Study On The Information Extraction Of Streetlight And Street Tree From Vehicle-borne LiDAR Point Cloud

Streetlight and street tree are the main components of road attachments in urban road scenes,bearing night lighting and urban greening functions respectively.The number of streetlights is large and the type is large too,street trees are scattered and different shapes in road scene.It is a cumbersome job to get the details of streetlights and street trees in a large scene.As a rapid acquisition of spatial information technology,vehicle-borne LiDAR has obvious advantages on high-precision of 3D information acquisition of objects in urban road environment and both sides of the street.Extract features’information such as streetlight and tree accurately and efficiently from vehicle-borne LiDAR is the goal of the researches’attempting for a long time,it is a difficult and hot issue to process the vehicle-borne LiDAR data.How to use the extracted information effectively,is also a problem researchers will face.This paper takes the urban road scene as an example to carry out the research on the extraction of the street light and the street tree information,and have done explorations of their application.The specific research work is as follows:1.Analysis the research status of domestic and international.By summarizing the shortcomings of research on streetlight and street tree information extraction based on LiDAR data both at home and abroad,study the research ideas and research route of this paper.2.Perform raw data preprocessing and ground handling.For this paper’s algorithm,do the data preprocessing such as data reduction and block.According to data characteristics,improve the ground point extraction algorithm based on mathematical morphology,and construct the regular DEM.3.Carry out the streetlight point cloud extraction algorithm based on the sample model.Through the establishment of streetlight model,using the L₁ skeleton extraction algorithm to extract the geometric parameters of streetlight samples,matching the geometric parameters of streetlight model and the suspected streetlight features which are according to the mathematical morphology method to extract,then determine the streetlight and extract the streetlight point cloud.At the same time of streetlight matching and judgment,recording the parameters of each street lamp information,that is the detailed streetlight parameter.4.Improve the street tree point cloud extraction algorithm based on the clustering method,and extract street tree’s information.Filter the low feature based on the regular DEM,increase the Euclidean clustering,and extract the street tree point cloud by the extraction strategy.The canopy segmentation of the connected street trees is carried out using the method based on the growth model.The information of the canopy and the trunk is extracted from the single street tree.5.Verify the feasibility of the proposed algorithm.Select the test area vehicle-borne point cloud data,does data preprocessing,ground point extracting and DEM construction,streetlight and street tree point cloud extracting,street trees information extracting and other tests.The results of the improvement of the ground point and the effect of the improved ground are compared.The results of the streetlight point cloud extraction are qualitatively and quantitatively analyzed,and the effect of the street tree point cloud and the information extraction result is analyzed.The results show that the method can effectively extract the streetlight and street tree information.6.Explore the application of streetlight and street tree information.The influence of the growth of the canopy on the streetlight is analyzed.The changes of the streetlight and the street tree in the two periods’data are also analyzed.