| With the development of the industry and agriculture,the demand for potassium salt has also increased.To ensure food security in our country,potash has been classified as a key shortage of minerals.The Lop Nur Yantian deposit is one of the most important potassium-bearing brine deposits in China.It is of great significance to scientifically plan Yantian's annual harvesting plan and ensure the salt content of salt mines to rationally develop non-renewable salt field resources and achieve sustainable development.Therefore,this paper studies the key technologies of the Lop Nur Yantian intelligent mining model.In this thesis,the denoising and simplification algorithms for point cloud data of large-scale brine underwater deposits are studied firstly,and the denoising algorithm based on grid method combined with bilateral filtering and a point cloud simplification based on fitting datum plane are designed respectively.Due to the large scale of the point cloud and the need to improve the efficiency of the algorithm,the multi-threaded parallel processing method is adopted to improve the algorithm.On this basis,multi-threaded parallel and GPU-based multi-thread parallelism based on multi-core CPU are implemented respectively,and two heterogeneous parallels are compared and analyzed.Then,the pre-processed feature point cloud is modeled by triangulation,and the process of network construction is studied in parallel.According to the characteristics that the point cloud of the deposit is not occluded from the horizontal plane,the point cloud data is first dimension-reduced,and the feature point set is delaunay triangulated in the plane.In the process of constructing triangular mesh,considering the mutual independence of point cloud data between channels,a parallel algorithm based on triangulation of non-coupling region decomposition and subnet merging is designed and implemented,and a complete Delaunay triangulation is obtained.The grid model satisfies the accuracy requirements of the deposit model while greatly improving the speed of modeling.This thesis also studies the problem of ore calculation.Considering the influence of different depths of salt ore in different depths of brine,in order to make the ore calculation in the deposit as accurate as possible,a layered model of the deposit is proposed,and each layer can independently reflect the salt mine information in corresponding elevation of the deposit.In addition,for the calculation of ore deposits,a multi-thread parallel algorithm based on Kriging interpolation is proposed,and the accuracy and operation efficiency of the algorithm are compared and analyzed.Finally,combined with the above research results,a salt pool ore estimation and ore visualization system was designed and implemented.The system can display multiple 3D models of the deposit by layer and can interact with the model.In addition,the system can calculate the total ore of the deposit,and can also calculate the ore amount in the designated area,which has practical value. |
PDF Full Text Request