岩心矿物立体填图系统研究与实现

发布时间：2018-03-25 14:32

本文选题：立体填图　切入点：克里金　出处：《南京理工大学》2017年硕士论文

【摘要】：岩心钻探分析地下矿物分布是地质和矿产勘探的重要手段。由于人力和财力有限,实际勘探区域不可能建立大量的岩心钻井,如何利用少量岩心数据进行地下勘探空间的矿物丰度预测和立体填图是地质和矿产勘探可视化的关键。本文基于少量的钻井岩心数据,研究和实现了岩心数据立体填图的关键技术,并研发了一个岩心数据矿物立体填图系统,主要工作如下:1)利用勘探区域有限钻井获取的岩心矿物数据,建立了一种克里金插值的矿物丰度立体填图可视化方法。该方法对勘探区域进行单元分解表示,采用克里插值方法对未知体素单元的矿物丰度进行预测,并通过建立矿物丰度和颜色映射关系,实现勘探区域的立体填图。实验分析比较了不同变差函数对矿体丰度预测结果的影响,给出了立体填图结果。2)设计了一种结合广义三棱柱(GTP)几何表达和BP神经网络学习模型的岩心矿物预测和立体填图方法。该方法使用不规则三角网(TIN)建立钻井之间的连接关系,基于广义三棱柱(GTP)对钻井勘探空间进行几何表达;把勘探空间进行单元分解,采用BP神经网络预测广义三棱柱内的体素矿物丰度。3)设计了一种基于TIN-GTP-TEN几何表示模型的矿物立体填图方法。该方法通过不规则三角网组成GTP模型,利用最小顶点标识法(SVID)对GTP模型对角剖分,形成四面体;进而利用距离反比法计算四面体的矿物丰度。最后给出了可视化实验,实验结果表明该方法较好克服了矿体剖分的不一致性,更准确地表达了矿体的外部轮廓。4)集成上述立体填图方法,基于OpenGL研发了一个岩心矿物立体填图系统。该系统可实现立体填图的三维观察和矿体剖切,通过若干测试数据进行验证演示,取得较好可视化结果。
[Abstract]:Core drilling analysis of underground mineral distribution is an important means of geological and mineral exploration. Due to the limited manpower and financial resources, it is impossible to establish a large number of core drilling in the actual exploration area. How to make use of a small amount of core data to predict the mineral abundance of underground exploration space and to make three-dimensional mapping is the key to the visualization of geological and mineral exploration. This paper is based on a small amount of drilling core data. The key technology of core data stereo mapping is studied and realized, and a core data mineral mapping system is developed. The main work is as follows: 1) the core mineral data obtained by using limited drilling in exploration area. A visualization method of mineral abundance by Kriging interpolation is established, in which the mineral abundance of unknown voxel units is predicted by the method of element decomposition and representation of exploration area, and the mineral abundance of unknown voxel units is predicted by Kerry interpolation method. By establishing the relationship between mineral abundance and color mapping, the three-dimensional mapping of exploration area is realized. The effects of different variation functions on the results of ore body abundance prediction are analyzed and compared experimentally. In this paper, a method of core mineral prediction and stereo mapping combined with generalized triangular prism (GTP) geometry expression and BP neural network learning model is designed. This method uses irregular triangulation network (TINs) to establish the connection between drilling wells. Based on the generalized triangular prism (GTP), the drilling exploration space is expressed in geometry, and the exploration space is decomposed into units. Using BP neural network to predict the bulk mineral abundance in the generalized triangular prism, a mineral stereo mapping method based on TIN-GTP-TEN geometric representation model is designed. The GTP model is made up of irregular triangulation networks. The minimum vertex marking method is used to divide the GTP model diagonally to form a tetrahedron, and the distance inverse ratio method is used to calculate the mineral abundance of the tetrahedron. Finally, the visualization experiment is given. The experimental results show that the method can overcome the inconsistency of ore body subdivision and more accurately express the external contour of ore body. A three-dimensional mapping system for core minerals is developed based on OpenGL. The system can realize the three-dimensional observation of three-dimensional mapping and the cutting of ore bodies. The system is verified and demonstrated by some test data and good visualization results are obtained.
【学位授予单位】：南京理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P628;TP183

【相似文献】