Related Experiment Video
Updated: Sep 3, 2025
05:48
Author Spotlight: Investigating the Impact of Emotional Prosodies on Voice Recognition and Perception
Published on: August 9, 2024
1.6K
A Study on the Applicability of the Impact-Echo Test Using Semi-Supervised Learning Based on Dynamic Preconditions
Young-Geun Yoon1, Chung-Min Kim1, Tae-Keun Oh1
1Department of Safety Engineering, Incheon National University, Incheon 22012, Korea.
Sensors (Basel, Switzerland)
|July 28, 2022
Summary
Semi-supervised learning (SSL) improves concrete crack detection using the Impact-Echo (IE) test. SSL enhances accuracy by 7-8% over supervised learning, offering better defect classification for structures.
Area of Science:
- Civil Engineering
- Materials Science
- Artificial Intelligence
Background:
- The Impact-Echo (IE) test is a non-destructive method for concrete defect detection, including cracks and delamination.
- Advancements in non-contact sensors enable rapid, multi-point IE testing, generating large datasets.
- Supervised learning (SL) faces challenges with accurate labeling and new specimen characteristics in large IE datasets.
Purpose of the Study:
- To evaluate the accuracy and applicability of semi-supervised learning (SSL) for analyzing air-coupled IE test data.
- To compare the performance of SSL models against SL models for concrete defect detection.
- To investigate the use of dynamic behavior analysis and principal component analysis (PCA) in conjunction with SSL.
Main Methods:
- Utilized air-coupled Impact-Echo (IE) testing with dynamic preconditions.
- Extracted 21 features in time and frequency domains, focusing on flexural modes for delamination detection.
- Applied Principal Component Analysis (PCA) to identify key features (real moment, real RMS, imaginary moment).
- Trained and evaluated both supervised learning (SL) and semi-supervised learning (SSL) models.
Main Results:
- SSL models demonstrated a 7-8% increase in accuracy compared to SL models for defect detection.
- Key principal components (PCs) derived from PCA were consistent across different concrete structures (slab, pavement, deck).
- SSL enabled a higher level of categorization for structural conditions (good, fair, poor).
Conclusions:
- Semi-supervised learning (SSL) is applicable and effective for analyzing Impact-Echo (IE) test data, outperforming supervised learning (SL).
- SSL enhances the accuracy and classification capabilities for concrete defect detection, particularly delamination.
- Future research should incorporate additional parameters to account for crack progression under varying field conditions.
Related Concept Videos
01:19Survival Tree
152
Survival trees are a non-parametric method used in survival analysis to model the relationship between a set of covariates and the time until an event of interest occurs, often referred to as the "time-to-event" or "survival time." This method is particularly useful when dealing with censored data, where the event has not occurred for some individuals by the end of the study period, or when the exact time of the event is unknown.
Building a Survival Tree
Constructing a...
Building a Survival Tree
Constructing a...
152
01:17Hypothesis: Accept or Fail to Reject?
28.3K
The outcome of any hypothesis testing leads to rejecting or not rejecting the null hypothesis. This decision is taken based on the analysis of the data, an appropriate test statistic, an appropriate confidence level, the critical values, and P-values. However, when the evidence suggests that the null hypothesis cannot be rejected, is it right to say, 'Accept' the null hypothesis?
There are two ways to indicate that the null hypothesis is not rejected. 'Accept' the null...
There are two ways to indicate that the null hypothesis is not rejected. 'Accept' the null...
28.3K