基于区域生长法的三维激光点云标靶提取方法研究

Research on Target Extraction Method of 3D Laser Point Cloud Based on Region Growing Algorithm

摘要:

三维激光点云中标靶核心区域的精准提取是后续靶心定位和点云数据对齐的关键。针对传统标靶提取方法依赖预设阈值、易受隧道复杂施工环境影响等问题，提出一种基于区域生长法的标靶提取方法。基于标靶局部颜色相似性准则，结合手动种子点选择与固定半径邻域生长策略，实现标靶区域的自适应提取。通过采用几何拟合法与重心法对提取的标靶核心区域进行靶心定位，并与激光放样的准确靶心坐标开展误差对比。结果表明：所提出的方法无需人工动态调整参数，在隧道复杂施工环境下仍能保持较高的提取精度与数据稳定性，定位偏差与基准靶心坐标高度吻合，显著提升了复杂施工环境下标靶提取的精度与工程适用性。

Abstract:

Accurate extraction of the core target area in 3D laser point clouds is the key to subsequent target center positioning and point cloud data registration. Aiming at the problems of traditional target extraction methods, such as reliance on preset thresholds and susceptibility to the complex construction environment of tunnels, a target extraction method based on the region growing algorithm is proposed. Based on the local color similarity criterion of targets, combined with the strategies of manual seed point selection and fixed-radius neighborhood growth, the adaptive extraction of target areas is realized. The geometric fitting method and the centroid method are used to locate the target center of the extracted core target area, and the error comparison is carried out with the accurate target center coordinates obtained by laser lofting. The results show that the proposed method does not require manual dynamic parameter adjustment, and can still maintain high extraction accuracy and data stability in the complex tunnel construction environment. The positioning deviation is highly consistent with the reference target center coordinates, which significantly improves the extraction accuracy and engineering applicability of targets in complex construction environments.