# US 2025/0148056 A1 · Computer vision inferencing for non-destructive testing

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2025/0148056 A1
- Document ID: US20250148056A1
- Status: Published application
- Kind: A1
- Application number: 18/924,194
- Filed: 2024-10-23
- Published: 2025-05-08
- Earliest claimed: 2021-04-16
- Applicants: Baker Hughes Holdings LLC
- Inventors: Sheldon McCrackin; Jeremy Vandam; Carlneil Domkam; Samvid Jhaveri
- Web page: https://patentgrep.com/patent/US20250148056A1

## CPC

- G06F 18/285: {Selection of pattern recognition techniques, e.g. of classifiers in a multi-classifier system}
- G06T 7/001: {using an image reference approach}
- G06F 18/24: Classification techniques
- G06T 7/0004: {Industrial image inspection}
- G06T 7/70: Determining position or orientation of objects or cameras (camera calibration G06T7/80)
- G06V 10/809: {of classification results, e.g. where the classifiers operate on the same input data}
- G06T 2207/10016: Video; Image sequence
- G06T 2207/20016: Hierarchical, coarse-to-fine, multiscale or multiresolution image processing; Pyramid transform
- G06T 2207/30164: Workpiece; Machine component
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

An inspection system is provided and includes a camera and controller. The controller can include one or more processors in communication with the camera and receive a plurality of images of a target captured by the camera. The controller can also determine, using a first computer vision algorithm, a first prediction and corresponding confidence level for substantially all of the images. The controller can select a subset of the images having the first prediction confidence level greater than or equal to a first prediction threshold value. The controller can additionally determine, using a second computer vision algorithm, a second prediction and corresponding second prediction confidence level for each of the selected images. The at least one second prediction can require more time to determine than the at least one first prediction. The controller can output the second prediction and the second prediction confidence level for each of the selected images.

## Claims

### Claim 1

An inspection system, comprising:
a camera; and
a controller including one or more processors in communication with the camera, the controller being configured to:
receive one or more images of a target captured by the camera;
determine, using a first computer vision algorithm, at least one first prediction of an object of interest within the target and a corresponding first prediction confidence level for each of the one or more images;
select an image for which the confidence level of the at least one first prediction is greater than or equal to a first prediction threshold value;
determine, using a second computer vision algorithm that is different from the first computer vision algorithm, at least one second prediction of the object of interest and a corresponding second prediction confidence level for the selected image;
output the at least one second prediction and the second prediction confidence level for the selected image.

### Claim 2 (depends on 1)

The system of claim 1, wherein the first computer vision algorithm determines the first prediction and the corresponding first prediction confidence level for all of the one or more images.

### Claim 3 (depends on 1)

The system of claim 1, wherein the at least one first prediction comprises at least one of a first classification of the object of interest or localization of the object of interest.

### Claim 4 (depends on 1)

The system of claim 1, wherein the object of interest is a defect within the target.

### Claim 5 (depends on 1)

The system of claim 1, wherein the first computer vision algorithm is at least one of a first image classification algorithm, a single shot detection algorithm, or an object tracking algorithm.

### Claim 6 (depends on 1)

The system of claim 1, wherein the second prediction comprises at least one of a second classification of the object of interest or quantification of the object of interest.

### Claim 7 (depends on 1)

The system of claim 1, wherein the second computer vision algorithm is at least one of a region proposal network, an instance segmentation algorithm, or a semantic segmentation algorithm.

### Claim 8 (depends on 1)

The system of claim 1, wherein the camera is configured to transmit the captured one or more images to the controller immediately after capture.

### Claim 9 (depends on 1)

The system of claim 1, wherein the one or more images are sequential frames of a video captured by the camera or a time ordered sequence of still images.

### Claim 10 (depends on 1)

The system of claim 1, wherein the inspection system is a borescope.

### Claim 11

A method of inspecting a target, comprising:
capturing, by a camera of an inspection system, one or more images of a target;
receiving, by one or more processors of the inspection system, the one or more images from the camera;
determining, by the one or more processors using a first computer vision algorithm, a first prediction of an object of interest within the target and a corresponding first prediction confidence level for each of the one or more images;
selecting, by the one or more processors, an image of the one or more images for which the confidence level of the at least one first prediction is greater than a first prediction threshold value;
determining, by the one or more processors using a second computer vision algorithm that is different from the first computer vision algorithm, a second prediction of the object of interest and a corresponding second prediction confidence level for the selected image; and
outputting, by the one or more processors, the at least one second prediction and the second prediction confidence level for the selected image.

### Claim 12 (depends on 11)

The method of claim 11, further comprising determining the first prediction and the corresponding first prediction confidence level for all of the one or more images.

### Claim 13 (depends on 11)

The method of claim 11, wherein the at least one first prediction comprises at least one of a first classification of the object of interest or localization of the object of interest.

### Claim 14 (depends on 11)

The method of claim 11, wherein the object of interest is a defect within the target.

### Claim 15 (depends on 11)

The method of claim 11, wherein the first computer vision algorithm is at least one of a first image classification algorithm, a single shot detection algorithm, or an object tracking algorithm.

### Claim 16 (depends on 11)

The method of claim 11, wherein the second prediction comprises at least one of a second classification of the object of interest or quantification of the object of interest.

### Claim 17 (depends on 11)

The method of claim 11, wherein the second computer vision algorithm is at least one of a region proposal network, an instance segmentation algorithm, or a semantic segmentation algorithm.

### Claim 18 (depends on 11)

The method of claim 11, wherein the one or more images are received substantially immediately after capture by the camera.

### Claim 19 (depends on 11)

The method of claim 11, wherein the one or more images are sequential frames of a video captured by the camera or a time ordered sequence of still images.

### Claim 20 (depends on 11)

The method of claim 11, wherein the inspection system is a borescope.
