# US 2021/0286133 A1 · Device Of Detection Of Surface Defects On At Least One Terminal Surface Of At Least One Optical Fiber

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

- Number: US 2021/0286133 A1
- Document ID: US20210286133A1
- Status: Published application
- Kind: A1
- Application number: 17/191,906
- Filed: 2021-03-04
- Published: 2021-09-16
- Earliest claimed: 2020-03-10
- Assignees: Data-Pixel
- Applicants: Data-Pixel
- Inventors: Yannick LAPERRIERE; Julien MAILLE; Eric TANGUY
- Web page: https://patentgrep.com/patent/US20210286133A1

## CPC

- G02B 6/385: {Accessories for testing or observation of connectors (means for centering or aligning the light guide within the ferrule with auxiliary facilities for movably aligning or adjusting the fibre within its ferrule, measuring position, eccentricity G02B6/3843; mechanical features associated with the optical testing of optical fibres G01M11/088)}
- G02B 21/0004: {specially adapted for specific applications}
- G01N 21/8851: {Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges (G01N21/8806 and G01N21/93 - G01N21/95692 take precedence; optical measurement of dimensions G01B11/00; optical scanning G02B26/10; image transformation G06T3/00; computerised image enhancement G06T5/00; image processing per se for flaw detection G06T7/0002)}
- G01N 21/94: Investigating contamination, e.g. dust (G01N21/85 takes precedence)
- G02B 21/365: {Control or image processing arrangements for digital or video microscopes (G02B21/361, G02B21/362 take precedence)}
- G06T 7/0004: {Industrial image inspection}
- G06T 7/136: involving thresholding
- G01N 2021/8883: {involving the calculation of gauges, generating models}
- G01N 2021/9511: {Optical elements other than lenses, e.g. mirrors}
- G06T 2207/10: Image acquisition modality
- G06T 2207/10056: Microscopic image
- G06T 2207/20081: Training; Learning
- G06T 2207/20084: Artificial neural networks [ANN]
- G06T 2207/30108: Industrial image inspection
- G02B: OPTICAL ELEMENTS, SYSTEMS OR APPARATUS
- G01N: INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES (measuring or testing apparatus or processes other than immunoassay, involving enzymes or microorganisms C12M, C12Q)
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

A device for detection of surface defects on a terminal surface of an optical fiber. The device includes a digital microscope configured to capture an image of a terminal surface; and a mechanism for analyzing the image configured to detect surface defects present on the terminal surface, the analysis mechanism integrating a “U-Net”-type neural network having had its training phase carried out via an enhancer. The enhancer is configured to create training images, intended to train the neural network, based on reference images. The training images are obtained from the reference images by only applying flips, rotations and/or luminosity, contrast, or shade variations to the reference images.

## Claims

### Claim 1

A surface defect detection device for detecting surface defects on at least one terminal surface of at least one optical fiber, said device comprising:
a digital microscope configured to capture an image of said at least one terminal surface; and
means for analyzing said image configured to detect surface defects present on said at least one terminal surface;
characterized in that said analysis means integrate a “U-Net”-type neural network having had its training phase carried out via an enhancer; said enhancer is configured to create training images, intended to train said neural network, based on reference images; and said training images are obtained from said reference images by only applying flips, rotations and/or luminosity, contrast, or shade variations to said reference images.

### Claim 2 (depends on 1)

The surface defect detection device according to claim 1, wherein one or a plurality of optical fibers is gathered in a connector, said digital microscope is configured to capture an image of the terminal surfaces of the optical fibers at the level of said connector; and said analysis means are configured to isolate the terminal surfaces of the optical fibers.

### Claim 3 (depends on 1)

The surface defect detection device according to claim 1, wherein said “U-Net”-type neural network comprises a contracting path and an expansive path with five processing layers each.

### Claim 4 (depends on 3)

The surface defect detection device according to claim 3, wherein said contracting path and said expansive path implement, for each processing layer, two processings performing a convolution of three pixels by three pixels followed by a linear rectification.

### Claim 5 (depends on 3)

The surface defect detection device according to claim 3, wherein said contracting path implements, between each processing layer, a processing performing a linear concatenation of the local maximum values.

### Claim 6 (depends on 3)

The surface defect detection device according to claim 3, wherein said expansive path implements, between each processing layer, a processing performing an up-convolution of two pixels by two pixels.

### Claim 7 (depends on 1)

The surface defect detection device according to claim 1, wherein the number of said reference images is in the range from 200 to 400, and the enhancer is configured to obtain a number of training images greater than 1,000.

### Claim 8 (depends on 1)

The surface defect detection device according to claim 1, wherein said enhancer randomly chooses to apply or not flips, rotations, and/or luminosity, contrast, or shade variations to said reference images.

### Claim 9 (depends on 1)

The surface defect detection device according to claim 1, wherein said enhancer randomly selects the parameters of the flips, rotations, and/or luminosity, contrast or shade variations applied to said reference images.