Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Comprehensive Survey on Machine Learning Driven Material Defect Detection

Jun Bai, Di Wu, Tristan Shelley, Peter Schubel, David Twine, John Russell, Xuesen Zeng, Ji Zhang

TL;DR

This survey addresses the rising role of machine learning in material defect detection (MDD) by organizing methods into unsupervised, supervised, semi-supervised, reinforcement, and generative learning, with a dedicated emphasis on composite materials. It provides a structured framework, comparing methodologies, datasets, and real-world performance while highlighting CM-specific challenges such as anisotropy and multimodal data fusion. Key contributions include a comprehensive taxonomy of techniques, a synthesis of benchmark datasets, and a discussion of open issues and future directions, including standardized datasets, small-sample learning, and real-time online adaptation. The work aims to guide both researchers and industry practitioners in selecting appropriate ML strategies for robust, scalable, and real-time MDD deployments, ultimately enhancing product quality, safety, and sustainability across sectors.

Abstract

Material defects (MD) represent a primary challenge affecting product performance and giving rise to safety issues in related products. The rapid and accurate identification and localization of MD constitute crucial research endeavors in addressing contemporary challenges associated with MD. In recent years, propelled by the swift advancement of machine learning (ML) technologies, particularly exemplified by deep learning, ML has swiftly emerged as the core technology and a prominent research direction for material defect detection (MDD). Through a comprehensive review of the latest literature, we systematically survey the ML techniques applied in MDD into five categories: unsupervised learning, supervised learning, semi-supervised learning, reinforcement learning, and generative learning. We provide a detailed analysis of the main principles and techniques used, together with the advantages and potential challenges associated with these techniques. Furthermore, the survey focuses on the techniques for defect detection in composite materials, which are important types of materials enjoying increasingly wide application in various industries such as aerospace, automotive, construction, and renewable energy. Finally, the survey explores potential future directions in MDD utilizing ML technologies. This survey consolidates ML-based MDD literature and provides a foundation for future research and practice.

A Comprehensive Survey on Machine Learning Driven Material Defect Detection

TL;DR

This survey addresses the rising role of machine learning in material defect detection (MDD) by organizing methods into unsupervised, supervised, semi-supervised, reinforcement, and generative learning, with a dedicated emphasis on composite materials. It provides a structured framework, comparing methodologies, datasets, and real-world performance while highlighting CM-specific challenges such as anisotropy and multimodal data fusion. Key contributions include a comprehensive taxonomy of techniques, a synthesis of benchmark datasets, and a discussion of open issues and future directions, including standardized datasets, small-sample learning, and real-time online adaptation. The work aims to guide both researchers and industry practitioners in selecting appropriate ML strategies for robust, scalable, and real-time MDD deployments, ultimately enhancing product quality, safety, and sustainability across sectors.

Abstract

Material defects (MD) represent a primary challenge affecting product performance and giving rise to safety issues in related products. The rapid and accurate identification and localization of MD constitute crucial research endeavors in addressing contemporary challenges associated with MD. In recent years, propelled by the swift advancement of machine learning (ML) technologies, particularly exemplified by deep learning, ML has swiftly emerged as the core technology and a prominent research direction for material defect detection (MDD). Through a comprehensive review of the latest literature, we systematically survey the ML techniques applied in MDD into five categories: unsupervised learning, supervised learning, semi-supervised learning, reinforcement learning, and generative learning. We provide a detailed analysis of the main principles and techniques used, together with the advantages and potential challenges associated with these techniques. Furthermore, the survey focuses on the techniques for defect detection in composite materials, which are important types of materials enjoying increasingly wide application in various industries such as aerospace, automotive, construction, and renewable energy. Finally, the survey explores potential future directions in MDD utilizing ML technologies. This survey consolidates ML-based MDD literature and provides a foundation for future research and practice.
Paper Structure (60 sections, 5 figures, 7 tables)

This paper contains 60 sections, 5 figures, 7 tables.

Table of Contents

  1. Introduction
  2. Preliminaries and Problem Identification
  3. Acronyms and Their Definitions
  4. Defect Detection
  5. Material Defect Detection
  6. Experimental Evaluation and Benchmarks
  7. Challenges in Material Defect Detection
  8. Survey Overview
  9. Techniques for Material Defect Detection
  10. Unsupervised Learning Techniques in Material Defect Detection
  11. Feature-Based Techniques
  12. Histogram-Based
  13. Principal Component Analysis (PCA)-Based
  14. Self-Organizing Maps (SOMs)-Based
  15. Clustering-Based Techniques
  16. ...and 45 more sections

Figures (5)

  • Figure 1: Significant technical milestones of defect detection based on ML.
  • Figure 2: Illustration of a material defect detection system based on machine vision.
  • Figure 3: Common machine learning methods for conducting the material defect detection.
  • Figure 4: Illustration of the general working process for conducting the material defect detection with UL techniques.
  • Figure 5: Illustration of the general working process for conducting the material defect detection with supervised learning techniques.