Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Deep Learning for Micro-Scale Crack Detection on Imbalanced Datasets Using Key Point Localization

Fatahlla Moreh, Yusuf Hasan, Bilal Zahid Hussain, Mohammad Ammar, Sven Tomforde

Abstract

Internal crack detection has been a subject of focus in structural health monitoring. By focusing on crack detection in structural datasets, it is demonstrated that deep learning (DL) methods can effectively analyze seismic wave fields interacting with micro-scale cracks, which are beyond the resolution of conventional visual inspection. This work explores a novel application of DL-based key point detection technique, where cracks are localized by predicting the coordinates of four key points that define a bounding region of the crack. The study not only opens new research directions for non-visual applications but also effectively mitigates the impact of imbalanced data which poses a challenge for previous DL models, as it can be biased toward predicting the majority class (non-crack regions). Popular DL techniques, such as the Inception blocks, are used and investigated. The model shows an overall reduction in loss when applied to micro-scale crack detection and is reflected in the lower average deviation between the location of actual and predicted cracks, with an average Intersection over Union (IoU) being 0.511 for all micro cracks (greater than 0.00 micrometers) and 0.631 for larger micro cracks (greater than 4 micrometers).

Deep Learning for Micro-Scale Crack Detection on Imbalanced Datasets Using Key Point Localization

Abstract

Internal crack detection has been a subject of focus in structural health monitoring. By focusing on crack detection in structural datasets, it is demonstrated that deep learning (DL) methods can effectively analyze seismic wave fields interacting with micro-scale cracks, which are beyond the resolution of conventional visual inspection. This work explores a novel application of DL-based key point detection technique, where cracks are localized by predicting the coordinates of four key points that define a bounding region of the crack. The study not only opens new research directions for non-visual applications but also effectively mitigates the impact of imbalanced data which poses a challenge for previous DL models, as it can be biased toward predicting the majority class (non-crack regions). Popular DL techniques, such as the Inception blocks, are used and investigated. The model shows an overall reduction in loss when applied to micro-scale crack detection and is reflected in the lower average deviation between the location of actual and predicted cracks, with an average Intersection over Union (IoU) being 0.511 for all micro cracks (greater than 0.00 micrometers) and 0.631 for larger micro cracks (greater than 4 micrometers).

Paper Structure

This paper contains 18 sections, 3 equations, 5 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Method
  4. Wide Convolutional Netoworks
  5. Model Architecture
  6. Loss Functions
  7. Training Procedure
  8. Convolutional Blocks
  9. Datasets
  10. Metrics
  11. Intersection Over Union:
  12. Comparison with the previous model
  13. A Novel Approach for Crack Detection in Numerical, Non-Visual Data Using Deep Learning
  14. Our Keypoint-Based Approach
  15. Crack Detection in Numerical, Non-Visual Data
  16. ...and 3 more sections

Figures (5)

  • Figure 1: Proposed model architecture for key point localization
  • Figure 2: Sample labels for the dataset
  • Figure 3: Predicted results where red bounding boxes are predicted and green bounding boxes are ground truth
  • Figure 4: Predicted results where red bounding boxes are predicted and green bounding boxes are ground truth
  • Figure 5: Predicted results where red bounding boxes are predicted and green bounding boxes are ground truth