面向密集点云的古城墙三维裂缝检测及三维形变提取方法研究

The safety of the existing ancient walls is to be threatened caused by the rapid development of infrastructure and aging effects. Therefore, it is extremely urgent to realize complete deformation monitoring for the ancient walls. The technique of terrestrial LiDAR is with the advantages of high speed, high accuracy and high resolution in extracting deformation information. In this project, the research on the three-dimensional deformation monitoring of ancient wall using dense terrestrial LiDAR point clouds is carried out from the perspective of theory and methodology. The main contents are summarized as follows: (1) an automatic approach combining digital image processing, improved watershed algorithm, principle component analysis and rectangular fitting is proposed to extract three-dimensional shape features, by which the brick three-dimensional shape can be extracted with good tolerance to noise interference and with high calculation efficiency. (2) the 3D deformation is identified and analyzed based on the baseline and baseline network that has the additional advantage that the registration is no longer required. In this approach, the accurate extraction of 3D deformation value, direction and local structural deformation information are realized. (3) the alpha-shapes algorithm is introduced to detect and analyze the 3D cracks in an adaptive way, in which the 3D geometrical features of the cracks are extracted effectively, the tendency of the cracks is obtained as well. The objective of this project will improve the reality and efficiency of the 3D deformation extraction for the ancient walls, and build the basic theories and approaches for the 3D cracks detection and tendency analysis. The research achievements will promote the application of terrestrial LiDAR in the field of deformation monitoring. In addition, it can be wildly applied in the protection of ancient walls and avoid the natural disasters.

由于基础设施的高速发展及时效等因素的影响，现存的古城墙安全性存在隐患，对古城墙进行完备的形变监测迫在眉睫。地面激光雷达提取形变具有高速、高精度、高分辨率等优点。本课题以古城墙为研究对象，利用密集地面激光雷达点云，从理论和方法上开展三维形变监测研究，主要研究内容如下： (1)融合改进分水岭算法、主成分分析及矩形拟合法，提出一种抗噪能力强、计算效率高的砖块三维形状提取方法，实现形特征的自动提取；(2)提出基于基线和基线网的三维形变提取与分析方法，消除配准误差，实现墙体三维形变大小、方向及局部结构形变的准确提取；(3)引入alpha-shapes算法，提出自适应的三维裂缝检测与分析方法，实现裂缝三维形态特征的有效提取与趋势分析。本项目的研究，将提高古城墙三维形变提取和三维裂缝检测的真实性和效率，促进地面激光雷达技术在形变监测领域的应用，可大范围应用于古城墙文物保护，预防自然灾害。

由于基础设施的高速发展及时效等因素的影响，现存的古城墙安全性存在隐患，对古城墙进行完备的形变监测迫在眉睫。地面激光雷达技术近年来高速发展，提取形变具有高速、高精度、高分辨率等优点。本课题以古城墙为研究对象，利用密集地面激光雷达点云，从理论和方法上开展了三维形变监测的研究，主要研究内容及成果如下： (1)针对点云分布不均、噪声等因素的影响，融合改进分水岭算法、主成分分析及矩形拟合法，提出一种抗噪能力强、计算效率高的砖块三维形状提取方法，实现了形特征的自动提取；(2)针对配准引入的不匹配误差直接影响了形变提取的可靠性，提出了一种基于基线和基线网的三维形变提取与分析方法，消除配准误差，实现了墙体三维形变大小、方向及局部结构形变的准确提取；(3)针对城墙三维裂缝的多样性、差异化等特征，深入分析裂缝点云的几何特性，引入alpha-shapes算法，提出了一种自适应的三维裂缝检测与分析方法，实现了裂缝三维形态特征的有效提取与趋势分析。本项目的研究成果，提高了古城墙三维形变提取和三维裂缝检测的真实性和效率，促进了地面激光雷达技术在形变监测领域的应用，可大范围应用于古城墙文物保护，预防自然灾害。