Scholarlink

Abstract <p>Glass fiber reinforced composite (GFRP) plays a vital role in aviation, automobile, and electronics due to its low weight, high corrosion resistance, and long life. Monitoring the condition of GFRP materials is an important task in many applications as it greatly enhances safety, serviceability, and operating limits. Non destructive testing and evaluation (NDT&amp;E) is becoming increasingly popular for condition monitoring (damage detection and characterization) of material in various fields because of its noncontact and nonintrusive nature. There are several NDT&amp;E methods available to investigate defects, flaws, voids, etc, in materials/systems. However, only a few methods satisfy the requirements to be noncontact and noninvasive in industrial applications. Frequency modulated thermal wave imaging (FMWTI) is a promising approach to detect delamination defects in GFRP. However, processing recorded thermograms is essential in NDT&amp;E to improve data usefulness. The thermograms are contaminated with noise. To visualize and extract features of the subsurface defects in the sample, temporal variations in contrast to each pixel with respect to the reference point at the non-defective area on the sample are to be analyzed. This paper proposes a new methodology to analyze raw data sequences using image fusion and segmentation. In the present work, a new segmentation approach-based clustering has been proposed, where defective regions are grouped into clusters based on their mutual-relationship.The experimental study was performed on a GFRP sample with Teflon inserts of various sizes placed at different depths. Relative foreground area error (RAE) is used to evaluate the performance of the proposed method. Comparison between existing and proposed methods has been made in terms of RAE and number of detects.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Novel Segmentation Based Clustering Technique for Defect detection in GFRP using Frequency Modulated Thermal Wave Imaging

  • 期刊论文

  • 发表日期：

    2025年11月1日

查看全文
  • Kante Murali,
  • D. V. Rama Koti Reddy,
  • Bhargav Appasani

摘要

Abstract

Glass fiber reinforced composite (GFRP) plays a vital role in aviation, automobile, and electronics due to its low weight, high corrosion resistance, and long life. Monitoring the condition of GFRP materials is an important task in many applications as it greatly enhances safety, serviceability, and operating limits. Non destructive testing and evaluation (NDT&E) is becoming increasingly popular for condition monitoring (damage detection and characterization) of material in various fields because of its noncontact and nonintrusive nature. There are several NDT&E methods available to investigate defects, flaws, voids, etc, in materials/systems. However, only a few methods satisfy the requirements to be noncontact and noninvasive in industrial applications. Frequency modulated thermal wave imaging (FMWTI) is a promising approach to detect delamination defects in GFRP. However, processing recorded thermograms is essential in NDT&E to improve data usefulness. The thermograms are contaminated with noise. To visualize and extract features of the subsurface defects in the sample, temporal variations in contrast to each pixel with respect to the reference point at the non-defective area on the sample are to be analyzed. This paper proposes a new methodology to analyze raw data sequences using image fusion and segmentation. In the present work, a new segmentation approach-based clustering has been proposed, where defective regions are grouped into clusters based on their mutual-relationship.The experimental study was performed on a GFRP sample with Teflon inserts of various sizes placed at different depths. Relative foreground area error (RAE) is used to evaluate the performance of the proposed method. Comparison between existing and proposed methods has been made in terms of RAE and number of detects.