3D岩石结构面粗糙度智能提取方法及应用  

Intelligent Extraction Method of 3D Rock Structure Surface Roughness and its Application

作　　者：王运森[1] 陆健 贾庸凡 WANG Yunsen;LU Jian;JIA Yongfan(Key Laboratory of Ministry of Education on Safe Mining of Deep Metal Mines,Northeastern University,Shenyang,Liaoning 110819,China)

机构地区：[1]东北大学深部金属矿山安全开采教育部重点实验室,辽宁沈阳110819

出　　处：《矿业研究与开发》2021年第10期168-173,共6页Mining Research and Development

基　　金：国家重点研发计划项目(2017YFC0602905);中央高校基本科研业务专项资金资助项目(N2101044).

摘　　要：针对岩石结构面粗糙度现场测量操作不方便和测量结果人为因素影响大的难题,提出一种3D岩石结构面粗糙度智能提取方法。该方法首先运用3D One工具构建标准结构面轮廓线的三维结构面模型,其次利用图像获取设备获取结构面图像并进行格式化处理,最后基于深度学习算法和RNN深度学习网络,对结构面数字化图像进行学习训练及分类处理,估算岩体结构表面的粗糙度系数。将此方法应用于焦家金矿取得了较好的效果,弥补了传统方法工作量繁重,受环境及主观影响大的缺陷,促进了矿山岩石力学现场调查工作的智能化。Aiming at the inconvenient operation of on-site measurement of rock structure surface roughness and the large influence of human factors on the measurement results, a intelligent extraction method of 3 D rock structure surface roughness was proposed. In this method, firstly the 3 D One tool was used to build a three-dimensional structural surface model of the standard structural surface contour, and secondly an image acquisition device was applied to obtain the structural surface image and formatting treatment was conducted. Finally, based on the deep learning algorithm and the RNN deep learning network, learning training and classification processing on the structural surface image were carried out to estimate the roughness coefficient of the rock structure surface. The application of this method in the Jiaojia Gold Mine has achieved good results. This method has made up for the heavy workload of the traditional method and the disadvantages of being greatly affected by the environment and subjective, which promotes the intelligentization of the mine rock mechanics field investigation work.

关 键 词：岩石结构面 粗糙度 数字化图像 RNN 智能提取 

分 类 号：TU45[建筑科学—岩土工程]