Orientador: Aluir Porfírio Dal Poz / Banca: Antônio Maria Garcia Tommaselli / Banca: Francisco Henrique de Oliveira / Resumo: Este trabalho apresenta uma metodologia para a predição de sombras projetadas por edifícios em vias urbanas usando imagens aéreas de alta-resolução e modelos digitais de elevações (MDE). Elementos de sombra podem ser usados na modelagem de informação contextual para, por exemplo, a extração semi-automática ou automática da malha viária a partir de imagens aéreas ou de satélite, principalmente em áreas urbanas densas. O contexto modela as relações entre os objetos numa imagem, como por exemplo, as relações entre edifícios e vias adjacentes. O uso de conhecimento contextual tem se tornado cada vez mais comum em processos de análise de imagem, principalmente em se tratando de cenas complexas. Este trabalho foi inspirado nas atuais possibilidades de se obter modelos digitais de elevações densos e acurados de áreas urbanas complexas a partir de dados de sistemas aerotransportados de varredura a laser. A metodologia proposta consiste de três etapas seqüenciais. Primeiramente, os contornos de telhados de edifícios são extraídos manualmente a partir de uma imagem altimétrica obtida pela transformação do MDE/laser. De maneira similar, os contornos dos limites das vias são extraídos, agora a partir da imagem de intensidade de retorno do pulso laser. Na etapa seguinte, os polígonos dos contornos de telhado são projetados nas vias adjacentes através do uso da projeção paralela. A direção das retas de projeção paralela é calculada a partir de dados de efemérides solares e do instante de tomada da imagem aérea. Finalmente, as partes dos polígonos de sombra que estão livres de obstruções perspectivas de edifícios são determinadas a partir da separação das regiões afetadas pela perspectiva de tomada. / Abstract: This research presents a methodology for prediction of building shadows cast on urban roads using high-resolution aerial images and Digital Elevation Models (DEM). Shadow elements can be used in the modeling of contextual information, as e. g., in the semiautomatic or automatic road network extraction from satellite or aerial images, mainly in dense urban areas. Context models the relations among objects in an image, as e. g., the relations among buildings and adjacent roads. The use of contextual knowledge has become more common in image analysis processes, mainly if complex scenes are present. This research drew inspiration from the present possibilities of obtaining accurate and dense DEM from airborne laser scanning data of complex urban areas. The proposed methodology consists in three sequential steps. First, the building roof contours are manually extracted from an intensity image generated from the laser DEM. In similar way, the road side contours are extracted, now from the radiometric information of the laser scanning data. Second, the roof contour polygons are projected onto the adjacent roads by using the parallel projection. The direction of the parallel straight lines is computed from the solar ephemeris, which depends on the taken time of the aerial photograph. Finally, the parts of shadow polygons that are freed of building perspective obstructions are determined, given rise to new shadow polygons. / Mestre

Resumo: Este trabalho propõe uma metodologia monoscópica para a determinação automática da altura de edifícios em fotografias aéreas digitais, baseando-se no deslocamento radial dos pontos projetados no plano imagem e na geometria formada no momento de obtenção da fotografia aérea. O conhecimento da altura de edifícios pode ser utilizado para a modelagem da superfície em áreas urbanas, estudos para a instalação de antenas de transmissão de sinais, planejamento de rotas para aeronaves, geração de “true-orthophoto” etc.. A metodologia emprega um conjunto de etapas para a detecção das bordas dispostas radialmente em relação ao sistema de coordenadas fotogramétrico, sendo que estas bordas representam as arestas laterais dos edifícios registradas na fotografia. Na primeira etapa realiza-se a redução do espaço de busca através da detecção das sombras projetadas pelos edifícios, gerando sub-imagens das áreas no entorno de cada sombra detectada. Na seqüência, é realizada a extração de bordas e conexão dos pixels destas sub-imagens. As bordas são analisadas individualmente para identificar qual par, dentre os pontos que compõem a borda, melhor corresponde aos extremos da aresta lateral do edifício. Os critérios utilizados para a seleção dos pares de pontos são os seguintes: análise do coeficiente linear da equação da reta formada pelo par de pontos; distância entre os pontos; e grau de sobreposição com a borda rotulada. Deste modo, a partir da seleção das arestas laterais, assim como o conhecimento da altura de vôo, são calculadas as alturas destes edifícios. Os resultados alcançados nos testes realizados com imagens reais se mostraram promissores para a identificação das arestas e cálculo da altura de edifícios. / Abstract: This work proposes an automatic extraction of buildings height in digital images through a monoscopic methodology. Surface modeling of urban areas planning of aircraft routes, true-orthophoto and so on, can be performed by knowing the buildings heights. In photographs taken from an aircraft the buildings edges appear as radial lines due to the relief displacement, which occurs radially from the principal point. The first step of the methodology consists on the sub-images creation with the aim to reduce the search space on image. Thus, its used a shadows extraction procedure, by selecting only those with the biggest area (projected at ground) according to threshold. The edge extraction and pixels connection are performed for each sub-image produced and analyzed individually in order to identify if they belong to a building. The adopted criteria are based on the linear coefficient of straight line, the distance and the overlapping degree with regard to the labeled edge. With the lateral edges selected and the knowledge of the flight height, the buildings heights can be calculated. The experimental results obtained with real images show that the proposed approach is suitable to perform the automatic identification of the buildings height in digital images. / Orientador: Júlio K. Hasegawa / Coorientador: Maurício Galo / Banca: Sergio Donizete Farias / Banca: Antônio Maria Garcia Tommaselli / Mestre

Les caméras de surveillance sont de plus en plus fréquemment présentes dans notre environnement (villes, supermarchés, aéroports, entrepôts, etc.). Ces caméras sont utilisées, entre autres, afin de pouvoir détecter des comportements suspects (intrusion par exemple) ou de reconnaître une catégorie d'objets ou de personnes (détection de genre, détection de plaques d'immatriculation par exemple). D'autres applications concernent également l'établissement de statistiques de fréquentation ou de passage (comptage d'entrée/sortie de personnes ou de véhicules) ou bien le suivi d'un ou plusieurs objets se déplaçant dans le champ de vision de la caméra (trajectoires d'objets, analyse du comportement des clients dans un magasin). Compte tenu du nombre croissant de caméras et de la difficulté à réaliser ces traitements manuellement, un ensemble de méthodes d'analyse vidéo ont été développées ces dernières années afin de pouvoir automatiser ces tâches. Dans cette thèse, nous nous concentrons essentiellement sur les tâches de détection et de suivi des objets mobiles à partir d'une caméra fixe. Contrairement aux méthodes basées uniquement sur les images acquises par les caméras, notre approche consiste à intégrer un certain nombre d'informations contextuelles à l'observation afin de pouvoir mieux interpréter ces images. Ainsi, nous proposons de construire un modèle géométrique et géolocalisé de la scène et de la caméra. Ce modèle est construit directement à partir des études de prédéploiement des caméras et peut notamment utiliser les données OpenStreetMap afin d'établir les modèles 3d des bâtiments proches de la caméra. Nous avons complété ce modèle en intégrant la possibilité de prédire la position du Soleil tout au long de la journée et ainsi pouvoir calculer les ombres projetées des objets de la scène. Cette prédiction des ombres a été mise à profit afin d'améliorer la segmentation des piétons par modèle de fond en supprimant les ombres du masque de mouvement. Concernant le suivi des objets mobiles, nous utilisons le formalisme des automates finis afin de modéliser efficacement les états et évolutions possibles d'un objet. Ceci nous permet d'adapter le traitement de chaque objet selon son état. Nous gérons les occultations inter-objets à l'aide d'un mécanisme de suivi collectif (suivi en groupe) des objets le temps de l'occultation et de ré-identification de ceux-ci à la fin de l'occultation. Notre algorithme s'adapte à n'importe quel type d'objet se déplaçant au sol (piétons, véhicules, etc.) et s'intègre naturellement au modèle de scène développé. Nous avons également développé un ensemble de "rétro-actions" tirant parti de la connaissance des objets suivis afin d'améliorer les détections obtenues à partir d'un modèle de fond. En particulier, nous avons abordé le cas des objets stationnaires, souvent intégrés à tort dans le fond, et avons revisité la méthode de suppression des ombres du masque de mouvement en tirant parti de la connaissance des objets suivis. L'ensemble des solutions proposées a été implémenté dans le logiciel de l'entreprise Foxstream et est compatible avec la contrainte d'exécution en temps réel nécessaire en vidéosurveillance. / Video-surveillance cameras are increasingly used in our environment. They are indeed present almost everywhere in the cities, supermarkets, airports, warehouses, etc. These cameras are used, among other things, in order to detect suspect behavior (an intrusion for instance) or to recognize a specific category of object or person (gender detection, license plates detection). Other applications also exist to count and/or track people in order to analyze their behavior. Due to the increasing number of cameras and the difficulty to achieve these tasks manually, several video analysis methods have been developed in order to address them automatically. In this thesis, we mainly focus on the detection and tracking of moving objects from a fixed camera. Unlike methods based solely on images captured by cameras, our approach integrates contextual pieces of information in order better interpret these images. Thus we propose to build a geometric and geolocalized model of the scene and the camera. This model is built directly from the pre-deployment studies of the cameras and uses the OpenStreetMap geographical database to build 3d models of buildings near the camera. We added to this model the ability to predict the position of the sun throughout the day and the resulting shadows in the scene. By predicting the shadows, and deleting them from the foreground mask, our method is able to improve the segmentation of pedestrians. Regarding the tracking of multiple mobile objects, we use the formalism of finite state machines to effectively model the states and possible transitions that an object is allowed to take. This allows us to tailor the processing of each object according to its state. We manage the inter-object occlusion using a collective tracking strategy. When taking part in an occlusion, objects are regrouped and tracked collectively. At the end of the occlusion, each object is re-identified and individual tracking resume. Our algorithm adapts to any type of ground-moving object (pedestrians, vehicles, etc.) and seamlessly integrates in the developed scene model. We have also developed several retro-actions taking advantage of the knowledge of tracked objects to improve the detections obtained with the background model. In particular, we tackle the issue of stationary objects often integrated erroneously in the background and we revisited the initial proposal regarding shadow removal. All proposed solutions have been implemented in the Foxstream products and are able to run in real-time.

Orientador: Luise Weiss / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Artes / Made available in DSpace on 2018-08-24T21:31:09Z (GMT). No. of bitstreams: 1 Silva_WilsonRobertoda_D.pdf: 16280272 bytes, checksum: 11413fbb861442a3e75dea59047ef186 (MD5) Previous issue date: 2014 / Resumo: Esta tese é resultado de uma pesquisa iniciada no mestrado e consiste na produção de estampas predominantemente provenientes de matrizes de gravura em metal e xilogravura, gravadas para serem capazes de permutar suas formas entre elas separadamente e formar combinações complementares, híbridas e aleatórias resultantes do processo de impressão. O principal objetivo é obter o maior número possível de estampas diferentes de um conjunto de matrizes e com isso, não só conciliar visualmente erotismo e anatomia humana na gravura, mas também ressaltar a variedade, como atributo dos organismos que se reproduzem sexualmente e perecem, para dar início a um novo ciclo / Abstract: This thesis is the result of a research started in the master's program and consists of the production of prints mainly from etching plates and woodcut, carved to be able to permute its shapes among them separately and create complimentary, hybrid and random combinations resulting from the printing process. The main goal is to obtain the largest possible amount of different prints from a set of plates and, with that, not only visually conciliate eroticism and human anatomy on the engraving, but also highlight the variety, as an attribute of the organisms that reproduce sexually and perish, in order to initiate a new cycle / Doutorado / Artes Visuais / Doutor em Artes

Orientador: Luise Weiss / Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Artes / Made available in DSpace on 2018-08-21T19:37:32Z (GMT). No. of bitstreams: 1 Cappellari_Angelo_M.pdf: 42290124 bytes, checksum: e52a0c51343e47c93b39760c3a3ba8f3 (MD5) Previous issue date: 2012 / Resumo: Esta dissertação percorre um duplo caminho: de um lado, alguns objetos do cotidiano sugerindo montagens, provocando novos olhares sobre estes objetos; de outro lado, algumas considerações sobre o Homem e seus objetos na Arte. Ambas as áreas dialogam com o objeto cotidiano na sua capacidade de transformação / Abstract: This work follows a double path: on one side, some everyday objects suggesting assemblies, provoking new looks on these objects; on the other hand, some considerations about Man and his objects in Art. Both areas dialogue as an everyday object in its changing capacity / Mestrado / Artes Visuais / Mestre em Artes

Körsimulatorer är idag en mycket viktig resurs för att utföra studier med fokus på förarbeteende. Så väl full kontroll överscenario och miljö som kostnad och säkerhet är aspekter som gör det fördelaktigt att utföra simulatorstudier gentemotstudier i den riktiga trafiken. Ett problem med körsimulatorer är att bilden projiceras på en tvådimensionell skärm, vilket begränsar förarens förmåga attuppskatta avstånd och hastighet. Det är allmänt känt att avstånd och hastighet underskattas i körsimulatorer. Målet med examensarbetet var att hitta metoder som kan ge förbättrad avståndsbedömning i körsimulatorer och underprojektet implementerades och testades rörelseparallax samt skuggor, med största fokus på det förstnämnda. I slutet av projektet genomfördes ett simulatorförsök för att utvärdera effekten av rörelseparallax. Tio försökspersoner fickgöra två körningar vardera i VTI:s Simulator III-anläggning, den ena med rörelseparallax aktiverat och den andra meddetsamma inaktiverat. Scenariot som utspelade sig under körningarna innehöll ett flertal omkörningssituationer samt etthastighetsuppfattningstest. Resultaten från simulatorförsöket visade att försökspersonerna tenderade att placera sig längre från mittlinjen närrörelseparallax var aktiverat i de situationer som sikten skymdes av framförvarande fordon. / Today, driving simulators are a very important resource for conducting studies whichconcern driver behavior and perception. Full control of the scenario andenvironment, costs and safety are all factors which makes simulator studiespreferable over real world studies. One issue for driving simulators is that the image is projected onto a twodimensionalscreen, which limits the driver's ability to correctly estimate distanceand speed. It is commonly known that distance and velocity are underestimated indriving simulators. The goal of this thesis was to find methods that could lead to better distanceestimation in driving simulators and in this project motion parallax and shadowswere implemented and tested, focusing mainly on the former.At the end of the project, a simulator study was conducted to evaluate the effect ofmotion parallax. Ten participants made two runs each in VTI's Simulator III facility,one with motion parallax enabled and one with it disabled. The scenario that tookplace during the two runs consisted of several overtaking situations and a speedperception test. The results from the simulator study showed that the participants tended to positionthemselves farther from the road center line when motion parallax was active insituations when the field of view was obscured by preceding vehicles.

This master's project focuses on basic techniques of creating shadows in 3D computer graphics. Two basic techniques are compared. Those are shadow maps and shadow volumes. Another technique combining previouse two is proposed.

This work concerns itself with choosen methods of shadows in computer graphics. In short it discuses both shadow maps and shadow volumes, with their advantages and disadvantages. Then it employs itself on shadow volumes and their implementation utilizing Open Scene Graph toolkit. It deals with z-pass and z-fail methods and some possible optimizations. Moreover, it contains measuring of FPS and comparison of fill consumption of the scene and shadow volumes on various scenes.

Cette thèse aborde la problématique du calcul d’ombre exact par pixel en temps réel. Ce mémoire propose une nouvelle méthode de rendu d’ombre dure avec une partition de volumes d’ombre : un arbre ternaire basés sur les plans d’ombre des volumes d’ombre de la scène est construit dans un premier temps, avant de l’interroger pour déterminer l’ombrage des pixels de l’image. Une des propriétés intéressantes de cette structure est la prédictibilité de son empreinte mémoire ; contrairement aux méthodes de calcul d’ombre, cette méthode supporte des scènes géométriquement complexes de l’ordre du million de triangles grâce au comportement logarithmique du parcours de la structure vis à vis de la complexité géométrique. / This thesis focuses on exact per pixel hard shadow computation. We propose a new method for rendering hard shadows using a partition of shadow volumes : first, a ternary tree based on the shadow planes of the shadow volumes, and then, it is traversed to determine the shading of each pixel in the image. An interesting property of this structure is the predictability of its memory footprint ; Unlike other geometric shadow methods, our approach supports complex scenes up to a million triangles thanks to the logarithmic behavior of the structure traversal with respect to the geometric complexity.

Les ombres sont un élément fondamental pour garantir le réalisme des images de synthèses. Elles donnent des informations nécessaires pour discerner correctement le positionnement relatif des objets. Le calcul d'ombre dures exactes en temps-réel reste toutefois un problème difficile en Informatique Graphique. Les Shadow Maps sont privilégiées pour leurs excellentes performances, mais elles ne peuvent garantir un résultat exact par pixel. Les volumes d'ombre sont plus lents que les Shadow Maps mais produisent des ombres exactes en toute circonstances. Bien qu'ils aient récemment atteint un haut niveau de performances, ils restent sensibles à la complexité géométrique et sont limités à des scènes de taille modeste. Dans cette thèse, nous proposons un nouvel algorithme de rendu d'ombres dures en temps réel basé sur les volumes d'ombre qui reste rapide sur des scènes de très grande tailles. De plus, nous introduisons une nouvelle stratégie de partitionnement jusque-là inexplorée en Informatique Graphique : les arbres métriques. / Shadows are a fundamental visual effect for computer generated images. They provide essential spatial hints allowing us to correctly perceive objects positions in the scene. Despite its importance, computing pixel accurate shadows in real-time is a challenging problem in Computer Graphics. Shadow Maps are still the favored solution due to their high efficiency, but they can't always provide pixel accurate results. Shadow Volumes are slower than Shadow Maps but always garantee pixel accurate shadows. While recent works significantly improved their performance, they remain sensitive to scalability, which limits their application scope. In this thesis, we propose a new algorithm based on Shadow Volumes that remains fast even on very large scenes. Furthermore, we introduce a new partitioning strategy that is yet unexplored in Computer Graphics : the metric tree.