The Ground-Based Synthetic Aperture Radar (GB-SAR) is a relatively new technique, which in the last ten years has gained interest as deformation measurement and monitoring tool. The GB-SAR technique is based on an imaging radar-based sensor, which o ers high sensitivity to small displacements, in the region of sub-millimetres to millimetres, long-range measurements, which can work up to some kilometres, and massive deformation measurement capability. These features confer to the GB-SAR technique interesting advantages with respect to other point-wise deformation measurement techniques. The process of estimating deformation from the GB-SAR data is not straightforward: it requires complex data processing and analysis tools. This dissertation is focused on these tools, covering the whole deformation estimation process. This thesis collects the main research results achieved on this topic during my work at the Active Remote Sensing Unit of the Institute of Geomatics. Two di erent approaches for measuring deformation with GB-SAR data are described and discussed. The irst one is the interferometric approach, based on the exploitation of the phase component of the GB-SAR data, which is the commonly used GB-SAR method. The second one is a non-interferometric approach, which exploits the amplitude component of the GB-SAR data, o ering an interesting alternative way to exploit the GB-SAR data. This dissertation has two main objectives. The first one is presenting, step by step, a complete interferometric GB-SAR procedure for deformation measuring and monitoring. The second one is presenting two new algorithms, which represent the most innovative part of this thesis. The first algorithm faces the phase unwrapping problem, providing an automatic solution for detecting and correcting unwrapping errors, which is called 2+1D phase unwrapping. The second algorithm is the base of the above mentioned non- interferometric approach, which overcomes some of the most critical limitation of GB-SAR interferometry, at the expense of getting less precise deformation estimates. The dissertation is divided in 6 chapters. The first one is the introduction, while the second one provides an overview of GB-SAR interferometry, introducing the main aspects that are the basics of the subsequent chapters. Chapter 3 describes a complete GB-SAR processing chain. Chapters 4 and 5 contain the most original part of the dissertation, i.e. the 2D+1 phase unwrapping algorithm, and the non-interferometric approach. Finally, in Chapter 6 the conclusions are discussed and further research is proposed. / El radar terrestre d’obertura sintètica (GB-SAR) és una tècnica relativament nova que, en els últims deu anys, ha guanyat interès com a eina per a mesurar i monitorar deformacions. La tècnica GB-SAR es basa en un sistema radar amb capacitat per proporcionar imatges, que ofereix una alta sensibilitat a petits desplaçaments, d’ordre mil·limètric o submil·limètric, que és capaç de mesurar a llargues distàncies (alguns km) i que té una alta capacitat per fer mesures massives. Aquestes característiques donen a la tècnica interessants avantatges respecte a altres tècniques clàssiques de mesura de deformacions, típicament basades en mesures puntuals. Derivar mesures de deformació a partir de dades GB-SAR no és un procés senzill, ja que requereix uns procediments complexos de processat i anàlisi de dades. Aquesta tesi es centra en aquests processos. Aquesta tesi recull alguns dels resultats més destacats de la investigació que he desenvolupat sobre aquest tema a la unitat de Teledetecció Activa de l'Institut de Geomàtica. Al llarg del document es descriuen dues aproximacions diferents per mesurar deformacions amb GB-SAR. Una es basa en la explotació de la tècnica de la interferometria, és a dir explotant la component de la fase de les imatges GB-SAR: és la tècnica GB-SAR usada habitualment. La segona, anomenada tècnica no-interferomètrica, es basa en la component de l’amplitud de les dades GB-SAR i ofereix una interessant alternativa a la primera. La tesi acompleix dos objectius principals. En primer lloc presenta un procediment complet per la mesura i monitoratge de deformacions mitjançant interferometria GB-SAR. En segon lloc, descriu dos nous algorismes que resolen problemes específics de la interferometria clàssica aplicada al GB-SAR i que representen la part més innovadora d’aquesta tesi. El primer algorisme aborda un dels problemes oberts de la interferometria, el phase unwrapping, proposant un mètode automàtic per detectar-ne i corregir-ne els errors. El segon algorisme proposa un nou mètode per a l'explotació de les dades GB-SAR per mesurar deformacions sense utilitzar la interferometria. La estructura de la tesi consisteix en sis capítols. Després de la introducció, el Capítol 2 proporciona una visió general de la interferometria GB-SAR, introduint els conceptes principals utilitzats en la tesi. En el tercer capítol es descriu una cadena de processament basada en GB-SAR interferomètric. Els capítols quart i cinquè contenen la part més original de la tesi: l'algorisme de phase unwrapping i el mètode no-interferomètric per la mesura de deformacions. Finalment, es discuteixen les conclusions principals i es proposen futures línies d’investigació.

Synthetic Aperture Radar (SAR) imaging is a well-known technique in the domain of remote sensing, aerospace surveillance, geography and mapping. To obtain images of high resolution under noise, taking into account of the characteristics of targets in the observed scene, the different uncertainties of measure and the modeling errors becomes very important.Conventional imaging methods are based on i) over-simplified scene models, ii) a simplified linear forward modeling (mathematical relations between the transmitted signals, the received signals and the targets) and iii) using a very simplified Inverse Fast Fourier Transform (IFFT) to do the inversion, resulting in low resolution and noisy images with unsuppressed speckles and high side lobe artifacts.In this thesis, we propose to use a Bayesian approach to SAR imaging, which overcomes many drawbacks of classical methods and brings high resolution, more stable images and more accurate parameter estimation for target recognition.The proposed unifying approach is used for inverse problems in Mono-, Bi- and Multi-static SAR imaging, as well as for micromotion target imaging. Appropriate priors for modeling different target scenes in terms of target features enhancement during imaging are proposed. Fast and effective estimation methods with simple and hierarchical priors are developed. The problem of hyperparameter estimation is also handled in this Bayesian approach framework. Results on synthetic, experimental and real data demonstrate the effectiveness of the proposed approach.

Texture analysis has been used extensively in the computer-assisted interpretation of SAR sea ice imagery. Provision of maps which distinguish relevant ice types is significant for monitoring global warming and ship navigation. Due to the abundance of SAR imagery available, there exists a need to develop an automated approach for SAR sea ice interpretation. Grey level co-occurrence probability (<i>GLCP</i>) texture features are very popular for SAR sea ice classification. Although these features are used extensively in the literature, they have a tendency to erode and misclassify texture boundaries. Proposed is an advancement to the <i>GLCP</i> method which will preserve texture boundaries during image segmentation. This method exploits the relationship a pixel has with its closest neighbors and weights the texture measurement accordingly. These texture features are referred to as <i>WGLCP</i> (weighted <i>GLCP</i>) texture features. In this research, the <i>WGLCP</i> and <i>GLCP</i> feature sets are compared in terms of boundary preservation, unsupervised segmentation ability, robustness to increasing boundary density and computation time. The <i>WGLCP</i> method outperforms the <i>GLCP</i> method in all aspects except for computation time, where it suffers. From the comparative analysis, an inconsistency with the <i>GLCP</i> correlation statistic was observed, which motivated an investigative study into using this statistic for image segmentation. As the overall goal of the thesis is to improve SAR sea ice segmentation accuracy, the concepts developed from the study are applied to the image segmentation problem. The results indicate that for images with high contrast boundaries, the <i>GLCP</i> correlation statistical feature decreases segmentation accuracy. When comparing <i>WGLCP</i> and <i>GLCP</i> features for segmentation, the <i>WGLCP</i> features provide higher segmentation accuracy.

The mapping of sea ice is an important task for understanding global climate and for safe shipping. Currently, sea ice maps are created by human analysts with the help of remote sensing imagery, including synthetic aperture radar (SAR) imagery. While the maps are generally correct, they can be somewhat subjective and do not have pixel-level resolution due to the time consuming nature of manual segmentation. Therefore, automated sea ice mapping algorithms such as the multivariate iterative region growing with semantics (MIRGS) sea ice image segmentation algorithm are needed. MIRGS was designed to work with one-channel single-polarization SAR imagery from the RADARSAT-1 satellite. The launch of RADARSAT-2 has made available two-channel dual-polarization SAR imagery for the purposes of sea ice mapping. Dual-polarization imagery provides more information for distinguishing ice types, and one of the channels is less sensitive to changes in the backscatter caused by the SAR incidence angle parameter. In the past, this change in backscatter due to the incidence angle was a key limitation that prevented automatic segmentation of full SAR scenes. This thesis investigates techniques to make use of the dual-polarization data in MIRGS. An evaluation of MIRGS with RADARSAT-2 data was performed and showed that some detail was lost and that the incidence angle caused errors in segmentation. Several data fusion schemes were investigated to determine if they can improve performance. Gradient generation methods designed to take advantage of dual-polarization data, feature space fusion using linear and non-linear transforms as well as image fusion methods based on wavelet combination rules were implemented and tested. Tuning of the MIRGS parameters was performed to find the best set of parameters for segmentation of dual-polarization data. Results show that the standard MIRGS algorithm with default parameters provides the highest accuracy, so no changes are necessary for dual-polarization data. A hierarchical segmentation scheme that segments the dual-polarization channels separately was implemented to overcome the incidence angle errors. The technique is effective but requires more user input than the standard MIRGS algorithm.

Since the 1980s, satellite-borne synthetic aperture radar (SAR) has been investigated for early warning and monitoring of marine oil spills to permit effective satellite surveillance in the marine environment. Automated detection of oil spills from satellite SAR intensity imagery consists of three steps: 1) Detection of dark spots; 2) Extraction of features from the detected dark spots; and 3) Classification of the dark spots into oil spills and look-alikes. However, marine oil spill detection is a very difficult and challenging task. Open questions exist in each of the three stages. In this thesis, the focus is on the first stage—dark spot detection. An efficient and effective dark spot detection method is critical and fundamental for developing an automated oil spill detection system. A novel method for this task is presented. The key to the method is utilizing the spatial density feature to enhance the separability of dark spots and the background. After an adaptive intensity thresholding, a spatial density thresholding is further used to differentiate dark spots from the background. The proposed method was applied to a evaluation dataset with 60 RADARSAT-1 ScanSAR Narrow Beam intensity images containing oil spill anomalies. The experimental results obtained from the test dataset demonstrate that the proposed method for dark spot detection is fast, robust and effective. Recommendations are given for future research to be conducted to ensure that this procedure goes beyond the prototype stage and becomes a practical application.

Ultra-wideband (UWB) technology and its use in imaging and sensing have drawnsignicant interest in the last two decades. Extensive studies have contributed toutilize UWB transient scattering for automated target recognition and imagingpurposes. In this thesis a near-eld UWB synthetic aperture radar (SAR) imagingalgorithm is presented.It is shown with measurements and simulation, that it is possible to reconstruct an imageof an object in the near eld region using UWB technology and SAR imaging algorithm.However the nal SAR image is highly aected by unwanted scattered elds at each pixelusually observed as an image artifact in the nal image. In this study these artifactsare seen as a smile around the main object. Two methods are suggested in this thesiswork to suppress this artifact. The rst method combines the scattered eld informationreceived from both rear and front of the object to reconstruct two separate images, onefrom rear view and one from front view of the object respectively. Since the scatteredelds from behind the object are mirrored, the pixel by pixel multiplication of thesetwo images for objects with simple geometry will cancel the artifact. This method isvery simple and fast applicable to objects with simple geometry. However this methodcannot be used for objects with rather complex geometry and boundaries. Thereforethe Range Point Migration (RPM) method is used along with the global characteristicsof the observed range map to introduce a new artifact rejection method based on thedirectional of arrival (DOA) of scattered elds at each pixel. DOA information can beused to calculate an optimum theta for each antenna. This optimum angle along withthe real physical direction of arrival at each position can produce a weighting factor thatlater can be used to suppress the eect of undesired scattered elds producing the smileshaped artifact. Final results of this study clearly show that the UWB SAR accompaniedwith DOA can produce an image of an object free of undesired artifact from scatteredeld of adjacent antennas.

Texture analysis has been used extensively in the computer-assisted interpretation of SAR sea ice imagery. Provision of maps which distinguish relevant ice types is significant for monitoring global warming and ship navigation. Due to the abundance of SAR imagery available, there exists a need to develop an automated approach for SAR sea ice interpretation. Grey level co-occurrence probability (<i>GLCP</i>) texture features are very popular for SAR sea ice classification. Although these features are used extensively in the literature, they have a tendency to erode and misclassify texture boundaries. Proposed is an advancement to the <i>GLCP</i> method which will preserve texture boundaries during image segmentation. This method exploits the relationship a pixel has with its closest neighbors and weights the texture measurement accordingly. These texture features are referred to as <i>WGLCP</i> (weighted <i>GLCP</i>) texture features. In this research, the <i>WGLCP</i> and <i>GLCP</i> feature sets are compared in terms of boundary preservation, unsupervised segmentation ability, robustness to increasing boundary density and computation time. The <i>WGLCP</i> method outperforms the <i>GLCP</i> method in all aspects except for computation time, where it suffers. From the comparative analysis, an inconsistency with the <i>GLCP</i> correlation statistic was observed, which motivated an investigative study into using this statistic for image segmentation. As the overall goal of the thesis is to improve SAR sea ice segmentation accuracy, the concepts developed from the study are applied to the image segmentation problem. The results indicate that for images with high contrast boundaries, the <i>GLCP</i> correlation statistical feature decreases segmentation accuracy. When comparing <i>WGLCP</i> and <i>GLCP</i> features for segmentation, the <i>WGLCP</i> features provide higher segmentation accuracy.

The mapping of sea ice is an important task for understanding global climate and for safe shipping. Currently, sea ice maps are created by human analysts with the help of remote sensing imagery, including synthetic aperture radar (SAR) imagery. While the maps are generally correct, they can be somewhat subjective and do not have pixel-level resolution due to the time consuming nature of manual segmentation. Therefore, automated sea ice mapping algorithms such as the multivariate iterative region growing with semantics (MIRGS) sea ice image segmentation algorithm are needed. MIRGS was designed to work with one-channel single-polarization SAR imagery from the RADARSAT-1 satellite. The launch of RADARSAT-2 has made available two-channel dual-polarization SAR imagery for the purposes of sea ice mapping. Dual-polarization imagery provides more information for distinguishing ice types, and one of the channels is less sensitive to changes in the backscatter caused by the SAR incidence angle parameter. In the past, this change in backscatter due to the incidence angle was a key limitation that prevented automatic segmentation of full SAR scenes. This thesis investigates techniques to make use of the dual-polarization data in MIRGS. An evaluation of MIRGS with RADARSAT-2 data was performed and showed that some detail was lost and that the incidence angle caused errors in segmentation. Several data fusion schemes were investigated to determine if they can improve performance. Gradient generation methods designed to take advantage of dual-polarization data, feature space fusion using linear and non-linear transforms as well as image fusion methods based on wavelet combination rules were implemented and tested. Tuning of the MIRGS parameters was performed to find the best set of parameters for segmentation of dual-polarization data. Results show that the standard MIRGS algorithm with default parameters provides the highest accuracy, so no changes are necessary for dual-polarization data. A hierarchical segmentation scheme that segments the dual-polarization channels separately was implemented to overcome the incidence angle errors. The technique is effective but requires more user input than the standard MIRGS algorithm.

Since the 1980s, satellite-borne synthetic aperture radar (SAR) has been investigated for early warning and monitoring of marine oil spills to permit effective satellite surveillance in the marine environment. Automated detection of oil spills from satellite SAR intensity imagery consists of three steps: 1) Detection of dark spots; 2) Extraction of features from the detected dark spots; and 3) Classification of the dark spots into oil spills and look-alikes. However, marine oil spill detection is a very difficult and challenging task. Open questions exist in each of the three stages. In this thesis, the focus is on the first stage—dark spot detection. An efficient and effective dark spot detection method is critical and fundamental for developing an automated oil spill detection system. A novel method for this task is presented. The key to the method is utilizing the spatial density feature to enhance the separability of dark spots and the background. After an adaptive intensity thresholding, a spatial density thresholding is further used to differentiate dark spots from the background. The proposed method was applied to a evaluation dataset with 60 RADARSAT-1 ScanSAR Narrow Beam intensity images containing oil spill anomalies. The experimental results obtained from the test dataset demonstrate that the proposed method for dark spot detection is fast, robust and effective. Recommendations are given for future research to be conducted to ensure that this procedure goes beyond the prototype stage and becomes a practical application.

The synthetic aperture radar (SAR) onboard Earth observing satellites has been acknowledged as an integral tool for many applications in monitoring the marine environment. Some of these applications include regional sea-ice monitoring and detection of illegal or accidental oil discharges from ships. Nonetheless, a practicality of the usage of SAR images is greatly hindered by the presence of speckle noises. Such noise must be eliminated or reduced to be utilized in real-world applications to ensure the safety of the marine environment. Thus this thesis presents a novel two-phase total variation optimization segmentation approach to tackle such a challenging task. In the total variation optimization phase, the Rudin-Osher-Fatemi total variation model was modified and implemented iteratively to estimate the piecewise smooth state by minimizing the total variation constraints. In the finite mixture model classification phase, an expectation-maximization method was performed to estimate the final class likelihoods using a Gaussian mixture model. Then a maximum likelihood classification technique was utilized to obtain the final segmented result. For its evaluation, a synthetic image was used to test its effectiveness. Then it was further applied to two distinct real SAR images, X-band COSMO-SkyMed imagery containing verified oil-spills and C-band RADARSAT-2 imagery mainly containing two different sea-ice types to confirm its robustness. Furthermore, other well-established methods were compared with the proposed method to ensure its performance. With the advantage of a short processing time, the visual inspection and quantitative analysis including kappa coefficients and F1 scores of segmentation results confirm the superiority of the proposed method over other existing methods.