Le projet INDESO (Développement de l'océanographie spatiale en Indonésie) en collaboration entre le gouvernement indonésien (Ministère des affaires maritimes et des pêches - MMAF) et la société française CLS (Collecte Localisation Satellite) promeut l'utilisation des technologies spatiales pour la surveillance des côtes et des mers indonésiennes. Cette thèse fait partie du volet sur la surveillance des récifs coralliens, mené par l'IRD (Institut de Recherche pour le Développement). L'objectif principal était de déterminer si les habitats des récifs coralliens dans l'île de Bunaken dans le nord de Sulawesi sont résilients, en utilisant i) des cartes d'habitat nouvellement conçues, ii) des données in situ et une série chronologique de 15 ans unique d'images satellites de différents capteurs très haute résolution (THR), iii) des données auxiliaires qui pourraient expliquer les changements détectés. Les résultats comprennent des cartes très détaillées de l'habitat des récifs de Bunaken (194 polygones cartographiés et un recensement de 175 habitats). L'influence de la chute du niveau de la mer sur la mortalité des coraux pendant l'événement El Niño de 2015-2016 est présentée en détail, et l'importance de ce processus est également discutée à partir de l'interprétation d'une série chronologique unique de 15 ans d'images THR. La série temporelle met en évidence des trajectoires très différentes des habitats coralliens. Nous avons conclu que le récif de Bunaken démontre une capacité de résilience et sans déphasage, mais qu'un diagnostic définitif de sa résilience reste difficile à déterminer par imagerie. Des trajectoires de l'habitat ne peuvent pas être totalement interprétées sans changer certains paradigmes de surveillance, et sans utiliser une combinaison d'observations de télédétection et de données in situ. / The INDESO (Infrastructure Development of Space Oceanography) project, in collaboration with the Indonesian Government (Ministry of Marine Affairs and Fisheries - MMAF) and the French company CLS (Collecte Localisation Satellites), promotes the use of space technologies for monitoring coastlines and Indonesian seas. This thesis is part of coral reef monitoring component, led by the IRD (Institute de Recherche pour le Développement). The main objective was to determine wether coral reef habitats on Bunaken Island in Northern Sulawesi are resilient, using (i) newly desgined habitat maps, (ii) in situ data, and a unique 15-year time series of satellite images of different very high resolution (VHR) sensors, and (iii) ancillary data that could explain the changes detected. The results include highly detailed maps of the Bunaken reefs habitat (194 polygons mapped and a census of 175 habitats). The influence of sea level fall on coral mortality during the El-Nino event of 2015 - 2016 is presented in detail, and the importance of this process is also discussed from the interpretation of a unique time series of 15 years of VHR images. The temporal series reveals very different trajectories of the coral habitats. We conclude that Bunaken reefs demonstrate an ability to resileince and without phase shift, but that a definitive diagnosis of their resilience remains difficult to determine by imagery. Habitat trajectories can not be fully interpreted without changing some monitoring paradigms and without using a combination of remote sensing and in situ data.

Detection of structural changes in time series is a topic with increasing pop- ularity among econometricians over the last decades. The main aim of this thesis was to review and compare the classical and modern econometric meth- ods of structural change detection and unit root testing. A recent method for testing a one-time break in at most linear trend function of a series without prior knowledge about the stationary or unit root nature of the error compo- nent proposed by Perron and Yabu (2009b) was studied. Subsequently, it was combined with the unit root test that allows for a break in trend proposed by Kim and Perron (2009) to examine the nature of the error component. All the methods for change detection and unit root testing were compared in a Monte Carlo simulation study that indicated significant improvement in power of the Perron-Yabu and Kim-Perron tests against most alternatives compared to the classical methods. However, all tests demonstrated poor performance in case of a quadratic trend function. Finally, the tests were employed in a practical ex- ample to examine the properties of the quarterly GDP time series of the Czech Republic. 1

Driving a motor vehicle is a complex activity, and errors in performing the driving task can result in crashes which cause property damage, injuries, and sometimes death. It is important that the road environment supports drivers in safe performance of the driving task. At present, increasing amounts of visual information from sources such as roadside advertising create visual clutter in the road environment. There has been little research on the effect of this visual clutter on driving performance, particularly for vulnerable groups such as novice and older drivers. The present work aims to fill this gap. Literature from a variety of relevant disciplines was surveyed and integrated, and a model of the mechanisms by which visual clutter could affect performance of the driving task was developed. To determine potential sources of clutter, focus groups with drivers were held and two studies involving subjective ratings of visual clutter in photographs and video clips of road environments were carried out. This resulted in a taxonomy of visual clutter in the road environment: ‘situational clutter’, including vehicles and other road users with whom drivers interact; ‘designed clutter’, including road signs, signals, and markings used by traffic authorities to communicate with users; and ‘built clutter’, including roadside development and any signage not originating from a road authority. The taxonomy of visual clutter was tested using the change detection paradigm. Drivers were slower to detect changes in photographs of road scenes with high levels of visual clutter than with low levels, and slower for road scenes including advertising billboards than road scenes without billboards. Finally, the effects of billboard presence and lead vehicles on vehicle control, eye movements and responses to traffic signs and signals were tested using a driving simulator. The number of vehicles included appeared to be insufficient to create situational clutter. However billboards had significant effects on driver speed (slower), ability to follow directions on road signs (slower with more errors), and eye movements (increased amount of time fixating on roadsides at the expense of scanning the road ahead). Older drivers were particularly affected by visual clutter in both the change detection and simulated driving tasks. Results are discussed in terms of implications for future research and for road safety practitioners. Visual clutter can affect driver workload as well as purely visual aspects of the driving task (such as hazard perception and search for road signs). When driver workload is increased past a certain point other driving tasks will also be performed less well (such as speed maintenance). Advertising billboards in particular cause visual distraction, and should be considered at a similar level of potential danger as visual distraction from in-vehicle devices. The consequences of roadside visual clutter are more severe for the growing demographic of older drivers. Currently, road environments do not support drivers (particularly older drivers) as well as they could. Based on the results, guidance is given for road authorities to improve this status when designing and location road signage and approving roadside advertising.

<p>Robust localization is a prerequisite for mobile robot autonomy. In many situations the GPS signal is not available and thus an additional localization system is required. A simple approach is to apply localization based on dead reckoning by use of wheel encoders but it results in large estimation errors. With exteroceptive sensors such as a laser range finder natural landmarks in the environment of the robot can be extracted from raw range data. Landmarks are extracted with the Hough transform and a recursive line segment algorithm. By applying data association and Kalman filtering along with process models the landmarks can be used in combination with wheel encoders for estimating the global position of the robot. If several robots can cooperate better position estimates are to be expected because robots can be seen as mobile landmarks and one robot can supervise the movement of another. The centralized Kalman filter presented in this master thesis systematically treats robots and extracted landmarks such that benefits from several robots are utilized. Experiments in different indoor environments with two different robots show that long distances can be traveled while the positional uncertainty is kept low. The benefit from cooperating robots in the sense of reduced positional uncertainty is also shown in an experiment. </p><p>Except for localization algorithms a typical autonomous robot task in the form of change detection is solved. The change detection method, which requires robust localization, is aimed to be used for surveillance. The implemented algorithm accounts for measurement- and positional uncertainty when determining whether something in the environment has changed. Consecutive true changes as well as sporadic false changes are detected in an illustrative experiment.</p>

This thesis, Earth satellites and detection of air andground based activities by Ulf Ekblad of the Physics departmentat the Royal Institute of Technology (KTH), addresses theproblem of detecting military activities in imagery. Examplesof various techniques are presented. In particular, problemsassociated with "novelties" and "changes" in an image arediscussed and various algorithms presented. The imagery usedincludes satellite imagery, aircraft imagery, and photos offlying aircraft. The timely delivery of satellite imagery is limited by thelaws of celestial mechanics. This and other information aspectsof imagery are treated. It is e.g. shown that dozens ofsatellites may be needed if daily observations of a specificsite on Earth are to be conducted from low Earth orbit. New findings from bioinformatics and studies of small mammalvisual systems are used. The Intersecting Cortical Model (ICM),which is a reduced variant of the Pulse-Coupled Neural Network(PCNN), is used on various problems among which are changedetection. Still much more could be learnt from biologicalsystems with respect to pre- and post-processing as well asintermediate processing stages. Simulated satellite imagery is used for determining theresolution limit for detection of tanks. The necessary pixelsize is shown to be around 6 m under the conditions of thissimulation. Difference techniques are also tested on Landsat satelliteimagery with the purpose of detecting underground nuclearexplosions. In particular, it is shown that this can easily bedone with 30 m resolution images, at least in the case studied.Satellite imagery from SPOT is used for detecting undergroundnuclear explosions prior to the detonations, i.e. under certainconditions 10 m resolution images can be used to detectpreparations of underground nuclear explosions. This type ofinformation is important for ensuring the compliance of nucleartest ban treaties. Furthermore, the necessity for havingcomplementary information in order to be able to interpretimages is also shown. Keywords: Remote sensing, reconnaissance, sensor,information acquisition, satellite imagery, image processing,image analysis, change detection, pixel difference, neuronnetwork, cortex model, PCNN, ICM, entanglement, Earthobservation, nuclear explosion, SPOT, Landsat, verification,orbit.

Robust localization is a prerequisite for mobile robot autonomy. In many situations the GPS signal is not available and thus an additional localization system is required. A simple approach is to apply localization based on dead reckoning by use of wheel encoders but it results in large estimation errors. With exteroceptive sensors such as a laser range finder natural landmarks in the environment of the robot can be extracted from raw range data. Landmarks are extracted with the Hough transform and a recursive line segment algorithm. By applying data association and Kalman filtering along with process models the landmarks can be used in combination with wheel encoders for estimating the global position of the robot. If several robots can cooperate better position estimates are to be expected because robots can be seen as mobile landmarks and one robot can supervise the movement of another. The centralized Kalman filter presented in this master thesis systematically treats robots and extracted landmarks such that benefits from several robots are utilized. Experiments in different indoor environments with two different robots show that long distances can be traveled while the positional uncertainty is kept low. The benefit from cooperating robots in the sense of reduced positional uncertainty is also shown in an experiment. Except for localization algorithms a typical autonomous robot task in the form of change detection is solved. The change detection method, which requires robust localization, is aimed to be used for surveillance. The implemented algorithm accounts for measurement- and positional uncertainty when determining whether something in the environment has changed. Consecutive true changes as well as sporadic false changes are detected in an illustrative experiment.

Land degradation is one of the most important drivers of landscape change around the globe. This dissertation examines land use-land cover change within a mosaic landscape in Eastern Terai, India, and shows evidence of anthropogenic factors contributing to landscape change. Land use and land cover change were examined within the Alipurduar Subdivision, a representative of the Eastern Terai landscape and the Jaldapara Wildlife Sanctuary, a protected area nested within Alipurduar through the use of multi-temporal satellite data over the past 28 years (1978 - 2006). This study establishes the potential of remote sensing technology to identify the drivers of landscape change; it provides an assessment of how regional drivers of landscape change influence the change within smaller local study extents and provides a methodology to map different types of grassland and monitor their loss within the region. The Normalized Difference Vegetation Index (NDVI) and a Normalized Difference Dry Index (NDDI) were found instrumental in change detection and the classification of different grasslands found inside the park based on their location, structure, and composition. Successful spectral segregation of different types of grasslands and their direct association with different grassland specialist species (e.g., hispid hare, hog deer, Bengal florican) clearly showed the potential of remote sensing technology to efficiently monitor these grasslands and assist in species conservation. Temporal analysis provided evidence of the loss of dense forest and grasslands within both study areas with a considerably higher rate of loss outside the protected area than inside. Results show a decline of forest from 40% in 1978 to 25% in 2006 across Alipurduar. Future trends project forest cover and grassland within Alipurduar to reduce to 15% and 5%, respectively. Within the Alipurduar, deforestation due to growth of tea industry was the primary driver of change. Flooding changed the landscape, but more intensely inside the wildlife preserve. Change of the river course inside Jaldapara during the flood of 1968 significantly altered the distribution of grassland inside the park. Unless, the direction of landscape change is altered, future trends predict growth of the tea industry within the region, increased forest loss, and homogenization of the landscape.

Knowledge about land use and land cover represents an important information basis for various planning applications. In particular, urban and suburban regions are subject to a high dynamic development. The detection and identification of changes is therefore an important instrument to follow and accompany the developments by planning. Here, aerial photography and, increasingly, satellite images serve as an important basis for information. The recognition and mapping of changes is still a time-consuming and cost-intensive matter which is mostly realized by visual interpretation of aerial photography and to an increasing degree of high- and ultra-high-resolution satellite images. Within the scope of the present work a new, robust and largely automated process based on a statistical change analysis is developed and presented. Basis for the data are multitemporal high-resolution satellite image data. The generated suspect areas, respectively areas of change, are supposed to function as clues in order to facilitate the process of the visual interpretation of multitemporal image datasets with regard to change mapping, since only marked areas of change have to undergo further examination. Consequently, this process can be used as a tool to ease and accelerate the updating of planning bases in general and maps in particular so far realised by visual interpretation. However, the automation of the process is not only supposed to serve the purpose of saving time and cost but also to bring the interpretation process to a higher level of objectivity. In order to improve the quality of the whole process, for the preprocessing of the image data selected methods of image processing have been integrated. Through the use of additional geo-information reference data for the automated calculation of the areas of change, a further refinement of the results can be reached. The obtained results in the first time-cut (1997-1998) can be proved and verified by a different data-take (1997-2000). To reach a convenient use and a good distribution of the developed method, the process has been implemented by means of the widespread image processing software ERDAS IMAGINE. This allows to make the developed method available for other users, since it can easily be integrated into the working environment of ERDAS IMAGINE. / Das Wissen um die Landnutzung und Landbedeckung ist für planerische Anwendungsgebiete eine wichtige Informationsgrundlage. Gerade urbane und suburbane Regionen unterliegen einer hohen Entwicklungsdynamik. Das Erkennen und Aufzeigen von Veränderungen ist somit ein wichtiges Instrument um Entwicklungen zu verfolgen und planerisch zu begleiten. Luft- und zunehmend Satellitenbilder dienen hierfür als wichtige Informationsgrundlage. Das Erkennen und Kartieren von Veränderungen ist nach wie vor eine zeitaufwändige und kostenintensive Angelegenheit, die überwiegend durch visuelle Interpretation von Luft- und zunehmend auch mit hoch- und höchstauflösenden Satellitenbildern realisiert wird. In dieser Arbeit wird ein neues, robustes, weitgehend automatisiertes, auf einem statistischen Ansatz beruhendes Verfahren der Veränderungsanalyse entwickelt und vorgestellt. Die Datengrundlage bilden multitemporale, hoch auflösende Satellitenbilddaten. Die generierten Verdachts- bzw. Veränderungsflächen sollen als Anhaltspunkte fungieren, um den Prozess der visuellen Interpretation von multitemporalen Bilddatensätzen in Hinsicht auf eine Veränderungskartierung zu erleichtern, da nur als Veränderungsflächen markierte Areale einer weiteren Untersuchung unterzogen werden müssen. Das Verfahren kann somit als Werkzeug dienen, die durch visuelle Interpretation realisierte Aktualisierung von Planungsgrundlagen bzw. Kartenwerken zu erleichtern und zu beschleunigen. Die Automatisierung des Verfahrens soll jedoch nicht allein dem Zweck der Zeit- und Kostenersparnis dienen, sondern auch den Interpretationsprozess objektiver gestalten. Um die Qualität des Verfahrens zu erhöhen, werden ausgewählte Methoden der Bildverarbeitung für die Vorverarbeitung der Bilder in das Verfahren integriert. Durch das Einbinden zusätzlicher Geobasisdaten in die automatisierte Berechnung der Veränderungsflächen kann eine weitere Verbesserung der Ergebnisse erzielt werden. Die Ergebnisse, der im ersten Zeitschnitt (1997-1998) untersuchten Datensätze, werden mit Hilfe eines weiteren Zeitschnitts (1997-2000) überprüft und verifiziert. Um eine unkomplizierte Anwendung und Verbreitung der Methode zu erreichen, wurde das Verfahren mit Hilfe der weit verbreiteten Bildverarbeitungssoftware ERDAS IMAGINE realisiert. Dies ermöglicht, das Verfahren auch anderen Nutzern zur Verfügung zu stellen, da es problemlos in die Arbeitsumgebung des Bildverarbeitungssystems ERDAS IMAGINE integriert werden kann

<p>This thesis, Earth satellites and detection of air andground based activities by Ulf Ekblad of the Physics departmentat the Royal Institute of Technology (KTH), addresses theproblem of detecting military activities in imagery. Examplesof various techniques are presented. In particular, problemsassociated with "novelties" and "changes" in an image arediscussed and various algorithms presented. The imagery usedincludes satellite imagery, aircraft imagery, and photos offlying aircraft.</p><p>The timely delivery of satellite imagery is limited by thelaws of celestial mechanics. This and other information aspectsof imagery are treated. It is e.g. shown that dozens ofsatellites may be needed if daily observations of a specificsite on Earth are to be conducted from low Earth orbit.</p><p>New findings from bioinformatics and studies of small mammalvisual systems are used. The Intersecting Cortical Model (ICM),which is a reduced variant of the Pulse-Coupled Neural Network(PCNN), is used on various problems among which are changedetection. Still much more could be learnt from biologicalsystems with respect to pre- and post-processing as well asintermediate processing stages.</p><p>Simulated satellite imagery is used for determining theresolution limit for detection of tanks. The necessary pixelsize is shown to be around 6 m under the conditions of thissimulation.</p><p>Difference techniques are also tested on Landsat satelliteimagery with the purpose of detecting underground nuclearexplosions. In particular, it is shown that this can easily bedone with 30 m resolution images, at least in the case studied.Satellite imagery from SPOT is used for detecting undergroundnuclear explosions prior to the detonations, i.e. under certainconditions 10 m resolution images can be used to detectpreparations of underground nuclear explosions. This type ofinformation is important for ensuring the compliance of nucleartest ban treaties. Furthermore, the necessity for havingcomplementary information in order to be able to interpretimages is also shown.</p><p>Keywords: Remote sensing, reconnaissance, sensor,information acquisition, satellite imagery, image processing,image analysis, change detection, pixel difference, neuronnetwork, cortex model, PCNN, ICM, entanglement, Earthobservation, nuclear explosion, SPOT, Landsat, verification,orbit.</p>

Les forêts de Bornéo sont inestimables. En plus d’une faune et d’une flore riche et diversifiée, ses milieux naturels constituent d’efficaces réservoirs de carbone. En outre, la matière ligneuse qui y est abondante fait l’objet d’une exploitation intensive. Par contre, c’est le potentiel agricole de l’île qui crée le plus d’enthousiasme, principalement en ce qui concerne la culture du palmier à huile. Pour tenter de mieux comprendre et surveiller le phénomène, nous avons développé des méthodes de détection de la déforestation et de la dégradation des forêts. Ces méthodes doivent tenir compte des caractéristiques propres à l’île. C’est que Bornéo est abondamment affectée par une nébulosité constante qui complexifie considérablement son observation à partir des satellites. Malgré ces contraintes, nous avons produit une série chronologique annuelle des points chauds de déforestation et de dégradation des forêts pour les années 2000 à 2009. / Borneo’s forests are priceless. Beyond the richness and diversity of its fauna and flora, its natural habitats constitute efficient carbon reservoirs. Unfortunately, the vast forests of the island are rapidly being cut down, both by the forestry industry and the rapidly expanding oil palm industry. In this context, we’ve developed methods to detect deforestation and forest degradation in order to better understand and monitor the phenomena. In doing so, the peculiarities of Borneo, such as the persistent cloud cover, had to be accounted for. Nevertheless, we succeeded in producing a time series of the yearly forest degradation and deforestations hotspots for the year 2000 through the year 2009. / Ce travail s’inscrit dans le cadre d’un programme de recherches appuyé par le Conseil de recherches en sciences humaines du Canada.