Use this URL to cite or link to this record in EThOS:
Title: An investigation into semi-automated 3D city modelling
Author: Kokkas, Nikolaos
ISNI:       0000 0004 2694 8289
Awarding Body: University of Nottingham
Current Institution: University of Nottingham
Date of Award: 2009
Availability of Full Text:
Access from EThOS:
Access from Institution:
Creating three dimensional digital representations of urban areas, also known as 3D city modelling, is essential in many applications, such as urban planning, radio frequency signal propagation, flight simulation and vehicle navigation, which are of increasing importance in modern society urban centres. The main aim of the thesis is the development of a semi-automated, innovative workflow for creating 3D city models using aerial photographs and LiDAR data collected from various airborne sensors. The complexity of this aim necessitates the development of an efficient and reliable way to progress from manually intensive operations to an increased level of automation. The proposed methodology exploits the combination of different datasets, also known as data fusion, to achieve reliable results in different study areas. Data fusion techniques are used to combine linear features, extracted from aerial photographs, with either LiDAR data or any other source available including Very Dense Digital Surface Models (VDDSMs). The research proposes a method which employs a semi automated technique for 3D city modelling by fusing LiDAR if available or VDDSMs with 3D linear features extracted from stereo pairs of photographs. The building detection and the generation of the building footprint is performed with the use of a plane fitting algorithm on the LiDAR or VDDSMs using conditions based on the slope of the roofs and the minimum size of the buildings. The initial building footprint is subsequently generalized using a simplification algorithm that enhances the orthogonality between the individual linear segments within a defined tolerance. The final refinement of the building outline is performed for each linear segment using the filtered stereo matched points with a least squares estimation. The digital reconstruction of the roof shapes is performed by implementing a least squares-plane fitting algorithm on the classified VDDSMs, which is restricted by the building outlines, the minimum size of the planes and the maximum height tolerance between adjacent 3D points. Subsequently neighbouring planes are merged using Boolean operations for generation of solid features. The results indicate very detailed building models. Various roof details such as dormers and chimneys are successfully reconstructed in most cases.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: TA 501 Surveying