NURBS曲面重构下水库高陡边坡岩体结构监测数据无人机识别方法  

UAV Identification Method for Monitoring Data of Rock Mass Structure on High and Steep Slopes of Reservoir Under NURBS Surface Reconstruction

作　　者：申世杰 邓伟 刘天榜 谢东 袁程文 姚经圣 彭泽伟 徐懋民 SHEN Shijie;DENG Wei;LIU Tianbang;XIE Dong;YUAN Chengwen;YAO Jingsheng;PENG Zewei;XU Maoming(Sichuan Zhongtu 3D Information Technology Co.,Ltd.,Chengdu,Sichuan 610083,China)

机构地区：[1]四川中图三维信息技术有限公司,四川成都610083

出　　处：《自动化应用》2024年第21期12-16,共5页Automation Application

摘　　要：面对复杂、不规则的岩体结构面,传统识别方法很难准确识别,因此,为了解水库高陡边坡岩体结构状态,保证水库安全,提出一种NURBS曲面重构下水库高陡边坡岩体结构监测数据无人机识别方法。通过无人机搭载激光扫描仪,获取高陡边坡的点云数据并实施预处理,为后期岩体结构识别打下了重要基础。利用NURBS曲面重构技术生成高陡边坡岩体的三维模型。通过构建CNN模型,自动从三维模型中提取出与岩体结构相关的特征,实现对水库高陡边坡岩体结构的准确分类和识别。结果表明,所研究方法的汉明损失更小,说明识别误差更小,准确性更高。In the face of complex and irregular rock mass structural planes,traditional identification methods are difficult to accurately identify.Therefore,in order to understand the structural state of high and steep slope rock mass in reservoirs and ensure reservoir safety,a method for UAV identification method for monitoring data of rock mass structure of high and steep slope in reservoir under NURBS surface reconstruction is proposed.By using UAV equipped with laser scanners,point cloud data of high and steep slopes is obtained and preprocessed,providing an important foundation for later rock structure identification.Using NURBS surface reconstruction technique to generate a three-dimensional model of high and steep slope rock mass.By constructing a CNN model,features related to rock mass structure are automatically extracted from the 3D model,achieving accurate classification and recognition of the rock mass structure of high and steep slopes in reservoirs.The results indicate that the Hamming loss of the studied method is smaller,indicating a smaller recognition error and higher accuracy.

关 键 词：NURBS曲面重构 水库 高陡边坡 点云监测 无人机 CNN识别方法 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]