Document Type

Theses, Masters

Rights

Available under a Creative Commons Attribution Non-Commercial Share Alike 4.0 International Licence

Publication Details

Successfully submitted for the award of Master of Philosophy (M.Phil.) to the Dublin Institute of Technology, 2006.

Abstract

Due to its speed and accuracy the Global Positioning System (GPS) is widely used as a data collection tool. Problems however can occur when this GPS data is used in conjunction with existing National Mapping Agencies (NMA) vector databases that are not of comparable accuracy. Shifts and misalignments of the datasets can occur. In talks with the Irish mapping agency, Ordnance Survey Ireland (OSi), prior to this project, it viewed with interest the possibility of using Airborne Laser Scanning (ALS) data as a general quality indicator of existing vector databases. The aim of this research was to extract the centre line of a small segment of straight road from triangulated ALS ground points. ALS data with a point density of 2 points per square metre was processed using TerraScan to yield a set of ground points. The extraction process was based on the creation and analysis of cross-sections taken at regular intervals from the triangulated ALS data. The cross-section widths and intervals were based on a search template developed from the start and end coordinates of an assumed centre line taken from an existing vector database. The cross-sections developed were based on individual triangles of the triangulation, groups of triangles and on interpolated data. Parameters of gradient, intensity and interpolated height are investigated. Algorithms were developed in MatLab to create and semi-automatically analyse the cross-sections. Cross-sections were generated for two different road sections and a ground truth survey was conducted for one of the roads. The most useful cross-sections were those based on Interpolated Heights from the triangulated ALS data using the road width as an additional parameter. Results demonstrated that it was possible to define the true road extent from the ALS data with accuracy equal to its point density of 2m by using a linear Least Squares best-fit algorithm. The Intensity of the return pulse was not used in the extraction process and formed a separate piece of research. The findings were that the most useful cross-section were those based on the Intensity Standard Deviations of the vertices of individual triangles in the triangulated ALS data and on Interpolated Intensity. Results show that it is possible to detect road markings from this information.

DOI

https://doi.org/10.21427/D7CW3Q


Share

COinS