Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Evaluation of signal disturbance and recovery in phased array ultrasonic inspection during welding.

Welding in the world·2026
Same author

Ultrasonic phased array measurement & compression for in-process weld bevel estimation.

Ultrasonics·2026
Same author

Urethral-sparing surgical management of primary urethral adenocarcinoma in a female patient with successful oncological outcome.

BMJ case reports·2025
Same author

Single-Bit Reception With Coded Excitation for Lightweight Advanced Ultrasonic Imaging Systems.

IEEE transactions on ultrasonics, ferroelectrics, and frequency control·2024
Same author

Unsupervised machine learning for flaw detection in automated ultrasonic testing of carbon fibre reinforced plastic composites.

Ultrasonics·2024
Same author

Three-Dimensional Residual Neural Architecture Search for Ultrasonic Defect Detection.

IEEE transactions on ultrasonics, ferroelectrics, and frequency control·2024

Related Experiment Video

Updated: Jun 24, 2025

Deep Learning-Based Segmentation of Cryo-Electron Tomograms
10:25

Deep Learning-Based Segmentation of Cryo-Electron Tomograms

Published on: November 11, 2022

8.7K

Advancing Carbon Fiber Composite Inspection: Deep Learning-Enabled Defect Localization and Sizing via 3-D U-Net

Shaun McKnight, Vedran Tunukovic, S Gareth Pierce

    IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control
    |June 3, 2024
    PubMed
    Summary
    This summary is machine-generated.

    A new deep learning method accurately locates and sizes defects in carbon fiber composites using 3-D ultrasonic data. This approach improves defect characterization, reducing reporting burdens for nondestructive evaluation operators.

    More Related Videos

    Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
    05:56

    Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

    Published on: April 14, 2023

    2.4K
    Detection and Quantification of Tunneling Nanotubes Using 3D Volume View Images
    12:45

    Detection and Quantification of Tunneling Nanotubes Using 3D Volume View Images

    Published on: August 31, 2022

    2.9K

    Related Experiment Videos

    Last Updated: Jun 24, 2025

    Deep Learning-Based Segmentation of Cryo-Electron Tomograms
    10:25

    Deep Learning-Based Segmentation of Cryo-Electron Tomograms

    Published on: November 11, 2022

    8.7K
    Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
    05:56

    Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

    Published on: April 14, 2023

    2.4K
    Detection and Quantification of Tunneling Nanotubes Using 3D Volume View Images
    12:45

    Detection and Quantification of Tunneling Nanotubes Using 3D Volume View Images

    Published on: August 31, 2022

    2.9K

    Area of Science:

    • Materials Science
    • Artificial Intelligence
    • Nondestructive Evaluation

    Background:

    • Accurate defect characterization is crucial for assessing component integrity in nondestructive evaluation (NDE).
    • Traditional methods for defect sizing and localization in composites can be labor-intensive and may lack precision.
    • Ultrasonic testing (UT) is a common NDE technique, but analyzing volumetric data for precise defect metrics remains challenging.

    Purpose of the Study:

    • To introduce a novel deep learning (DL) methodology for precise defect localization and sizing in carbon fiber reinforced polymer (CFRP) composites.
    • To leverage 3-D U-Net for volumetric segmentation of UT data, enabling automated defect analysis.
    • To compare the DL approach against the industry-standard 6 dB drop method for defect characterization.

    Main Methods:

    • Development of a 3-D U-Net model for volumetric segmentation of UT data from CFRP components.
    • Generation of synthetic training data with ground truth segmentation masks representative of experimental UT data.
    • Validation against 40 fabricated defects, comparing DL performance with the conventional 6 dB drop analysis.

    Main Results:

    • Excellent through-thickness localization accuracy (0.08 mm MAE) and good in-plane localization (0.57 mm MAE) achieved by the DL model.
    • Initial DL sizing overestimates defects, but a correction factor significantly improves accuracy, reducing MAE by 35% compared to the 6 dB drop method.
    • The volumetric approach eliminates the need for extensive preprocessing like signal gating.

    Conclusions:

    • The 3-D U-Net methodology offers a robust and automated solution for defect localization and sizing in CFRP composites using UT data.
    • This DL approach enhances accuracy, particularly in through-thickness localization, and streamlines the NDE process.
    • The ability to generate 3-D defect masks facilitates direct use in CAD systems, significantly reducing reporting workload.