Automatic extraction of bridges and footbridges from airborne LiDAR point cloud

The product of the Airborne LIDAR (Light Detection and Ranging) scanning is a point cloud on the 3D platform. Airborne LIDAR Scanning is capable of delivering a high density and high accuracy point clouds in a relatively short time. However, despite the ability to measure objects with high positional accuracy, automatic detection and interpretation of individual objects in the landscapes is still a problem. As an example of these problems, the classification of point clouds produced by Airborne LIDAR Scan. The classification of Airborne LIDAR scanning point clouds consists primarily of labeling the points as objects or bare earth. The labeled object points are then also labeled as buildings or vegetation etc. Bridges, overpass and footbridges, one of the objects to be limited, have played an important role in human transport networks. Moreover, automatic and rapid extraction of bridge information from LIDAR data is of great importance in GIS (Geographical Information System) applications, DTM (Digital Terrain Model) and intelligent transportation systems. Especially, the determination of the bridge and footbridge outline is a very important problem. This is the focus of many scientists' research topics. In this study, the researches about the fundamentals of the LIDAR system to the LIDAR data format, the programs used to perform the LIDAR data analysis and LIDAR filtering algorithms which have been done to date are explained clearly. In this thesis, it is aimed to extract automatic bridge and footbridge from 3D airborne LIDAR point cloud data belonging to two areas selected from Şişli district of Istanbul and one areas from United States of Utah George City. For this purpose, firstly, The ground filtering process was carried out by using ATIN algorithm, which is differentiate ground and non-ground data and improved with commercial software and presented to users. Bare earth which was obtained by ground filtration was eliminated from the first point cloud, and classification was made using over the remaining LIDAR points and the bridge object was obtained. The bridge outer boundary of the obtained bridge point cloud was automatically plotted using LIDAR points. Bridge areas created by drawing are compared with reference bridge area for accuracy analysis. In this thesis, point cloud analysis and statistical results were obtained with Global Mapper and LAStools software. Keywords: Bridge, DTM, Footbridge, LIDAR, Overpass, Point Cloud, Transportation