The topic of my diploma thesis is a comparison of speculation about a film from two perspectives. Firstly, it deals with the approach of Jan Mukařovský, a Czech structuralist, who was one of the first in the Czech context to explore the time concept in a film. Secondly, it considers the approach of Gillese Deleuze, who was accredited with the new view on the time-image in a film, however, much later than Jan Mukařovský. The aim of the paper is to explain the way the approaches of Mukařovský and Deluze relate to other approaches of thinking about time in films, the influence of other concepts of their time that they both theoretically based their views on, as well as specifying the films that became the starting points of exploring the time in films. The method of comparison is divided into three aspects, as three comparison levels.

Satellite Image Time Series (SITS) are becoming available at high spatial, spectral and temporal resolutions across the globe by the latest remote sensing sensors. These series of images can be highly valuable when exploited by classification systems to produce frequently updated and accurate land cover maps. The richness of spectral, spatial and temporal features in SITS is a promising source of data for developing better classification algorithms. However, machine learning methods such as Random Forests (RFs), despite their fruitful application to SITS to produce land cover maps, are structurally unable to properly handle intertwined spatial, spectral and temporal dynamics without breaking the structure of the data. Therefore, the present work proposes a comparative study of various deep learning algorithms from the Convolutional Neural Network (CNN) family and evaluate their performance on SITS classification. They are compared to the processing chain coined iota2, developed by the CESBIO and based on a RF model. Experiments are carried out in an operational context using with sparse annotations from 290 labeled polygons. Less than 80 000 pixel time series belonging to 8 land cover classes from a year of Sentinel- 2 monthly syntheses are used. Results show on a test set of 131 polygons that CNNs using 3D convolutions in space and time are more accurate than 1D temporal, stacked 2D and RF approaches. Best-performing models are CNNs using spatio-temporal features, namely 3D-CNN, 2D-CNN and SpatioTempCNN, a two-stream model using both 1D and 3D convolutions. / Tidsserier av satellitbilder (SITS) blir tillgängliga med hög rumslig, spektral och tidsmässig upplösning över hela världen med hjälp av de senaste fjärranalyssensorerna. Dessa bildserier kan vara mycket värdefulla när de utnyttjas av klassificeringssystem för att ta fram ofta uppdaterade och exakta kartor över marktäcken. Den stora mängden spektrala, rumsliga och tidsmässiga egenskaper i SITS är en lovande datakälla för utveckling av bättre algoritmer. Metoder för maskininlärning som Random Forests (RF), trots att de har tillämpats på SITS för att ta fram kartor över landtäckning, är strukturellt sett oförmögna att hantera den sammanflätade rumsliga, spektrala och temporala dynamiken utan att bryta sönder datastrukturen. I detta arbete föreslås därför en jämförande studie av olika algoritmer från Konvolutionellt Neuralt Nätverk (CNN) -familjen och en utvärdering av deras prestanda för SITS-klassificering. De jämförs med behandlingskedjan iota2, som utvecklats av CESBIO och bygger på en RF-modell. Försöken utförs i ett operativt sammanhang med glesa annotationer från 290 märkta polygoner. Mindre än 80 000 pixeltidsserier som tillhör 8 marktäckeklasser från ett års månatliga Sentinel-2-synteser används. Resultaten visar att CNNs som använder 3D-falsningar i tid och rum är mer exakta än 1D temporala, staplade 2D- och RF-metoder. Bäst presterande modeller är CNNs som använder spatiotemporala egenskaper, nämligen 3D-CNN, 2D-CNN och SpatioTempCNN, en modell med två flöden som använder både 1D- och 3D-falsningar.

New satellite missions have provided High Resolution (HR) Satellite Image Time Series (SITS), offering detailed spatial, spectral, and temporal information for effective monitoring of diverse Earth features including weather, landforms, oceans, vegetation, and agricultural practices. SITS can be used for an accurate understanding of the Land Cover (LC) behavior and providing the possibility of precise mapping of LCs. Moreover, HR SITS presents an unprecedented possibility for the creation and modification of HR Land Cover Change (LCC) and Land Cover Transition (LCT) maps. For the long-term scale, spanning multiple years, it becomes feasible to analyze LCC and the LCTs occurring between consecutive years. Existing methods in literature often analyze bi-temporal images and miss the valuable multi-temporal/multi-annual information of SITS that is crucial for an accurate SITS analysis. As a result, HR SITS necessitates a paradigm shift in processing and methodology development, introducing new challenges in data handling. Yet, the creation of techniques that can effectively manage the high spatial correlation and complementary temporal resolutions of pixels remains paramount. Moreover, the temporal availability of HR data across historical and current archives varies significantly, creating the need for an effective preprocessing to account for factors like atmospheric and radiometric conditions that can affect image reflectance and their applicability in SITS analysis. Flexible and automatic SITS analysis methods can be developed by paying special attention to handling big amounts of data and modeling the correlation and characterization of SITS in space and time. Novel methods should deal with data preparation and pre-processing at large-scale from end-to-end by introducing a set of steps that guarantee reliable SITS analysis while upholding the computational efficiency for a feasible SITS analysis. In this context, the recent strides in deep learning-based frameworks have demonstrated their potential across various image processing tasks, and thus the high relevance for addressing SITS analysis. Deep learning-based methods can be supervised or unsupervised considering their learning process. Supervised deep learning methods rely on labeled training data, which can be impractical for large-scale multi-temporal datasets, due to the challenges of manual labeling. In contrast, unsupervised deep learning methods are favored as they can automatically discover temporal patterns and changes without the need for labeled samples, thereby reducing the computational load, making them more suitable for handling extensive SITS. In this scenario, the objectives of this thesis are mainly three. Firstly, it seeks to establish a robust and reliable framework for the precise mapping of LCs by designing novel techniques for time series analysis. Secondly, it aims to utilize the capacities of unsupervised deep learning methods, such as pretrained Convolutional Neural Networks (CNNs), to construct a comprehensive methodology for Change Detection (CD), thereby mitigating complexity and reducing computational requirements in comparison with supervised methods. This involves the efficient extraction of spatial, spectral, and temporal features from complex multi-temporal, multi-spectral SITS. Lastly, the thesis endeavors to develop novel methods for analyzing LCCs occurring over extended time periods, spanning multiple years. This multifaceted approach encompasses the detection of changes, timing identification, and classification of the specific types of LCTs. The efficacy of the innovative methodologies and associated techniques is showcased through a series of experiments conducted on HR SITS datasets, including those from Sentinel-2 and Landsat. These experiments reveal significant enhancements when compared to existing methods that represent the current state-of-the-art.

Made available in DSpace on 2016-04-26T18:16:09Z (GMT). No. of bitstreams: 1 Iara H Magalhaes.pdf: 1225338 bytes, checksum: ea1a4ace24b813af850a8da1762a94c7 (MD5) Previous issue date: 2007-05-22 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / This paper aims to study the compositional technique of cinema, researching the technique of creation of the movie São Jerônimo, by Júlio Bressane, based in the design of the time. São Jerônimo was constructed from a series of documents, readings, paintings, thesis that traverse between fascination for iconography linked to the saint and the immanence of the film construction. A process of translation of the process of translating. The focus of this research is the time. The time that arises from the emancipation of the movement through light and sound time in the cinematographic image and drama. This paper was based theoretically in two works by Gilles Deleuze The image-movement and The image-time. The study of the concept image-time time emancipated from movement as the original regime of imagine and sign of the modern cinema, made it possible for us to see, hear and feel the time in São Jerônimo. We also studied the Peircian semiotics, the principles of the filmic analysis, the role of the sound in the cinema, the theoretical production of Brazilian researchers on critique of movies and unpublished works about photography in the creation of films.The method used to do the reading of the film involved three stages: in the first we studied the qualities, the shapes, the codes and their materialization in the composition of the audio-visual images, observing the framing, decoupage, the movement of the camera, the montage and materialization of the rhythmical ideas and their interactions; in the second stage we investigated the audiovisual images as aesthetical objects through the image-time concept; in the third stage we analyised the disjunction between the visual and audio images of the movie.The analysis of São Jerônimo made us to come to the following conclusion: its images immobilized the visual and sound aspects. The sound tries to take the film out of the visual field framed in the screen, in a centrifugal movement, but it detains itself in several deviations, tending to the purity of sound, while the visual aspect rapidly drags us to territories of excess or shortage of light the desert and Rome in a centripetal movement that leads, however, to disconnected and decentralized places, resisting to the photo grams and the images in movement. São Jerônimo, design of a moving material that cuts the cinematographic plan and rearrange the cuts that present the time. Jerônimo, Bressane and Deleuze s intercessor / Esse trabalho destina-se a estudar a técnica composicional no cinema, pesquisando a técnica de criação do filme São Jerônimo, de Júlio Bressane, segundo um desenho do tempo. São Jerônimo foi construído a partir de uma série de documentos, leituras, pinturas, teses que transitam entre o fascínio pela iconografia ligada ao santo e a imanência do filme sendo realizado. Um processo de tradução do processo de traduzir. O foco dessa pesquisa é o tempo. O tempo que nasce da emancipação do movimento pela luz e pelo som. O tempo na imagem e no drama cinematográficos. Esse estudo orientou-se pelas obras A imagem-movimento e A imagem-tempo, de Gilles Deleuze. O estudo do conceito imagem-tempo, tempo emancipado do movimento, como o regime original das imagens e signos da modernidade do cinema, possibilitou-nos pensar as técnicas de ver, ouvir e sentir o tempo em São Jerônimo. Ancorando a pesquisa, estudamos a semiótica Peirciana, os fundamentos da análise fílmica, o papel do som no cinema, a produção teórica de pesquisadores brasileiros sobre análise e crítica de filmes, e trabalhos inéditos sobre a fotografia na criação de filmes. A análise da técnica de luteria do filme compreendeu três etapas: na primeira estudamos as qualidades, as formas, os códigos e suas materializações nas composições das imagens audiovisuais, observando os enquadramentos, as decupagens, os movimentos de câmera, as montagens e as materializações de idéias rítmicas e suas interações; na segunda etapa investigamos as imagens audiovisuais como objetos estéticos através do conceito imagem-tempo; e na terceira etapa estudamos a disjunção entre as imagens visuais e sonoras do filme. A análise de São Jerônimo nos conduziu a uma constatação: suas imagens imobilizam o visual e o sonoro. O sonoro ensaia levar o filme para fora do campo visual enquadrado na tela, num movimento centrífugo, mas se detém em diversos desvios, tendendo à pureza do próprio som; e o visual se apressa a tragar-nos para territórios fundados pelo excesso ou escassez de luz, o deserto e Roma, num movimento centrípeto, todavia para lugares desconexos e sem centros, resistindo aos fotogramas e às imagens em movimento. São Jerônimo, desenho de uma matéria movente a cortar os planos e re-encadeando os cortes a nos apresentar o tempo. Jerônimo intercessor de Bressane e de Deleuze

Les Séries Temporelles d'Images Satellitaires (STIS), visant la même scène en évolution, sont très intéressantes parce qu'elles acquièrent conjointement des informations temporelles et spatiales. L'extraction de ces informations pour aider les experts dans l'interprétation des données satellitaires devient une nécessité impérieuse. Dans ce mémoire, nous exposons comment on peut adapter l'extraction de motifs séquentiels fréquents à ce contexte spatio-temporel dans le but d'identifier des ensembles de pixels connexes qui partagent la même évolution temporelle. La démarche originale est basée sur la conjonction de la contrainte de support avec différentes contraintes de connexité qui peuvent filtrer ou élaguer l'espace de recherche pour obtenir efficacement des motifs séquentiels fréquents groupés (MSFG) avec signification pour l'utilisateur. La méthode d'extraction proposée est non supervisée et basée sur le niveau pixel. Pour vérifier la généricité du concept de MSFG et la capacité de la méthode proposée d'offrir des résultats intéressants à partir des SITS, sont réalisées des expérimentations sur des données réelles optiques et radar. / The Satellite Image Time Series (SITS), aiming the same scene in evolution, are of high interest as they capture both spatial and temporal information. The extraction of this infor- mation to help the experts interpreting the satellite data becomes a stringent necessity. In this work, we expound how to adapt frequent sequential patterns extraction to this spatiotemporal context in order to identify sets of connected pixels sharing a same temporal evolution. The original approach is based on the conjunction of support constraint with different constraints based on pixel connectivity that can filter or prune the search space in order to efficiently ob- tain Grouped Frequent Sequential (GFS) patterns that are meaningful to the end user. The proposed extraction method is unsupervised and pixel level based. To verify the generality of GFS-pattern concept and the proposed method capability to offer interesting results from SITS, real data experiments on optical and radar data are presented.

La réussite du lancement d'un grand nombre des satellites Radar à Synthèse d'Ouverture (RSO - SAR) de nouvelle génération a fourni régulièrement des images SAR et SAR polarimétrique (PolSAR) multitemporelles à haute et très haute résolution spatiale sur de larges régions de la surface de la Terre. Le système SAR est approprié pour des tâches de surveillance continue ou il offre l'avantage d'être indépendant de l'éclairement solaire et de la couverture nuageuse. Avec des données multitemporelles, l'information spatiale et temporelle peut être exploitée simultanément pour rendre plus concise, l'extraction d'information à partir des données. La détection de changement de structures spécifiques dans un certain intervalle de temps nécessite un traitement complexe des données SAR et la présence du chatoiement (speckle) qui affecte la rétrodiffusion comme un bruit multiplicatif. Le but de cette thèse est de fournir une méthodologie pour simplifier l'analyse des données multitemporelles SAR. Cette méthodologie doit bénéficier des avantages d'acquisitions SAR répétitives et être capable de traiter différents types de données SAR (images SAR mono-, multi- composantes, etc.) pour diverses applications. Au cours de cette thèse, nous proposons tout d'abord une méthode générale basée sur une matrice d'information spatio-temporelle appelée Matrice de détection de changement (CDM). Cette matrice contient des informations de changements obtenus à partir de tests croisés de similarité sur des voisinages adaptatifs. La méthode proposée est ensuite exploitée pour réaliser trois tâches différentes: 1) la détection de changement multitemporel avec différents types de changements, ce qui permet la combinaison des cartes de changement entre des paires d'images pour améliorer la performance de résultat de détection de changement; 2) l'analyse de la dynamicité de changement de la zone observée, ce qui permet l'étude de l'évolution temporelle des objets d'intérêt; 3) le filtrage nonlocal temporel des séries temporelles d'images SAR/PolSAR, ce qui permet d'éviter le lissage des informations de changement dans des séries pendant le processus de filtrage.Afin d'illustrer la pertinence de la méthode proposée, la partie expérimentale de la thèse est effectuée sur deux sites d'étude: Chamonix Mont-Blanc, France et le volcan Merapi, Indonésie, avec différents types de changements (i.e. évolution saisonnière, glaciers, éruption volcanique, etc.). Les observations de ces sites d'étude sont acquises sur quatre séries temporelles d'images SAR monocomposantes et multicomposantes de moyenne à haute et très haute résolution: des séries temporelles d'images Sentinel-1, ALOS-PALSAR, RADARSAT-2 et TerraSAR-X. / A large number of successfully launched and operated Synthetic Aperture Radar (SAR) satellites has regularly provided multitemporal SAR and polarimetric SAR (PolSAR) images with high and very high spatial resolution over immense areas of the Earth surface. SAR system is appropriate for monitoring tasks thanks to the advantage of operating in all-time and all-weather conditions. With multitemporal data, both spatial and temporal information can simultaneously be exploited to improve the results of researche works. Change detection of specific features within a certain time interval has to deal with a complex processing of SAR data and the so-called speckle which affects the backscattered signal as multiplicative noise.The aim of this thesis is to provide a methodology for simplifying the analysis of multitemporal SAR data. Such methodology can benefit from the advantages of repetitive SAR acquisitions and be able to process different kinds of SAR data (i.e. single, multipolarization SAR, etc.) for various applications. In this thesis, we first propose a general framework based on a spatio-temporal information matrix called emph{Change Detection Matrix} (CDM). This matrix contains temporal neighborhoods which are adaptive to changed and unchanged areas thanks to similarity cross tests. Then, the proposed method is used to perform three different tasks:1) multitemporal change detection with different kinds of changes, which allows the combination of multitemporal pair-wise change maps to improve the performance of change detection result;2) analysis of change dynamics in the observed area, which allows the investigation of temporal evolution of objects of interest;3) nonlocal temporal mean filtering of SAR/PolSAR image time series, which allows us to avoid smoothing change information in the time series during the filtering process.In order to illustrate the relevancy of the proposed method, the experimental works of the thesis is performed on four datasets over two test-sites: Chamonix Mont-Blanc, France and Merapi volcano, Indonesia, with different types of changes (i.e., seasonal evolution, glaciers, volcanic eruption, etc.). Observations of these test-sites are performed on four SAR images time series from single polarization to full polarization, from medium to high, very high spatial resolution: Sentinel-1, ALOS-PALSAR, RADARSAT-2 and TerraSAR-X time series.

Les moyens mis en œuvre pour surveiller la surface de la Terre, notamment les zones urbaines, en cas de catastrophes naturelles telles que les inondations ou les tremblements de terre, et pour évaluer l’impact de ces événements, jouent un rôle primordial du point de vue sociétal, économique et humain. Dans ce cadre, des méthodes de classification précises et efficaces sont des outils particulièrement importants pour aider à l’évaluation rapide et fiable des changements au sol et des dommages provoqués. Étant données l’énorme quantité et la variété des données Haute Résolution (HR) disponibles grâce aux missions satellitaires de dernière génération et de différents types, telles que Pléiades, COSMO-SkyMed ou RadarSat-2 la principale difficulté est de trouver un classifieur qui puisse prendre en compte des données multi-bande, multi-résolution, multi-date et éventuellement multi-capteur tout en gardant un temps de calcul acceptable. Les approches de classification multi-date/multi-capteur et multi-résolution sont fondées sur une modélisation statistique explicite. En fait, le modèle développé consiste en un classifieur bayésien supervisé qui combine un modèle statistique conditionnel par classe intégrant des informations pixel par pixel à la même résolution et un champ de Markov hiérarchique fusionnant l’information spatio-temporelle et multi-résolution, en se basant sur le critère des Modes Marginales a Posteriori (MPM en anglais), qui vise à affecter à chaque pixel l’étiquette optimale en maximisant récursivement la probabilité marginale a posteriori, étant donné l’ensemble des observations multi-temporelles ou multi-capteur / The capabilities to monitor the Earth's surface, notably in urban and built-up areas, for example in the framework of the protection from environmental disasters such as floods or earthquakes, play important roles in multiple social, economic, and human viewpoints. In this framework, accurate and time-efficient classification methods are important tools required to support the rapid and reliable assessment of ground changes and damages induced by a disaster, in particular when an extensive area has been affected. Given the substantial amount and variety of data available currently from last generation very-high resolution (VHR) satellite missions such as Pléiades, COSMO-SkyMed, or RadarSat-2, the main methodological difficulty is to develop classifiers that are powerful and flexible enough to utilize the benefits of multiband, multiresolution, multi-date, and possibly multi-sensor input imagery. With the proposed approaches, multi-date/multi-sensor and multi-resolution fusion are based on explicit statistical modeling. The method combines a joint statistical model of multi-sensor and multi-temporal images through hierarchical Markov random field (MRF) modeling, leading to statistical supervised classification approaches. We have developed novel hierarchical Markov random field models, based on the marginal posterior modes (MPM) criterion, that support information extraction from multi-temporal and/or multi-sensor information and allow the joint supervised classification of multiple images taken over the same area at different times, from different sensors, and/or at different spatial resolutions. The developed methods have been experimentally validated with complex optical multispectral (Pléiades), X-band SAR (COSMO-Skymed), and C-band SAR (RadarSat-2) imagery taken from the Haiti site