中国安全科学学报 ›› 2025, Vol. 35 ›› Issue (S2): 131-138.doi: 10.16265/j.cnki.issn1003-3033.2025.S2.0015

• 安全工程技术 • 上一篇    下一篇

基于无人机贴近摄影的高陡边坡危岩体结构面识别

丁兴龙(), 于正兴, 张亦海**(), 石志磊   

  1. 中安国泰(北京)科技发展有限公司, 北京 102209
  • 收稿日期:2025-06-10 出版日期:2026-02-04
  • 通信作者:
    **张亦海(1989—),男,河北沧州人,博士,高级工程师,主要从事边坡安全方面的工作。E-mail:
  • 作者简介:

    丁兴龙 (1995—),男,甘肃定西人,本科,工程师,主要从事边坡稳定性等方面的工作。E-mail:

    于正兴 高级工程师

    张亦海 高级工程师

    石志磊 工程师

Identification of structural planes of dangerous rock mass on high and steep slopes based on close-range UAV photography

DING Xinglong(), YU Zhengxing, ZHANG Yihai**(), SHI Zhilei   

  1. Cathay Safety Technology Co., Ltd., Beijing 102209, China
  • Received:2025-06-10 Published:2026-02-04

摘要:

为解决高陡岩质边坡地质调查困难、危岩体稳定性判别受限的问题,利用无人机灵活、机动、安全、高效的技术优势快速精准获取危岩体结构特征信息,以新井煤矿滑坡为对象,运用无人机贴近摄影获取滑坡边界高分辨率、高精度影像;采用运动结构恢复算法(SFM)生成细节和纹理丰富的三维点云和三维实体模型;精确估计点云法向量后,使用HSV色彩空间对点云赋色,得到结构面分布特征;运用随机抽样一致性(RANSAC)点云多平面分割算法识别主断壁岩体结构面,并利用赤平极射投影分析主断壁岩体稳定性。结果表明: RANSAC可有效从含有噪声和离群点的点云数据中识别结构面,主断壁岩体结构面呈3组产出。岩体发生平面滑动、楔形体滑动和倾倒破坏的概率分别为14.89%、37.76%和11.83%。

关键词: 无人机, 贴近摄影, 高陡边坡, 结构面识别, 赤平极射投影

Abstract:

To address the challenges of geological surveys and the limited stability assessment of unstable rock masses in high and steep rock slopes, UAVs were used for their flexibility, mobility, safety, and efficiency in rapidly and accurately acquiring structural characteristic information of unstable rock masses, thereby supporting slope stability evaluation. By taking the Xinjing Coal Mine landslide as a case study, close-range photography via UAVs was employed to obtain high-resolution and high-precision images of the landslide boundary. The structure from motion (SFM) algorithm was applied to generate detailed and texture-rich 3D point clouds and solid models. After accurately estimating point cloud normals, the HSV color space was used to colorize the point clouds, revealing the distribution characteristics of structural planes. The random sample consensus (RANSAC) point cloud multi-plane segmentation algorithm was applied to identify the structural planes of the main broken wall's rock mass, and stereographic projection was utilized to analyze the stability of the main broken wall's rock mass. The results indicate that RANSAC can effectively identify structural planes from point cloud data containing noise and outliers, and the structural planes in the main broken wall's rock mass are grouped into three sets. The probabilities of planar sliding, wedge sliding, and toppling failure are 14.89%, 37.76%, and 11.83%, respectively.

Key words: unmanned aerial vehicle (UAV), close-range photography, high and steep slope, structural plane identification, stereographic projection

中图分类号: