Research On Technology Of Building Reconstruction Under LiDAR Point Cloud

Since the concept of smart city was first put forward,city 3D construction causes wide attention.Since the 21 st century,laser scanning technology is improved significantly,hence it can obtain a wide range of urban scene of 3D data in a quickly and accurately way,providing a data basis for 3D reconstruction of the city.As an important part of the city scene,3D reconstruction of artificial buildings is a hotspot of current research,and also facing data processing difficulties and so on.In the reconstruction of complex buildings,we need to solve the model reconstruction of complex detailed structure to recover its appearance model truthfully and completely.In view of challenges and the urgent needs of 3D reconstruction,the research work of this paper is as follows:1.Through deep discussion of research status about 3D reconstruction,this paper introduces several kinds of commonly used 3 d reconstruction algorithm based on laser point cloud,namely knowledge based reconstruction of building models,syntax segmentation based of 3D reconstruction and reconstruction technology based on Glob Fit.Through detailed analysis of advantages and disadvantages of three kinds of algorithm,it can provide theoretical foundation for the following work.2.Since terrestrial laser scanner exists scanning corner,leading to problems such as lacking of point cloud,uneven density,which makes it hard for complete segmentation of building facade and brings great difficulty for sequent 3D reconstruction.Existing RANSAC and Mutil-GS having obvious advantage in sampling strategy,but there exists shortcomings for model selection and subsequent optimization.Therefore,this paper puts forward an algorithm based on a guidance of sampling point density and optimizing the extracted model,namely Global Sample and Model Optimization Sampling and Consensus(GSMOSAC)algorithm.Comparing with the traditional RANSAC and Mutil-GS,the algorithm obtains a better segmentation quality in the light of the experiment results under three types of Li DAR point cloud data.3.On account of the existing algorithms of RANSAC?PEARL and Glob Fit having shortages in water-flow?scalability and automation and so on,we propose a 3D reconstruction algorithms based on regular sets.The algorithm has good scalability,and can automatically accomplish geometry reconstruction.Through experimental analysis under four groups of different types and scale of lidar data,and compared with Ransac,PEARL and Glob Fit algorithms,our algorithm gets better reconstruction effect,and solves the exists generally closed questions in the process of building the geometric reconstruction.