JACKSON, PHILIP,THOMAS,GABRIEL (2016) Automatic Update of Airport GIS by Remote Sensing Image Analysis. Masters thesis, Durham University.
|PDF - Accepted Version |
Available under License Creative Commons Attribution 3.0 (CC BY).
This project investigates ways to automatically update Geographic Information Systems (GIS) for airports by analysis of Very High Resolution (VHR) remote sensing images. These GIS databases map the physical layout of an airport by representing a broad range of features (such as runways, taxiways and roads) as georeferenced vector objects. Updating such systems therefore involves both automatic detection of relevant objects from remotely sensed images, and comparison of these objects between bi-temporal images. The size of the VHR images and the diversity of the object types
to be captured in the GIS databases makes this a very large and complex problem. Therefore we must split it into smaller parts which can be framed as instances of image processing problems. The aim of this project is to apply a range of methodologies to these problems and compare their results, providing quantitative data where possible. In this report, we devote a chapter to each sub-problem that was focussed on.
Chapter 1 begins by introducing the background and motivation of the project, and describes the problem in more detail.
Chapter 2 presents a method for detecting and segmenting runways, by detecting their distinctive markings and feeding them into a modified Hough transform. The algorithm was tested on a dataset of six bi-temporal remote sensing image pairs and validated against manually generated ground-truth GIS data, provided by Jeppesen.
Chapter 3 investigates co-registration of bi-temporal images, as a necessary precursor to most direct change detection algorithms. Chapter 4 then tests a range of bi-temporal change detection algorithms (some standard, some novel) on co-registered images of airports, with the aim of producing a change heat-map which may assist a human operator in rapidly focussing attention on areas that have changed significantly.
Chapter 5 explores a number of approaches to detecting curvilinear AMDB features such as taxilines and stopbars, by means of enhancing such features and suppressing others, prior to thresholding. Finally in Chapter 6 we develop a method for distinguishing between AMDB lines and other curvilinear structures that may occur in an image, by analysing the connectivity between such features and the runways.
|Item Type:||Thesis (Masters)|
|Award:||Master of Science|
|Keywords:||Remote sensing, runway detection, change detection, image processing, very high resolution|
|Faculty and Department:||Faculty of Science > Engineering and Computing Science, School of (2008-2017)|
|Copyright:||Copyright of this thesis is held by the author|
|Deposited On:||14 Jun 2016 14:42|