Scholarlink

<p>In this study, a new image segmentation framework for infrared thermographic inspection, utilizing texture encoding integrated with thermal constraints, is developed to enhance defect localization. The method, namely, advanced local binary pattern with thermal constraint (ALBP-TC), incorporates median-filtered grayscale pre-processing, ALBP-based texture feature extraction, and adaptive thermal thresholding. These elements are combined by using an extraction and adaptive thresholding, and the segmentation accuracy is improved by using a weighted approach. To automate parameter selection and maximize diagnostic performance, a non-dominated sorting genetic algorithm-II (NSGA-II) is employed to optimize thermal threshold (<InlineEquation ID="IEq1"> <EquationSource Format="TEX">\({T}_{t}\)</EquationSource> <EquationSource Format="MATHML"><math> <msub> <mi>T</mi> <mi>t</mi> </msub> </math></EquationSource> </InlineEquation>) and fusion weight (<InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(\omega\)</EquationSource> <EquationSource Format="MATHML"><math> <mi>ω</mi> </math></EquationSource> </InlineEquation>) by using multi-objective criteria by considering multi-performance parameters, namely, Accuracy, Dice similarity index (DSI), and Jaccard similarity index (JSI), etc. The experimental validation of thermal images of roller bearings demonstrates a robust condition monitoring ability by achieving 99.16% accuracy, 100% sensitivity, and 99.15% specificity, with DSI and JSI measures of 82% and 69.48%, respectively. A comparative analysis of the proposed method with other experimental methods from recent literature confirms the superiority of the proposed method in terms of accuracy and spatial coherence.</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

Fault Diagnosis in Roller Bearings by Using an Advanced Artificial Intelligence-Based Metrology

  • 期刊论文

  • 发表日期：

    2026年1月27日

查看全文
  • Ekta Yadav,
  • V. K. Chawla,
  • Navdeep,
  • Sanjay Yadav

摘要

In this study, a new image segmentation framework for infrared thermographic inspection, utilizing texture encoding integrated with thermal constraints, is developed to enhance defect localization. The method, namely, advanced local binary pattern with thermal constraint (ALBP-TC), incorporates median-filtered grayscale pre-processing, ALBP-based texture feature extraction, and adaptive thermal thresholding. These elements are combined by using an extraction and adaptive thresholding, and the segmentation accuracy is improved by using a weighted approach. To automate parameter selection and maximize diagnostic performance, a non-dominated sorting genetic algorithm-II (NSGA-II) is employed to optimize thermal threshold ( \({T}_{t}\) T t ) and fusion weight ( \(\omega\) ω ) by using multi-objective criteria by considering multi-performance parameters, namely, Accuracy, Dice similarity index (DSI), and Jaccard similarity index (JSI), etc. The experimental validation of thermal images of roller bearings demonstrates a robust condition monitoring ability by achieving 99.16% accuracy, 100% sensitivity, and 99.15% specificity, with DSI and JSI measures of 82% and 69.48%, respectively. A comparative analysis of the proposed method with other experimental methods from recent literature confirms the superiority of the proposed method in terms of accuracy and spatial coherence.