IP Library Granted Patent US 12711608
Granted Patent B2
US 12711608 · App. 18/437,350 · Granted Aug 18, 2026

Inspecting parts using geometric models

Inventors: Srinivas Ravela (Farmington, CT); Alan Matthew Finn (Chapel Hill, NC); Ozgur Erdinc (Mansfield, CT); Sergio S. Frutuoso (Avon, CT); Eugene B. Donskoy (Marblehead, MA); Joseph A. Sylvestro (Avon, CT)
Assignee: RTX CORPORATION
G06T7/001B25J19/021G06T7/70G06T17/00G06T2207/30164G06T2207/30244G06V10/751G06V10/764
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12711608
App. No.
18/437,350
Granted
Aug 18, 2026
Kind
B2
Abstract

Examples described herein provide a method that includes comparing features in an image of a part with features of models in a database and creating a list of candidate models for the part. The method further includes rendering a three-dimensional (3D) model at a pose/scale of an imaging system to generate a synthetic image that contains features of the part. The method further includes creating a part representation based on the features of the part, wherein the part representation depicts a defect of the part. The method further includes classifying a region of the part based on a set of rules for the part to determine whether the defect is within an allowable tolerance for the part.

Claims (26)

1 . A method comprising:

comparing features in an image of a part having a defect located in part region with features of models in a database and creating a list of candidate models for the part;

estimating a pose/scale of an imaging system to determine an actual pose/scale of the part appearing in the image;

rendering a three-dimensional (3D) model at the pose/scale of the imaging system to generate a synthetic image that has a synthetic pose/scale matching the actual pose/scale of the image and that contains features of the part;

creating a part representation based on the features of the part, wherein the part representation depicts a synthesized defect at a synthesized region of the synthesized image that matches a part region of the part appearing in the image; and

classifying a part region of the part appearing in the image based on a set of rules for the part and the synthesized defect at the synthesized region of the synthesized image to determine whether the defect is within an allowable tolerance for the part,

wherein the allowable tolerance is assigned to the part region and the allowable tolerance is selected based on a location of the part region on the part.

2 . The method of claim 1 , further comprising capturing the image of the part using the imaging system.

3 . The method of claim 1 , further comprising determining values for a camera using the part representation.

4 . The method of claim 1 , wherein the synthetic image comprises an indicium associated with a defect of the part.

5 . The method of claim 1 , wherein different tolerances are assigned to different part regions containing a defect and wherein an assigned tolerance among the different tolerances is selected based on a location of a part region among the different part regions.

6 . A processing system comprising:

a memory comprising computer readable instructions; and

a processing device for executing the computer readable instructions, the computer readable instructions controlling the processing device to perform operations comprising:

comparing features in an image of the part with features of models in a database and creating a list of candidate models for the part;

retrieving a three-dimensional (3D) model for the part from the list of candidate models based on the comparison;

estimating a pose and scale (pose/scale) of an imaging system to determine an actual pose/scale of the part appearing in the image;

rendering a three-dimensional (3D) model at the pose/scale of the imaging system to generate a synthetic image that has a synthetic pose/scale matching the actual pose/scale of the image and that contains features of the part; and

creating a part representation based on the features of the part, wherein the part representation depicts a synthesized defect at a synthesized region of the synthesized image that matches a part region of the part appearing in the image;

performing a stress analysis for the part based on the 3D model for the part to determine whether the part meets a performance specification by classifying a part region of the part appearing in the image based on a set of rules for the part and the synthesized defect at the synthesized region of the synthesized image,

wherein an allowable tolerance is assigned to the part region and the allowable tolerance is selected based on a location of the part region on the part.

7 . The processing system of claim 6 , wherein the image of the part is captured using the imaging system.

8 . The processing system of claim 6 , wherein the performance specification for the part defines a limit on a size of a defect for the part.

9 . The processing system of claim 6 , wherein the performance specification for the part defines a limit on a number of defects for the part.

10 . The processing system of claim 6 , wherein the performance specification for the part defines a limit a total amount of part deformity that is acceptable.

11 . The processing system of claim 6 , wherein different tolerances are assigned to different part regions containing a defect and wherein an assigned tolerance among the different tolerances is selected based on a location of a part region among the different part regions.