Unlock instant, AI-driven research and patent intelligence for your innovation.

Composite material defect detection method based on thermal image analysis of generative nuclear principal components

A composite material and defect detection technology, used in material defect testing, image analysis, neural learning methods, etc., can solve problems such as limiting the ability of models to perform expected results, and achieve the effect of improving visibility

Active Publication Date: 2022-06-03
ZHEJIANG UNIV OF TECH
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

A single test produces only a few dozen thermal images, which often limits the model's ability to perform as expected

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Composite material defect detection method based on thermal image analysis of generative nuclear principal components
  • Composite material defect detection method based on thermal image analysis of generative nuclear principal components
  • Composite material defect detection method based on thermal image analysis of generative nuclear principal components

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0078] IRT was performed on artificially fabricated CFRP specimens containing multiple defects of varying shapes and depths. running pulse

[0080] Step 2.1: The SNGAN generation network G sets four deconvolution layers, and the first three layers use a linear unit (ReLU) excitation

[0081]

[0084] is the weight matrix after spectral normalization.

[0085] The SNGAN model converges until the discriminator D cannot discriminate the fake thermal image generated by the generator G. Finally, the trained

[0088]

[0091] x

[0092] x

[0094]

[0097] x

[0099] Step 3.1: Feature Space Mapping:

[0101]

[0104] m is the number of samples, m=n

[0105] Step 3.2: Calculation of the projection matrix T:

[0107] Kw=λw

[0109] w is the corresponding eigenvector matrix.

[0111]

[0113] S

[0117]

[0125] The signal-to-noise ratio is the contrast between defective and non-defective areas. The higher the signal-to-noise ratio, the better the ability of the method to identify defect...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a composite material defect detection method based on generating nuclear principal component thermal image analysis, and belongs to the technical field of thermal imaging non-destructive detection of composite materials. It includes the following steps: Step 1. Acquire thermal image data sets of composite materials; Step 2, amplification and preprocessing of thermal image data: establish spectral normalization to generate adversarial networks to generate thermal images; Step 3, establish kernel principal component analysis Model: feature space mapping and projection matrix calculation; step 4, image reconstruction and defect visualization; step 5, model performance evaluation. The present invention adopts the data amplification strategy based on generative confrontation network and the nonlinear dimensionality reduction technology based on kernel mapping to analyze the thermal image data with nonlinear characteristics; when the original thermal image data is less, the thermal image distribution is generated and experimentally recorded The same data; the nuclear principal component thermal imaging analysis model is used to solve the problem of difficult separation of defects and background in thermal image analysis, and improve the visibility of defects.

Description

Defect detection method of composite materials based on thermal image analysis of generative core principal components technical field The invention belongs to the nondestructive testing technical field of composite material thermal imaging, be specifically related to the thermal A composite defect detection method based on image analysis. Background technique Infrared thermal imaging camera (IRT) has the advantages of fast response, wide measurement range, intuitive results, etc., and is widely used in composite In the quality evaluation of non-destructive testing of material products or structures. However, especially in the defect assessment task, due to the experimental setting and environmental influences, it is difficult for IRT to achieve the expected detection results. Background inhomogeneity and noise in infrared thermal images are often The main factors that affect the detection accuracy and efficiency. Therefore, it has become common to employ data ana...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G01N25/72G06T7/00G06N3/04G06N3/08
CPCG01N25/72G06T7/0002G06N3/08G06T2207/10048G06T2207/20068G06N3/045Y02E30/30
Inventor 刘毅刘凯新娄维尧汤宇炜杨建国
Owner ZHEJIANG UNIV OF TECH