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

Generating R2<sup>*</sup> maps from T1W and T2W images using image-to-image translation for Parkinson's disease.

Medical physics·2026
Same author

Deep-learning based electroencephalogram denoising: A literature review.

Journal of neural engineering·2026
Same author

The discovery of novel 3-nitropyridin-2-amine-based 2,3-dihydroimidazo[1,2-a]pyridin-5(1H)-one insecticides.

Pest management science·2026
Same author

The Discovery of HNPC-A0073: A Novel Thiazol-5-ylmethyl-Based 3-Nitropyridin-2-amine Fungicide.

Journal of agricultural and food chemistry·2026
Same author

Lineage tracing of soma-to-primordial germ cell-like conversion in human tumor cell line.

iScience·2026
Same author

Antibacterial activity and multi-target mechanism of naringenin against Alicyclobacillus acidoterrestris and its application in apple juice.

International journal of food microbiology·2026

Related Experiment Video

Updated: Sep 11, 2025

Surrogate Model Development for Digital Experiments in Welding
09:17

Surrogate Model Development for Digital Experiments in Welding

Published on: March 28, 2025

1.2K

A Welding Defect Detection Model Based on Hybrid-Enhanced Multi-Granularity Spatiotemporal Representation Learning.

Chenbo Shi1, Shaojia Yan1, Lei Wang1

  • 1College of lntelligent Equipment, Shandong University of Science and Technology, Taian 271019, China.

Sensors (Basel, Switzerland)
|August 14, 2025
PubMed
Summary

This study introduces a novel algorithm for real-time molten pool monitoring in automated welding, enhancing quality control by accurately identifying defects and improving model interpretability for intelligent manufacturing.

Keywords:
deep learningimage interferencekey interference-free framesmulti-granularity spatiotemporal featuresporosity defect

More Related Videos

Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography
11:34

Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography

Published on: May 15, 2017

11.2K
Author Spotlight: Efficient Image Recognition Using Directional Gradient Histogram Technique and Support Vector Machines
08:27

Author Spotlight: Efficient Image Recognition Using Directional Gradient Histogram Technique and Support Vector Machines

Published on: January 5, 2024

1.2K

Related Experiment Videos

Last Updated: Sep 11, 2025

Surrogate Model Development for Digital Experiments in Welding
09:17

Surrogate Model Development for Digital Experiments in Welding

Published on: March 28, 2025

1.2K
Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography
11:34

Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography

Published on: May 15, 2017

11.2K
Author Spotlight: Efficient Image Recognition Using Directional Gradient Histogram Technique and Support Vector Machines
08:27

Author Spotlight: Efficient Image Recognition Using Directional Gradient Histogram Technique and Support Vector Machines

Published on: January 5, 2024

1.2K

Area of Science:

  • Materials Science and Engineering
  • Robotics and Automation
  • Computer Vision and Machine Learning

Background:

  • Real-time molten pool monitoring is crucial for high-quality automated welding.
  • Existing methods face challenges with image interference (e.g., spatter reflections) and limited deep learning interpretability.
  • Accurate defect detection, such as distinguishing spatter from porosity, is essential for intelligent welding systems.

Purpose of the Study:

  • To develop a robust algorithm for real-time molten pool quality monitoring in complex welding scenarios.
  • To address interference issues in molten pool images and enhance the interpretability of deep learning models.
  • To achieve a balance between accuracy, inference speed, and interpretability for intelligent automated welding.

Main Methods:

  • Proposed a multi-granularity spatiotemporal representation learning algorithm combining handcrafted and deep learning features.
  • Utilized a MobileNetV2 backbone with Temporal Shift Module (TSM) for capturing dynamic molten pool features and temporal information.
  • Implemented a multi-granularity attention-based feature aggregation module with cross-frame attention and Convolutional Block Attention Module (CBAM).
  • Integrated handcrafted Histogram of Oriented Gradient (HOG) and Scale-Invariant Feature Transform (SIFT) features for enhanced interpretability.

Main Results:

  • Achieved a high accuracy of 99.187% on a self-constructed dataset for molten pool quality monitoring.
  • Attained a real-time inference speed of 20.983 ms per sample on a platform with Intel i9 CPU and RTX 3060 GPU.
  • Demonstrated effective balancing of accuracy, speed, and interpretability in the proposed method.

Conclusions:

  • The proposed hybrid feature learning algorithm effectively addresses interference and interpretability issues in molten pool monitoring.
  • The method enables accurate and efficient real-time quality assessment for intelligent automated welding systems.
  • This approach offers a promising solution for advancing the capabilities of automated welding technology.