Fluorescent Coating Thickness Inspection Without Full OCT Scans

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for inspecting coating thickness, such as optical coherence tomography and fluorescence brightness, are inefficient or inaccurate, particularly in assessing coating thickness on objects with varying backgrounds.

Innovation Solution

An apparatus and method combining ultraviolet light, image sensors, and optical coherence tomography with machine learning models to determine coating thickness based on fluorescence brightness, compensating for background effects and using correlation models trained on learning data to enhance accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If optical coherence tomography is used to measure coating thickness, then measurement precision is improved, but productivity deteriorates due to scan time proportional to inspection area

Engineering Contradiction:
Improvecoating thickness measurement precisionVSAvoidinspection speed
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent creates a virtual copy of the coating thickness information by training a neural network model to predict OCT measurements from fluorescence brightness images. Once trained, the model generates thickness maps instantly without performing actual OCT scans, thus copying the measurement results without the time cost.

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The patent performs preliminary training of the correlation model using OCT scan data and fluorescence images before actual inspection. This preliminary action creates a predictive model that can instantly estimate thickness without requiring full OCT scans during production inspection.

Inventive Principle:
Principle #10Preliminary action

2Productivity

If fluorescence brightness is used to inspect coating presence, then productivity is improved, but measurement precision deteriorates as it cannot accurately determine coating thickness

Engineering Contradiction:
Improveinspection speedVSAvoidcoating thickness measurement precision
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent introduces a neural network correlation model as an intermediary that translates fluorescence brightness data into accurate thickness predictions. The model learns the complex relationship between fluorescence intensity and coating thickness, enabling thickness measurement without direct OCT scanning.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent transforms the measurement parameter from direct optical interference patterns (OCT) to fluorescence brightness intensity, using a trained model to map between these parameters. This allows using the simpler fluorescence method while achieving OCT-level thickness measurement accuracy.

Inventive Principle:
Principle #35Parameter changes

3Adaptability or versatility

If correlation models are trained on diverse learning data to improve accuracy across different backgrounds, then adaptability is improved, but device complexity increases

Engineering Contradiction:
Improvebackground variation adaptabilityVSAvoidmodel training complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent trains a single universal correlation model that can handle multiple background types and coating conditions. The model learns to generalize across different scenarios during training, making it adaptable to various backgrounds without requiring separate models for each condition.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

Accurately measures coating thickness in a short amount of time, even on objects with diverse backgrounds, by leveraging fluorescence brightness and optical coherence tomography, with high accuracy and efficiency.

Implementation Method 1

an object, which is coated with a coating comprising a fluorescent material, with ultraviolet rays

Methodology Applied
Scientific EffectFluorescence: Fluorescence

Implementation Method 2

control the optical coherence tomography module to obtain coating thicknesses measured through optical coherence tomography

Methodology Applied
Scientific EffectOptical coherence tomography: Tomography

Data Source

PatentEP4729881A1Apparatus and method for inspecting coating thickness and recording medium storing command
Publication Date: 2026.04.22 KOHYOUNG TECH
  • EP4729881A1 patent drawingFigure 1
  • EP4729881A1 patent drawingFigure 2
  • EP4729881A1 patent drawingFigure 3

AI summary

An apparatus according to an aspect of the present disclosure may comprise: an ultraviolet light source; an image sensor; an optical coherence tomography module; a memory storing one or more correlation models; and a processor. The processor may: control the ultraviolet light source to irradiate, with ultraviolet rays, an object to which a coating including a fluorescent material is applied; control the image sensor to obtain a fluorescent image of the object with which the ultraviolet rays are irradiated; determine, on the basis of the fluorescent image, fluorescence brightness of one or more sampling points included in the object; control the optical coherence tomography module to obtain coating thickness measured via optical coherence tomography at the one or more sampling points; train the one or more correlation models on the basis of first learning data including the fluorescence brightness of the one or more sampling points and the coating thickness measured at the one or more sampling points; determine, on the basis of the fluorescent image, a fluorescence brightness of a target area of the object; and determine, on the basis of the fluorescence brightness of the target area and the one or more correlation models, a coating thickness of the target area.