School of Infrastructure Engineering, Nanchang University
China Railway Water Conservancy and Hydropower Planning and Design Group Co., Ltd.
School of Resources and Environment, Nanchang University

准确识别岩质高边坡结构面和获取产状统计信息是进行边坡稳定性分析的重要前提。无人机摄影测量技术为解决高边坡结构面准确勘测难题提供了可能，但缺少高效准确的影像后处理方法，且现有研究没有考虑结构面产状信息特征的不确定性，致使结构面识别准确性差、效率低。针对该问题，以江西省南昌市某露天矿高边坡为研究背景，提出了融合无人机摄影、后处理算法及统计分析的一体化结构面识别与产状统计信息采集方法。首先，通过Phantom 4 Pro V2.0无人机获取边坡表面影像；其次，利用Context Capture软件进行处理，得到高密度三维点云数据；然后，采用K近邻（KNN）算法中的确定近邻点数量法构建相似点集，采用基于密度的聚类（DBSCAN）算法进行聚类分析，从而实现边坡结构面识别，获得结构面产状信息并进行统计特征分析；最后，通过现场勘测数据进行对比验证。结果表明：该方法能够快速获取完整的高密度点云数据，准确高效地识别岩质高边坡大部分结构面，识别结果与边坡工程现场实际情况基本吻合；该方法可获取高边坡结构面数量、产状信息及其统计特征，大部分结构面倾角和倾向概率分布与实测数据拟合较好，为高边坡裂隙网络模型构建及稳定性分析提供了重要数据来源。

Accurately recognizing the structural planes of high rock slopes and obtaining occurrence information are important prerequisites for conducting slope stability analysis. Unmanned aerial vehicle photogrammetry technology provides the possibility to solve the problem of accurate surveying of high slope structural planes. But it lacks efficient and accurate image post-processing methods. The existing research has not considered the uncertainty of structural plane occurrence information features, resulting in poor accuracy and efficiency in structural plane recognition. In order to solve the above problems, taking a high slope of an open-pit mine in Nanchang City, Jiangxi Province as the research background, an integrated method for recognizing structural planes and collecting occurrence information by integrating unmanned aerial vehicle photography, post-processing algorithms, and statistical analysis is proposed. Firstly, the method obtains surface images of the slope using the Phantom 4 Pro V2.0 unmanned aerial vehicle. Secondly, the method uses Context Capture software for processing, and the high-density 3D point cloud data is obtained. Secondly, the K-nearest neighbor (KNN) algorithm is used to determine the number of nearest neighbor points to construct a set of similar points. The density-based spatial clustering of applications with noise (DBSCAN) algorithm is used for clustering analysis to recognize slope structural planes, obtain structural plane occurrence information, and perform statistical feature analysis. Finally, comparative verification is conducted through on-site survey data. The results show that this method can quickly obtain complete high-density point cloud data, accurately and efficiently recognize most structural planes of high rock slopes. The recognition results are basically consistent with the actual situation of slope engineering sites. This method can obtain information on the number, occurrence, and statistical features of high slope structural planes. The probability distribution of most structural plane dip angles and inclinations fits well with the measured data, providing an important data source for the construction of high slope fracture network models and stability analysis.

国家自然科学基金优秀青年科学基金项目(52222905)；江西省自然科学基金资助项目(20232ACB204031, 20224ACB204019)；江西省水利科技计划资助项目(202325ZDKT07, 202426ZDKT12)。

蒋水华，余琦，黄河，等. 岩质高边坡结构面识别及产状统计信息采集方法[J]. 工矿自动化，2024，50（7）：156-164.

JIANG Shuihua, YU Qi, HUANG He, et al. Structural planes recognition and occurrence statistics information collection method for high rock slopes[J]. Journal of Mine Automation，2024，50（7）：156-164.