爆破岩块自动识别与块度特征提取方法  

Method of Automatic Identification and Extraction of the Fragmentation Characteristics of Blasting Rock Blocks

作　　者：谢博 施富强[1,2] 赵建才 朱登高 蒋建德 李徐然 廖学燕[2] 李锋[1] XIE Bo;SHI Fu-qiang;ZHAO Jian-cai;ZHU Deng-gao;JIANG Jian-de;LI Xu-ran;LIAO Xue-yan;LI Feng(School of Mechanical Engineering,Southwest Jiaotong University,Chengdu 610031,China;Sichuan Academy of Safety Science and Technology,Chengdu 610045,China;DARCH CONSTRUCTION GROUP CO LTD,Zhoushan 316000,China)

机构地区：[1]西南交通大学机械工程学院,成都610031 [2]四川省安全科学技术研究院,成都610045 [3]大昌建设集团有限公司,舟山316000

出　　处：《爆破》2019年第3期43-49,共7页Blasting

基　　金：四川省科技计划项目[2016]8号;四川省省级科研院所基本科研业务费项目(2018YSKY0038)

摘　　要：爆破块度是评价爆破效果的主要评价指标之一。为了快速、准确获取这一指标,以爆破后部分岩块为研究对象,利用三维激光扫描仪获取岩块点云数据,在原始点云数据预处理的基础上,通过设定体素聚类控制标准与权重因子w的变化区间计算得到超体聚类点云。然后通过凹凸性判据中阈值(βThresh、θThresh)的设定对聚类结果融合,利用区域生长的思想进行岩块点云分割,达到岩块轮廓面自动识别的目的。从岩块点云的分割结果中依次提取出单个岩块点云,应用PCA算法自动计算岩块粒径并统计块度分布结果。结果表明:岩块分割效果好,自动计算法与手动量取法的平均相对误差约为4.90%,且块度分布的统计结果对爆破质量评价有一定的指导作用,满足对爆破质量分析与评价的要求。The blasting fragmentation is one of the main evaluation indexes for evaluating the blasting effect.In order to obtain this index quickly and accurately,the blasting fragments were taken as research objects.And the point cloud data of the rock blocks was obtained by a 3D laser scanner.Based on the pretreatment of the original point cloud data,a supervoxel clustering point cloud was obtained by setting the voxel clustering control standard and the weighting factor w.Then the clustering results were integrated by the threshold(β Thresh 、θ Thresh )setting in the concavity and convexity criterion.The idea of regional growth was used to complete the rock point cloud segmentation,which achieved the purpose of automatic identification of blasting rock blocks.Then,extracting each rock point cloud from the segmentation results of the rock point cloud in turn,the PCA algorithm was used to automatically calculate the rock mass size and count the fragmentation distribution.The results showed that the rock segmentation effect was good,and the relative error between the automatic calculation method and the manual measurement method was 4.90 %.Therefore,the statistical results have a certain guiding effect on the blasting quality evaluation,which meets the requirements for blasting quality analysis and evaluation.

关 键 词：爆堆块度 三维激光扫描 点云分割 超体素体 岩块粒径 

分 类 号：TV542[水利工程—水利水电工程]