:…
G01N 23/046
And forming images of the material: using tomography, e.g. computed tomography [CT]
G01N 2223/401
Imaging: image processing
G01N 2223/426
Imaging: image comparing, unknown with known substance