Research On 3d Reconstruction And Optimization Of Bulk Material Pile

With the innovation and development of the Industrial Internet,the intelligent and digitalization of dry bulk ports has become the development direction of port and shipping companies.Among them,the three-dimensional reconstruction and automatic stacking of accumulations have become key issues in the intelligent operation of dry bulk cargo.This requires three-dimensional reconstruction and optimization research on large dry bulk deposits.This paper focuses on the scanning strategy,reconstruction and optimization method of the pile surface in the bulk cargo port yard,and conducts experiments and analysis on various types of pile shapes.Firstly,based on the current port automatic stacker equipment in the process of developing the technology of complicated process,poor measurement device environment adaptability,high cost and the problem of low precision of model,based on 77 GHZ millimeter wave radar,Bei Dou positioning sensors,posture sensor material piling machine scanning system and reclaimer scanning system,and formulate the corresponding strategy;The formula of multi-sensor data fusion to obtain discrete point cloud data is deduced,and the preprocessing method of millimeter wave radar point cloud data in the storage yard is given.Secondly,aiming at the problem of self-occlusion in the process of pile scanning,it is proposed to use the quantized pigeon swarm algorithm to improve the kriging interpolation algorithm to fill in the holes and grid point cloud data.The cross-validation method is used to compare and analyze the improved kriging interpolation algorithm and other common interpolation algorithms,and give the error curve of various types of point cloud data,which lays the foundation for further free-form surface reconstruction of the stockpile.Again,in view of the low accuracy of free-form surface reconstruction of the curved surface of a variety of large bulk cargo stockpiles,a NURBS surface fitting algorithm optimized by particle swarms is proposed.After adopting Fourier transform to reduce the noise,the model value points required for NURBS fitting are obtained.The particle swarm optimization algorithm is used to configure the node vector to improve the cusp and the location with large fluctuations and the poor fitting effect,and the root mean square error of the shortest distance from the point to the complex NURBS curve and surface is obtained by the search algorithm based on region division.Finally,by testing and scanning stockpiles in the automatic stacking and reclaiming projects of Coal Port and Ore Port,it is proved that the method of this paper is used to process multi-sensor data,three-dimensional reconstruction and optimization to verify the correctness and feasibility of this method.