复杂空区三维激光等距扫描与多站点云建模算法研究

It is vital for mine safety, efficient mining and sufficient recycling to obtain accurate 3D morphology of complex cavity and actual boundary. On the basis of analysis of the drawbacks of equiangular 3D laser scanning, the issue closely integrated geometrical features of cavity own space, the method of equidistant 3D laser scanning in complex cavity, point clouds data modeling and matching algorithms of multi-station laser scanning was explored, the basic idea is: through pre-scanning several cross-sectional of cavity, the contour information was acquired, the reference contour was analyzed and extracted; on the basis of the former, radial angle and axial angle control command of equidistant 3D laser scanning was studied, in order to drive the operation of the 3D laser scanner to obtain a substantially uniform distribution of point clouds data of cavity; combined the characteristics of single station laser scanning point clouds data can be projected to the sphere, modeling algorithm of point clouds data based on sphere-plane joint projection have been put forward; considering the complexity of spatial form of deep cavity, single station could not be finished probing, the fast positioning method of multi-station clouds data overlap region need to be studied, matching algorithm between the grid and the surface was proposed, and then the methods of determining grid of the same name, avoiding abnormal matching and matching error allocation was researched. The research results provide the overall solution on obtaining accurate 3D information of complex cavity and modeling in domestic metal mines, provide technical support on safety assessment and disaster control in cavity of metal mines.

复杂空区三维空间形态和实际边界的精确获取关系到矿山安全、高效开采、资源有效回收。本课题在分析传统三维激光等角扫描弊端的基础上，紧密结合空区自身空间几何特点，探索复杂空区三维激光等距扫描方法、多站激光扫描点云数据建模和匹配算法，其基本思路为：预扫描空区若干剖面获取轮廓信息，分析并提取基准轮廓线；在此基础上研究三维激光等距扫描的径向角度和轴向角度控制指令，以此驱动激光扫描仪运作，获取分布均匀的空区点云数据；结合单站激光扫描点云数据可投影至球面上这一特性，提出基于球面-平面联合投影的点云数据建模算法；考虑空区空间形态复杂性无法单站完成探测，研究多站点云数据重叠区域快速定位方法，并提出基于网格与面之间的匹配算法，进而研究同名网格判定、避免异常匹配和匹配误差分配方法等。研究成果为我国金属矿复杂空区三维空间信息精确获取与建模提供整体解决方案，为金属矿空区安全性评价及灾害控制提供技术支撑。

采空区是我国金属矿山安全生产所面临的主要灾源之一。对采空区进行探测及建模，精确获取其三维形态、实际边界、体积大小等空间信息，进而对其实施有效管理和控制是我国矿山安全生产可持续发展所面临的重大课题之一。采用理论研究、实证分析、数值模拟等多种方法，对金属矿采空区激光扫描空间信息精确建模及其应用开展了深入研究，主要研究内容及取得的成果如下：.1）提出了一种三维激光扫描仪数据等距扫描采集方法，利用三维激光扫描仪可扫描得到分布均匀的空区点云数据，解决因复杂空区边界与扫描中心点距离差异导致扫描数据分布不均的问题；.2）针对单站激光扫描点云数据，提出了一种基于球面-平面联合投影的三维激光扫描数据建模方法，通过该方法实现空间点云数据快速准确建模；.3）提出了多点扫描点云数据拼接方法，解决多站激光扫描数据拼接问题。.4）复杂贯通采空区的探测实践表明：对复杂采空区（群）运用激光设备开展多点探测，然后进行点云数据拼接，能有效提高探测效果和建模精度。.本文理论分析与工程实践结合，综合运用激光探测技术、点云处理、三维建模等手段，对采空区三维空间信息精确建模及其应用进行了深入研究，对金属矿山地下工程精密探测、安全评价与管理、资源安全高效回采等具有重要理论意义和工程应用价值。研究成果已成功应用于国内金属矿山企业的采空区建模及管理实践中，取得了良好的实际效果。.