Classification And Building Extraction Based On In-vehicle 3D Laser Scanning Data

In recent years,the development and application of 3D spatial information has become more and more mature,and has been widely used in urban construction and planning.As a new method of data acquisition,vehicle mounted laser scanning technology has been used in the field of geographic information industry.When the vehicle scanning system obtains the data,it can get the 3D information of the objects in the city block.In the city construction,the building runs through the whole city,it is very important to segment and extract the building point cloud data.This paper introduces the composition and working principle of vehicle scanning system;introduces the process of the acquisition process and point cloud data of point cloud data,classification and summarizes other scholars based on the laser scanning data,this paper adopts the method of residual analysis and region growing based on the point cloud data were coarse classification and fine classification.First of all,in this paper,by using the local neighborhood vector and the flatness property based on the plane fitting residuals,the point cloud is divided into different areas such as buildings,surface,poles,vegetation,etc..In the subdivision of the use of the plane of the building property,the building point cloud is divided into different details,such as windows,doors,walls,etc..In the detail classification to extract the details of the construction of the local area of the residual error is.used to determine whether a point is in a plane area,the normal vector angle to determine the similarity of neighborhood points.By calculating the two parameters and St to limit the region growing process,the St value can be used to eliminate the noise points in the growth process.