Scholarlink

<p>Terrestrial Laser Scanning (TLS) provides high-resolution 3D point clouds, revolutionizing structural integrity assessment. This study presents a comprehensive, automated framework for the direct analysis of these point clouds to quantify and evaluate structural damage. Unlike conventional approaches that rely on singular features or manual inspections, our method leverages a synergistic combination of well-established algorithms. Specifically, we use Principal Component Analysis (PCA) to derive a novel suite of eight geometric features that capture subtle surface irregularities and complex damage patterns with high fidelity. A unique contribution lies in the subsequent application of K-means clustering, which effectively segments these features into damaged and undamaged regions, enabling the direct computation of a new quantitative metric: the Damage Severity Index (DSI). This index is then visualized through spatial heat maps, providing a powerful and intuitive tool for assessment. The proposed framework demonstrates its exceptional capability in identifying subtle damage zones by leveraging inherent geometric characteristics, an advance over methods relying on manual inspection or intensity analysis. The accuracy of this approach was then validated using metrics including sensitivity, precision, and F1-score. On two representative wall datasets, the method achieved F1-scores of 0.85 and 0.67, confirming a high degree of precision in our segmentation and classification results. Thus, this work demonstrates that a tailored integration of existing methods can yield a robust, automated tool for structural integrity assessment and informed restoration planning, offering a significant improvement in efficiency and reliability.</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

Geometric features extraction and automated classification for damage assessment using LiDAR data

  • 期刊论文

  • 发表日期：

    2026年2月11日

查看全文
  • M R Apoorva,
  • N Vaani

摘要

Terrestrial Laser Scanning (TLS) provides high-resolution 3D point clouds, revolutionizing structural integrity assessment. This study presents a comprehensive, automated framework for the direct analysis of these point clouds to quantify and evaluate structural damage. Unlike conventional approaches that rely on singular features or manual inspections, our method leverages a synergistic combination of well-established algorithms. Specifically, we use Principal Component Analysis (PCA) to derive a novel suite of eight geometric features that capture subtle surface irregularities and complex damage patterns with high fidelity. A unique contribution lies in the subsequent application of K-means clustering, which effectively segments these features into damaged and undamaged regions, enabling the direct computation of a new quantitative metric: the Damage Severity Index (DSI). This index is then visualized through spatial heat maps, providing a powerful and intuitive tool for assessment. The proposed framework demonstrates its exceptional capability in identifying subtle damage zones by leveraging inherent geometric characteristics, an advance over methods relying on manual inspection or intensity analysis. The accuracy of this approach was then validated using metrics including sensitivity, precision, and F1-score. On two representative wall datasets, the method achieved F1-scores of 0.85 and 0.67, confirming a high degree of precision in our segmentation and classification results. Thus, this work demonstrates that a tailored integration of existing methods can yield a robust, automated tool for structural integrity assessment and informed restoration planning, offering a significant improvement in efficiency and reliability.