Parallel Spectral Unmixing for Coating Identification

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional methods for identifying and matching coating compositions, such as those used in automotive coatings, are inefficient and prone to errors due to their reliance on brute force approaches and neural networks, which are slow, inflexible, and resource-intensive, often resulting in inaccurate and time-consuming processes.

Innovation Solution

A system and method for parallel processing of spectrometric data using probabilistic colorant analysis, which initiates multiple independent decision points to calculate the probability of different pigment types in a target coating, allowing for accurate and reproducible identification of final colorant probabilities and formulation of matching coatings.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional brute force approaches and neural networks are used for coating composition identification, then comprehensive analysis can be performed, but processing time increases and accuracy decreases

Engineering Contradiction:
Improvecoating identification accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent segments the coating identification process into distinct spectral region analyses (UV, visible, IR regions processed separately by different experts), allowing parallel processing of multiple aspects simultaneously. This segmentation enables comprehensive analysis without requiring sequential brute force evaluation of all possible pigment combinations.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies preliminary action by using spectral unmixing and expert analysis to identify candidate pigments before final formulation determination. This preliminary identification narrows down the search space significantly, avoiding time-consuming exhaustive searches while maintaining high accuracy in pigment detection.

Inventive Principle:
Principle #10Preliminary action

2Adaptability or versatility

If conventional neural network systems are used, then coating analysis can be performed, but the systems are inflexible and resource-intensive

Engineering Contradiction:
Improvesystem flexibilityVSAvoidsystem complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent implements a universal expert system framework that can handle multiple coating types (automotive, industrial, artistic) and various spectral regions (UV, visible, IR) using the same core methodology. This multi-functional approach provides flexibility across different applications without requiring separate specialized systems for each coating type.

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

Solution Approach 2:

The patent replaces conventional neural network mechanical systems with a knowledge-based expert system that uses spectral unmixing algorithms and rule-based reasoning. This substitution eliminates the inflexibility and high resource consumption of neural networks while providing transparent, interpretable analysis through explicit spectral decomposition methods.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Reliability

If brute force approaches are used for pigment identification, then all possible combinations can be evaluated, but the process becomes error-prone and time-consuming

Engineering Contradiction:
Improveidentification reliabilityVSAvoidanalysis speed
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent performs preliminary spectral unmixing to decompose the target coating spectrum into constituent pigment spectra before final identification. This preliminary decomposition reliably identifies present pigments by comparing unmixed spectral components against a database, avoiding errors from evaluating all possible pigment combinations while maintaining high analysis speed.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent creates spectral copies or representations of candidate pigments from a database and compares these copied spectral signatures against the unmixed target spectrum. This copying approach enables reliable identification through pattern matching without requiring physical manipulation or exhaustive testing of all pigment combinations.

Inventive Principle:
Principle #26Copying

Data Source

PatentEP3785181B1Systems and methods for rapid coating composition determinations
Publication Date: 2024.03.27 PPG INDUSTRIES OHIO INC
  • EP3785181B1 patent drawingFigure 1~2
  • EP3785181B1 patent drawingFigure 3~4
  • EP3785181B1 patent drawingFigure 5~6

AI summary

Modern coatings provide several important functions in industry and society. Coatings can protect a coated material from corrosion, such as rust. Coatings can also provide an aesthetic function by providing a particular color and/or texture to an object. For example, most automobiles are coated using paints and various other coatings in order to protect the metal body of the automobile from the elements and also to provide aesthetic visual effects. In view of the wide-ranging uses for different coatings, it is often necessary to identify a target coating composition. For instance, it might be necessary to identify a target coating on an automobile that has been in an accident. If the target coating is not properly identified, any resulting repair to the automobile's coating will not match the original coating. As used herein, a target coating comprises any coating of interest that has been applied to any physical object. There are many opportunities for new methods and systems that improve the identification of coatings.