White Rose University Consortium logo
University of Leeds logo University of Sheffield logo York University logo

Image registration for sonar applications

Henson, Benjamin (2017) Image registration for sonar applications. PhD thesis, University of York.

This is the latest version of this item.

[img]
Preview
Text
thesis.pdf - Examined Thesis (PDF)
Available under License Creative Commons Attribution-Noncommercial-No Derivative Works 2.0 UK: England & Wales.

Download (25Mb) | Preview

Abstract

This work develops techniques to estimate the motion of an underwater platform by processing data from an on-board sonar, such as a Forward Looking Sonar (FLS). Based on image registration, a universal algorithm has been developed and validated with in field datasets. The proposed algorithm gives a high quality registration to a fine (sub-pixel) precision using an adaptive filter and is suitable for both optical and acoustic images. The efficiency and quality of the result can be improved if an initial estimate of the motion is made. Therefore, a coarse (pixel-wide) registration algorithm is proposed, this is based on the assumption of local sparsity in the pixel motion between two images. Using a coarse and then fine registration, large displacements can be accommodated with a result that is to a sub-pixel precision. The registration process produces a displacement map (DM) between two images. From a sequence of DMs, an estimation of the sensor's motion is made. This is performed by a proposed fast searching and matching technique applied to a library of modelled DMs. Further, this technique exploits regularised splines to estimate the attitude and trajectory of the platform. To validate the results, a mosaic has been produced from three sets of in field data. Using a more detailed model of the acoustic propagation has the potential to improve the results further. As a step towards this a baseband underwater channel model has been developed. A physics simulator is used to characterise the channel at waymark points in a changing environment. A baseband equivalent representation of the time varying channel is then interpolated from these points. Processing in the baseband reduces the sample rate and hence reduces the run time for the model. A comparison to a more established channel model has been made to validate the results.

Item Type: Thesis (PhD)
Related URLs:
Academic Units: The University of York > Electronics (York)
Identification Number/EthosID: uk.bl.ethos.737917
Depositing User: Mr Benjamin Henson
Date Deposited: 28 Mar 2018 13:21
Last Modified: 24 Jul 2018 15:24
URI: http://etheses.whiterose.ac.uk/id/eprint/19536

Available Versions of this Item

You do not need to contact us to get a copy of this thesis. Please use the 'Download' link(s) above to get a copy.
You can contact us about this thesis. If you need to make a general enquiry, please see the Contact us page.

Actions (repository staff only: login required)