Urban ground motion due to natural or man-made geological processes is an issue of major importance for local authorities, property developers, planners and buyers. Increased knowledge of this phenomena would benefit all involved but the measurement techniques in common use have either spatial or temporal inadequacies. A technique known as Persistent Scatterer Interferometry (PSI) has been developed which can map ground motion to high precision over large areas with a temporal scale measured in years. PSI takes advantage of the high number of Synthetic Aperture Radar (SAR) images available to mitigate the atmospheric effects that inhibit standard Interferometric SAR (InSAR) techniques. This however involves assumptions about the nature of atmospheric variability, such as its randomness over time, or its spatial extent. In addition, little is known about the Persistent Scatterers (PS) themselves and PSI is only able to provide results relative to a reference PS. The reference PS point is often arbitrarily chosen and may itself be in an area undergoing ground motion, thus adding a degree of ambiguity to any relatively derived motion. The purpose of this work is to investigate possible solutions to these shortfalls and quantify any improvements made. A corner reflector network is established in the Nottingham area of the UK. A data archive is collated over three years containing Global Positioning System (GPS) data at the corner reflector sites, data from surrounding Continuous GPS (CGPS) sites and levelling data. Due to conflicts with the European Space Agency (ESA) Environmental Satellite (ENVISAT), there were insufficient SAR images to com- pute a fully integrated corner reflector PSI study. Instead, the project focussed on atmospheric correction of PSI results using absolute ZWD estimates. Zenith Wet Delay (ZWD) estimates are derived from a Precise Point Positioning (PPP) GPS processing method which does not rely on a network of ground stations and therefore produces absolute ZWD estimates which are less prone to biases and noise. These are interpolated across a PSI study area and used to mitigate the long wavelength effects of atmopheric water vapour in the PSI differential interferograms. The corrected PSI results are then compared to uncorrected results, GPS derived motion and levelling data. Results between the ZWD corrected PSI study and the uncorrected study show statistical improvements in some areas and reductions in others. Correlation factors between double-differenced levelling observations and double-differenced PSI results improve from 0.67 to 0.81. PSI deformation rates also show improvement when compared to GPS deformation rates, although some results do not satisfy statistical tests.

A Shack Hartmann wavefront sensor is used to detect the distortion of light in an optical wavefront. It does this by sampling the wavefront with an array of lenslets and measuring the displacement of focused spots from reference positions. These displacements are linearly related to the local wavefront tilts from which the entire wavefront can be reconstructed. In most Shack Hartmann wavefront sensors, a CCD is used to sample the entire wavefront, typically at a rate of 25 to 60 Hz, and a whole frame of light spots is read out before their positions are processed. This results in a data bottleneck. In this design, parallel processing is achieved by incorporating local centroid processing for each focused spot, thereby requiring only reduced bandwidth data to be transferred off-chip at a high rate. To incorporate centroid processing at the sensor level requires high levels of circuit integration not possible with a CCD technology. Instead a standard 0.7J..lmCMOS technology was used but photodetector structures for this technology are not well characterised. As such characterisation of several common photodiode structures was carried out which showed good responsitivity of the order of 0.3 AIW. Prior to fabrication on-chip, a hardware emulation system using a reprogrammable FPGA was built which implemented the centroiding algorithm successfully. Subsequently, the design was implemented as a single-chip CMOS solution. The fabricated optical centroid processor successfully computed and transmitted the centroids at a rate of more than 2.4 kHz, which when integrated as an array of tilt sensors will allow a data rate that is independent of the number of tilt sensors' employed. Besides removing the data bottleneck present in current systems, the design also offers advantages in terms of power consumption, system size and cost. The design was also shown to be extremely scalable to a complete low cost real time adaptive optics system.

One of the most exciting areas of research in optics is rare-earth doped glasses and fibres with capacity for near-infrared to mid-infrared operations. In particular, there is great interest in optimising parameters like ion concentration, fibre length/geometry, and pump conditions for applications in photoluminescence, amplification and lasing. Round trip investigation from material fabrication, experimental setup and actual device can be laborious, expensive and come with some uncertainties. Some of these uncertainties are accurate identification of ion-ion interactions, impact of such interactions on device performance, correct extraction of phenomenological material properties and the prediction of combination of properties with numerical methods. In this thesis, the spectroscopic behaviour of rare-earth doped materials are theoretically studied via numerical simulations and experimentally verified. The models developed are applicable to steady-state and transient behaviour of rare-earths under different excitation conditions. For the simulation, a couple of spectroscopic parameters are needed which have to be obtained in advance from bulk glasses. Parameters like radiative and non-radiative lifetimes are calculated by complementing theoretical analysis with a few experimental measurements. The first part of the research concentrates on the study of ion-ion interactions in different concentrations of erbium doped sol-gel SiO2 prepared by the sol-gel method. The work includes continuous-wave (CW) and pulsed excitation spectroscopic measurement on the glasses that provide data for the model. These measurements together with the rate-equation modelling are used to obtain a physical understanding of the processes responsible for the fluorescence features observed. A particle swarm optimisation technique was used to predict the values of the ion-ion interactions. The behaviour of the 488 nm and 800 nm excitations were consistent with the predictions of the model. Indeed, the agreement between the calculated photoluminescence and the measured emission indicates that the six important processes that influence the ion-ion interactions in the bulk material have been correctly identified and included. With this model of photoluminescence at hand, it was possible to extend it to laser or amplifier configurations. Subsequently, erbium doped ZBLAN glass fibre with lower phonon energy were explored for lasing in the mid-infrared for application to 2.73 µm high-power delivery for tissue surgery. Accurate laser characteristics were predicted for two different designs, including the ultimate thermal designs. Optimum boundary conditions of mirror end-facet reflectivity, fibre length and effects of modelling parameters were addressed. The study is complimented with experimental data of double-clad fibre and the results reported were a clear documentation of the design of erbium doped ZBLAN fiber laser. Finally, the potential of P r3+ doped chalcogenide (GeAs(Ga/In)Se) glass for photoluminescence and lasing at 4.73 µm is studied. This is to answer the research question - Can we extract the spectroscopic parameters and also model the superior property of these novel glasses?. The laboratory facilities and availability of experimental data were decisive in the choice of praseodymium ions as well as inclusion of Gallium or Indium for this part of the research. The superior characteristics of Indium over Gallium for hotoluminescence and consequently device characteristics were studied with the aid of a rate equation model. The phenomenon of photon reabsorption in the chalcogenide fibres were also simulated and verified with experiment. The work has produced a comprehensive numerical model for the simulation of photoluminescence in P r3+doped selenide based chalcogenide glass and fibre from NIR to mid-IR especially in the Gallium and Indium based analogues.

This thesis presents the study of a relatively new class of photonic structures in-voking Parity-Time (PT)-symmetry. PT-symmetric structures in photonics, as a realisation of PT-symmetric Quantum Mechanics problems, are constructed by a judicious design of refractive index modulation which requires the real part of the refractive index to be an even function and the imaginary part of the refractive index to be an odd function in space. PT-symmetric structures in the form of Bragg gratings, coupled resonators and chain resonators are the main configurations studied in this thesis. These PT-symmetric structures feature a spontaneous symmetry breaking at which interesting wave behaviour such as an asymmetric response depending on the direction of the incident wave, unidirectional invisibility, simultaneous coherent-perfect absorber lasing and localised termination modes are observed; these behaviours are presented in this thesis. Theoretical and numerical studies of these PT-symmetric structures are undertaken which assume realistic material parameters,including material dispersion and material non-linearity. Moreover,in this thesis, potential applications of these PT-symmetric structures are explored. The first part of the thesis considers PT-symmetric Bragg grating structures which are formed by introducing a PT-symmetric refractive index modulation into a Bragg grating structure. If gain/loss dispersion is considered, it is shown that dispersion limits the PT-symmetric operation to just a single frequency. As such spontaneous symmetry breaking can only be achieved by varying the gain/loss parameter. Nevertheless, it is shown that by switching the gain/loss in the system, a switching operation can be achieved by using the PT-Bragg grating at a single frequency. Subsequently,anon-linear PT-Bragg grating is investigated by using a time-domain numerical method, namely the Transmission-Line modelling (TLM) method. For the present work a TLM code is developed from scratch in order to ensure full-flexibility when modelling a dispersive and non-linear material. Using the TLM solver, it is demonstrated that gain/loss saturation is an important material property which should be considered as it may impact the practical applications of a PT-symmetry-based device. In the context of a non-linear PT-Bragg grating (NPTBG), the gain/loss saturation affects the interplay between the PT-symmetric opearation and the Kerr non-linear effect. It is further shown that gain/loss saturation plays a crucial role in securing a stable operation of non-linear PT-based devices. For practical applications, it is demonstrated that a non-linear PT-symmetric Bragg grating offers an additional degree of freedom in their operation,by modulating the gain/loss and the intensity of the input signal,compared to a passive structure which can only be manipulated by the input signal intensity. Two applications based on the interplay of PT-symmetric behaviour and Kerr non-linearity are demonstrated,namely a memory device and a logic-gate device. The second part of the thesis studies PT-symmetric resonator structures as a coupled system and as a periodic chain system. For these studies, a semi-analytical method based on the Boundary Integral Equation (BIE) method is developed and used together with a two-dimensional TLM method. The impact of realistic material parameter on the spectral properties of the structure is again investigated. It is shown that the PT-symmetric behaviour can be observed at a single frequency. Moreover, it is shown that PT-symmetry-like behaviour is observed but with complex eigenfrequencies due to the radiation losses; this is a deviation of the strict definition of a PT-symmetric structure with balanced gain and loss. Lowering lasing threshold by increasing loss in the system is demonstrated; this occurs due to induced early symmetry breaking. The final part of the thesis studies the spectral properties of an infinite and finite chain of PT-symmetric resonators. It is shown that the type of modulation along the PT-chain affects the position of the breaking point of the PT-structure. For a finite PT-chain structure, and for a particular type of refractive index modulation, early PT-symmetry breaking is observed and shown to cause the presence of termination states which are localised at the edge of the finite-chain resulting in localised lasing and dissipative modes at each end of the chain.

This thesis presents an investigation into the properties of III-VI metal chalcogenide semiconductor nanosheets and demonstrates their capability to enhance graphene-based optoelectronics. Strong quantization effects and tunable near-infrared-to-visible (NIR-to-VIS) photoluminescence emission are reported in mechanically exfoliated crystals of gamma-rhombohedral semiconducting InSe at room temperature. The optical properties of InSe nanosheets differ qualitatively from those reported for transition metal dichalcogenides and indicate a crossover from a direct-to-indirect band gap semiconductor when the InSe nanosheet thickness, L, is reduced to a few nanometres, corresponding to the emergence of a ‘Mexican hat’ energy dispersion for the valence band. At low temperature, radiative recombination of photoexcited carriers bound at native donors and acceptors in nominally undoped InSe nanosheets is observed. A two-dimensional hydrogenic model for impurities is used to describe the increase in binding energy with decreasing L and reveals a strong sensitivity of the binding energy on the position of the impurities within the nanolayer. The application of a magnetic field, B, perpendicular to the plane of InSe nanosheets induces a marked change of the observed optical spectrum. A transfer of intensity from a low-to-high energy component at high B corresponds to an indirect-to-direct band gap crossover, which arises from the Landau quantisation of the in-plane carrier motion and crossover between hole cyclotron orbits centred on closed edges of the valence band. High broad-band (NIR-to-VIS) photoresponsivity is achieved in mechanically formed InSe–graphene van der Waals heterostructures, which exploit the broad-band transparency of graphene, the direct bandgap of InSe, and the favourable band line up of n-InSe with graphene. The photoresponse is dependent on the electron transit time through the InSe layer, as evaluated by a semiclassical model.

Owing to their vitreous nature and mid-infrared (MIR) transparency, chalcogenide glasses are a promising material for remote hyperspectral imaging. For medical applications, such as in-vivo cancer diagnosis, Ge-Sb-Se glasses are a particularly interesting material since, it is believed that Sb-containing chalcogenides are less toxic than their As-containing equivalents. For passive optical fibres which aim to deliver and collect MIR light to and from tissue samples, the main challenge which faces their performance is the removal of extrinsic optical losses. Hence, this Project explores and develops high purity Ge-Sb-Se/S bulk glasses and optical fibres. Focussing on the GexSb10Se90-x atomic % (at. %) glass series, bulk samples are initially characterised before the fibre-drawing capability of each composition is assessed. Although stoichiometric Ge25Sb10Se65 at. % and non-stoichiometric Ge20Sb10Se70 at. % glasses both exist within the same two-dimensional, overconstrained network, results from their fibre-drawing investigations reveal a significant difference in their resistance against crystallisation. Whereas, non-stoichiometric Ge20Sb10Se70 at. % is shown to produce stable optical fibres with promising low losses, it is found that stoichiometric Ge25Sb10Se65 at. % repeatedly crystallises into a single phase of monoclinic GeSe2. To produce a low numerical aperture (NA) step-index fibre (SIF), it is suggested that a Ge20Sb10Se70 at. % core glass is paired with a Ge20Sb10Se67S3 at. % cladding glass. Substituting 3 at. % Se for 3 at. % S in the Ge20Sb10Se70-xSx at. % series, is found to increase the glass transition temperature (Tg) by 10 °C and decrease the refractive index by 0.01. It is calculated, that for a SIF consisting of a Ge20Sb10Se70 at. % core and a Ge20Sb10Se67S3 at. % cladding, the NA would be 0.25 at 3.1 µm wavelength. Co-extruded at 267±0.1 °C, and then subsequently drawn into 200 ±5 µm diameter SIF, optical loss measurements demonstrate that MIR light can be successfully guided through a large, circular Ge20Sb10Se70 at. % core. The core-cladding ratio is found to be 95 %. Calculations using the Antoine equation are used to investigate the optimal conditions required for the bake-out of Se, S and Sb impurities prior to batching. For a high-purity Ge20Sb10Se70 at. % core glass, a distillation technique is developed using 1000 ppm wt. TeCl4 as a hydrogen getter [H] and 700 ppm wt. Al as an oxygen getter [O]. It is shown, that to successfully distil Ge-Sb-Se glass, with [H] and [O] getters, two primary challenges must be overcome. The first suggests that there must be sufficient removal of HCl(g), prior to the start of distillation, in order to avoid a vapour barrier once the silica glass distillation rig is sealed. The second advises precise temperature control, with necessary monitoring, so that there is no separation of Ge-Sb-Se material, either before or after it has distilled. Successful distillation is eventually achieved in an open system i.e. under flowing vacuum, at a temperature close to 693 °C. Optical fibre loss measurements, conducted on 18 m length of 200 ±10 µm diameter fibre, reveal that the distillation of Ge20Sb10Se70 at. % core glass with 1000 ppm wt. TeCl4 and 700 ppm wt. Al, removes the Ge-O absorption peak at 7.9 µm and significantly reduces, if not removes, all of the Se-H peaks as well. The lowest background loss is also found as 0.44 dB/m at 6.4 µm wavelength. As a preliminary investigation into the biocompatibility of Ge-Sb-Se glasses for medical applications, two in-vitro cytotoxicity test are explored viz.: a direct contact protocol with an alamarBlue® assay and an elution protocol with a neutral red assay. Due to contradictory results between Trial 1 and Trial 2, it is suggested that further work is required to confirm the cytotoxicity of etched vs. non-etched Ge-Sb-Se fibres. Overall, there has been significant progress made during this Project, towards the fabrication of high purity Ge-Sb-Se/S SIFs for use in a MIR imaging probe for early cancer diagnosis.

Multi-atlas segmentation (MAS) has become an established technique for the automated delineation of anatomical structures. The often manually annotated labels from each of multiple pre-segmented images (atlases) are typically transferred to a target through the spatial mapping of corresponding structures of interest. The mapping can be estimated by pairwise registration between each atlas and the target or by creating an intermediate population template for spatial normalisation of atlases and targets. The former is done at runtime which is computationally expensive but provides high accuracy. In the latter approach the template can be constructed from the atlases offline requiring only one registration to the target at runtime. Although this is computationally more efficient, the composition of deformation fields can lead to decreased accuracy. Our goal was to develop a MAS method which was both efficient and accurate. In our approach we create a target-specific template (TST) which has a high similarity to the target and serves as intermediate step to increase registration accuracy. The TST is constructed from the atlas images that are most similar to the target. These images are determined in low-dimensional manifold spaces on the basis of deformation fields in local regions of interest. We also introduce a clustering approach to divide atlas labels into meaningful sub-regions of interest and increase local specificity for TST construction and label fusion. Our approach was tested on a variety of MR brain datasets and applied to an in-house dataset. We achieve state-of-the-art accuracy while being computationally much more efficient than competing methods. This efficiency opens the door to the use of larger sets of atlases which could lead to further improvement in segmentation accuracy.

Biological vision and computational models of vision can be split into three independent components (image description, decision process, and image set). The thesis presented here aimed to investigate the influence of each of these core components on computational model’s similarity to human behaviour. Chapter 3 investigated the similarity of different computational image descriptors to their biological counterparts, using an image matching task. The results showed that several of the computational models could explain a significant amount of the variance in human performance on individual images. The deep supervised convolutional neural net explained the most variance, followed by GIST, HMAX and then PHOW. Chapter 4 investigated which computational decision process best explained observers’ behaviour on an image categorization task. The results showed that Decision Bound theory produced behaviour the closest to that of observers. This was followed by Exemplar theory and Prototype theory. Chapter 5 examined whether the naturally differing image set between computational models and observers could partially account for the difference in their behaviour. The results showed that, indeed, the naturally differing image set between computational models and observers was affecting the similarity of their behaviour. This gap did not alter which image descriptor best fit observers’ behaviour and could be reduced by training observers on the image set the computational models were using. Chapter 6 investigated, using computational models of vision, the impact of the neighbouring (masking) images on the target images in a RSVP task. This was done by combining the neighbouring images with the target image for the computational models’ simulation for each trial. The results showed that models behaviour became closer to that of the human observers when the neighbouring mask images were included in the computational simulations, as would be expected given an integration period for neural mechanisms. This thesis has shown that computational models can show quite similar behaviours to human observers, even at the level of how they perform with individual images. While this shows the potential utility in computational models as a tool to study visual processing, It has also shown the need to take into account many aspects of the overall model of the visual process and task; not only the image description, but the task requirements, the decision processes, the images being used as stimuli and even the sequence in which they are presented.

Handling large displacement optical flow is a remarkably arduous task. For instance, standard coarse-to-fine techniques often struggle to adequately deal with moving objects whose motion exceeds their size. Here we propose a learning approach to the estimation of large displacement between two non-consecutive images in a sequence on the basis of a learning set of optical flows estimated a priori between different consecutive images in the same sequence. Our method refines an initial estimate of the flow field by replacing each displacement vector by a linear combination of displacement vectors at the center of similar patches taken from a code-book built from the learning set. The key idea is to use the accurate flows estimated a priori between consecutive images to help improve the potentially less accurate flows estimated online between images further apart. Experimental results suggest the ability of a purely data-driven learning approach to handle fine scale structures with large displacements.

Computer technology has rapidly improved over the last few years, with more powerful machines becoming ever smaller and cheaper. The latest growth area is in portable personal computers, providing powerful facilities to the mobile business person. Alongside this development has been the vast improvement to the human computer interface, allowing noncomputer- literate users access to computing facilities. These two aspects are now being combined into a portable computer that can be operated with a stylus, without the need for a keyboard. Handwriting is the obvious method for entering data and cursive script recognition research aims to comprehend unconstrained, natural handwriting. The ORCHiD system described in this thesis recognises connected handwriting collected on-line, in real time, via a digitising pad. After preprocessing, to remove any hardware-related errors, and normalising, the script is segmented and features of each segment measured. A new segmentation method has been developed which appears to be very consistent across a large number of handwriting styles. A statistical template matching algorithm is used to identify the segments. The system allows ambiguous matching, since cursive script is an ambiguous communications medium when taken out of context, and a probability for each match is calculated. These probabilities can be combined across the word to produce a ranked list of possible interpretations of the script word. A fast dictionary lookup routine has been developed enabling the sometimes very large list of possible words to be verified. The ORCHiD system can be trained, if desired, to a particular user. The training routine, however, is automatic since the untrained recognition system is used as the basis for the trained system. There is therefore very little start-up time before the system can be used. A decision-directed training approach is used. Recognition rates for the system vary depending on the consistency of the writing. On average, the untrained system achieved 75% recognition. After some training, average recognition rates of 91% were achieved, with up to 96% observed after further training.