Many contemporary types of 3D design software use some form of environment mapping, where the environment surrounding the 3D object is rendered to create a panoramic view similar to what we would see in the real world. This not only serves to make the scene look more realistic, but also helps to calculate light effects within the scene. Although environment mapping has been around for a while, and today's methods are highly optimized, they are still not trivially cheap to do, especially on lower end hardware. Our thesis investigates how an environment mapping method currently in use by our client can be further optimized. This optimization intends to serve two main purposes - to improve performance for clients with lower quality hardware and to allow users to increase texture resolution without extra cost on performance. Two alternative methods are presented, both based somewhat loosely on the idea of occlusion culling. The two approaches are then tested and compared to the original solution in terms of speed, memory utilization and network performance. Although both approaches show promise and outperform the original solution in some of the tests, they still lack the versatility of the original solution and suffer from some major flaws, making them less appealing alternatives for the general customer. The first approach managed to perform well in all three areas of measurement, but suffers a drawback which limits its use in a real-world scenario. The second approach did not have the same drawback which may make it a more viable option. However, the results of the second approach were not as positive as the first one. With that said, it showed some promise for users who do their rendering on a separate server. While this solution may not yet be viable for the general user, it may serve well for users with more unique needs. To make this approach a viable solution for the general user improvements in regards to rendering speed and GPU utilization will have to be investigated further.

Spatial mapping or environment mapping is the process of exploring a real world environment and creating its digital representation. To create convincing mixed reality programs, an environment mapping device must be able to detect a user's position and map the user's environment. Currently available commercial spatial mapping devices mostly use infrared camera to obtain a depth map which is effective only for short to medium distances (3-4 meters). This work describes an extension to the existing environment mapping devices and techniques to enable mapping of larger architectural environments using a combination of a camera, Inertial Measurement Unit (IMU), and Light Detection and Ranging (LIDAR) devices supported by sensor fusion and computer vision techniques. There are three main parts to the proposed system. The first part is data collection and data fusion using embedded hardware, the second part is data processing (segmentation) and the third part is creating a geometry mesh of the environment. The developed system was evaluated against its ability to determine the dimension of the room and of objects within the room. This low cost system can significantly expand the mapping range of the existing mixed reality devices such as Microsoft HoloLens device. / Master of Science / Mixed reality is the mixing of computer generated graphics and real world objects together to create an augmented view of the space. Environmental mapping, the process of creating a digital representation of an environment, is used in mixed reality applications so that its virtual objects can logically interact with the physical environment. Most of the current approaches to this problem work only for short to medium distances. This work describes an extension to the existing devices and techniques to enable mapping of larger architectural spaces. The developed system was evaluated against its ability to determine the dimension of the room and of objects within the room. With certain conditions the system was able to evaluate the dimensions of a room with an error less than twenty percent and is capable of determining the dimensions of objects with an error less than five percent in adequate conditions. This low cost system can significantly expand the mapping range of the existing mixed reality devices such as the Microsoft HoloLens device, allowing for more diverse mixed reality applications to be developed and used.

In the journey toward optimal spectrum usage, techniques and concepts such as Cognitive Radio and Dynamic Spectrum Access have enjoyed increasing attention in many research projects. Dynamic Spectrum Access introduces the need for real-time RF spectrum information in the form of Radio Environment Maps. This need motivates an investigation into a hybrid approach of sample measurements and spatial interpolation as opposed to using conventional propagation models. Conventional propagation models, both path-general and path-specific, require information of transmitters within the area of interest. Irregular Terrain Models such as the Longley-Rice model, further require topographic information in order to consider the effects of obstacles. The proposed spatial interpolation technique, kriging, requires no information regarding transmitters. Furthermore, Ordinary Kriging requires nothing other than measured samples whereas other kriging variants such as Universal Kriging and Regression Kriging can use additional information such as topographic data to aid in prediction accuracy. This dissertation investigates the performance of the three aforementioned kriging variants in producing Radio Environment Maps of received power. For practical and financial reasons, the received power measurement samples are generated using the Longley-Rice Irregular Terrain Model and are, therefore, simulated measurements. The experimental results indicate that kriging shows great promise as a tool to generate Radio Environment Maps. It is found that Ordinary Kriging produces the most accurate predictions of the three kriging methods and that prediction errors of less than 10 dB can be achieved even when using very low sampling densities. / MSc (Computer and Electronic Engineering), North-West University, Potchefstroom Campus, 2015

In the journey toward optimal spectrum usage, techniques and concepts such as Cognitive Radio and Dynamic Spectrum Access have enjoyed increasing attention in many research projects. Dynamic Spectrum Access introduces the need for real-time RF spectrum information in the form of Radio Environment Maps. This need motivates an investigation into a hybrid approach of sample measurements and spatial interpolation as opposed to using conventional propagation models. Conventional propagation models, both path-general and path-specific, require information of transmitters within the area of interest. Irregular Terrain Models such as the Longley-Rice model, further require topographic information in order to consider the effects of obstacles. The proposed spatial interpolation technique, kriging, requires no information regarding transmitters. Furthermore, Ordinary Kriging requires nothing other than measured samples whereas other kriging variants such as Universal Kriging and Regression Kriging can use additional information such as topographic data to aid in prediction accuracy. This dissertation investigates the performance of the three aforementioned kriging variants in producing Radio Environment Maps of received power. For practical and financial reasons, the received power measurement samples are generated using the Longley-Rice Irregular Terrain Model and are, therefore, simulated measurements. The experimental results indicate that kriging shows great promise as a tool to generate Radio Environment Maps. It is found that Ordinary Kriging produces the most accurate predictions of the three kriging methods and that prediction errors of less than 10 dB can be achieved even when using very low sampling densities. / MSc (Computer and Electronic Engineering), North-West University, Potchefstroom Campus, 2015

Οι δορυφορικές εικόνες ASTER αποτελούν το πιο πρόσφατο προϊόν της διαστημικής τεχνολογίας, Έχει πολύ μεγάλη φασματική διακριτική ικανότητα σε σημείο που να χαρακτηρίζεται από αρκετούς επιστήμονες σαν ένα υπερφασματικό καταγραφικό σύστημα στο υπέρυθρο. Έτσι δίνονται νέες δυνατότητες για την ερμηνεία είτε ποιοτικά, είτε ποσοτικά καλύψεων γης και φυσικών καταστάσεων της γήινης επιφάνειας, αφού οι προηγούμενοι δορυφόροι. Ο σκοπός είναι η περιβαλλοντική χαρτογράφηση του N. Ηλείας από δορυφορικές εικόνες ASTER προκειμένου να χαρτογραφηθούν οι φυσικοί πόροι και οι καλύψεις Γης. Έτσι θα μπορούμε να αξιολογήσουμε τις δυνατότητες και τις εφαρμογές που έχουν αυτά τα δεδομένα. Μετά τις διορθώσεις (αποζωνοποίηση, μετατροπή σε τιμές ενέργειας, κ.α.) η ταξινόμηση της δορυφορικής εικόνας μας επέτρεψε να διακρίνουμε τις παρακάτω κατηγορίες: καλλιέργειες/ποώδη βλάστηση, δύο τύπους δάσους, αστική γη, υδάτινες επιφάνειες, σύννεφα και εντελώς γυμνό από βλάστηση έδαφος (πχ κοίτη ποταμού που καταλήγει στην τεχνητή λίμνη του Πηνειού) και να δημιουργήσουμε ένα θεματικό χάρτη. Εάν χρησιμοποιηθούν περισσότερα κανάλια και όχι μόνο 3, τότε θα γίνει δυνατή η χαρτογράφηση περισσότερων κατηγοριών μόνο που η ταυτοποίηση της κάθε κατηγορίας θα μπορεί να πραγματοποιηθεί μόνο με εργασίες πεδίου στο ύπαιθρο. / -

Scene perception and traversability analysis are real challenges for autonomous driving systems. In the context of off-road autonomy, there are additional challenges due to the unstructured environments and the existence of various vegetation types. It is necessary for the Autonomous Ground Vehicles (AGVs) to be able to identify obstacles and load-bearing surfaces in the terrain to ensure a safe navigation (McDaniel et al. 2012). The presence of vegetation in off-road autonomy applications presents unique challenges for scene understanding: 1) understory vegetation makes it difficult to detect obstacles or to identify load-bearing surfaces; and 2) trees are usually regarded as obstacles even though only trunks of the trees pose collision risk in navigation. The overarching goal of this dissertation was to study traversability analysis in unstructured forested terrains for off-road autonomy using LIDAR data. More specifically, to address the aforementioned challenges, this dissertation studied the impacts of the understory vegetation density on the solid obstacle detection performance of the off-road autonomous systems. By leveraging a physics-based autonomous driving simulator, a classification-based machine learning framework was proposed for obstacle detection based on point cloud data captured by LIDAR. Features were extracted based on a cumulative approach meaning that information related to each feature was updated at each timeframe when new data was collected by LIDAR. It was concluded that the increase in the density of understory vegetation adversely affected the classification performance in correctly detecting solid obstacles. Additionally, a regression-based framework was proposed for estimating the understory vegetation density for safe path planning purposes according to which the traversabilty risk level was regarded as a function of estimated density. Thus, the denser the predicted density of an area, the higher the risk of collision if the AGV traversed through that area. Finally, for the trees in the terrain, the dissertation investigated statistical features that can be used in machine learning algorithms to differentiate trees from solid obstacles in the context of forested off-road scenes. Using the proposed extracted features, the classification algorithm was able to generate high precision results for differentiating trees from solid obstacles. Such differentiation can result in more optimized path planning in off-road applications.

A maior parte dos mapas empregados em tarefas de navegação por robôs móveis representam apenas informações espaciais do ambiente. Outros tipos de informações, que poderiam ser obtidos dos sensores do robô e incorporados à representação, são desprezados. Hoje em dia é comum um robô móvel conter sensores de distância e um sistema de visão, o que permitiria a princípio usá-lo na realização de tarefas complexas e gerais de maneira autônoma, dada uma representação adequada e um meio de extrair diretamente dos sensores o conhecimento necessário. Uma representação possível nesse contexto consiste no acréscimo de informação semântica aos mapas métricos, como por exemplo a segmentação do ambiente seguida da rotulação de cada uma de suas partes. O presente trabalho propõe uma maneira de estruturar a informação espacial criando um mapa semântico do ambiente que representa, além de obstáculos, um vínculo entre estes e as imagens segmentadas correspondentes obtidas por um sistema de visão omnidirecional. A representação é implementada por uma descrição relacional do domínio, que quando instanciada gera um campo aleatório condicionado, onde são realizadas as inferências. Modelos que combinam probabilidade e lógica de primeira ordem são mais expressivos e adequados para estruturar informações espaciais em semânticas. / Most maps used in navigational tasks by mobile robots represent only environmental spatial information. Other kinds of information, that might be obtained from the sensors of the robot and incorporated in the representation, are negleted. Nowadays it is common for mobile robots to have distance sensors and a vision system, which could in principle be used to accomplish complex and general tasks in an autonomously manner, given an adequate representation and a way to extract directly from the sensors the necessary knowledge. A possible representation in this context consists of the addition of semantic information to metric maps, as for example the environment segmentation followed by an attribution of labels to them. This work proposes a way to structure the spatial information in order to create a semantic map representing, beyond obstacles, an anchoring between them and the correspondent segmented images obtained by an omnidirectional vision system. The representation is implemented by a domains relational description that, when instantiated, produces a conditional random field, which supports the inferences. Models that combine probability and firstorder logic are more expressive and adequate to structure spatial in semantic information.

Construir um robô capaz de realizar tarefas sem qualquer interferência humana é um dos maiores desafios da Robótica Move!. Dispondo apenas de sensores, um robô autônomo precisa explorar ambientes desconhecidos e, simultaneamente, construir um mapa confiável a fim de se localizar e realizar a tarefa. Na presença de erros de odometria, o robô não consegue se auto-localizar corretamente em seu mapa interno e acaba por construir um mapa deformado e não condizente com a realidade. Este trabalho apresenta uma solução para o problema da auto-localização de robô moveis autônomos. Esta solução faz use de um método linear de calculo de estimativas chamado Filtro de Kalman para corrigir a posição do robô em seu mapa intern° do ambiente enquanto realiza a exploração. A proposta leva em consideração que toda entidade que se movimenta em um ambiente conta sempre com alguns pontos de referencia para se localizar. Estes pontos são implementados como objetos especiais chamados marcas de Kalman. Em simulação, o reconhecimento das marcas pode ser feito de duas maneiras: através de sua posição no mapa ou através de sua identidade. Nos experimentos realizados em simulação, o método é testado para diferentes erros no angulo de orientação do robô. Os resultados são comparados levando em consideração as deformações no mapa gerado, com e sem marcas de Kalman, e o erro médio da posição do robô durante todo o processo exploratório. / Build a robot capable of performing tasks without any human interference is one of the biggest challenges of the Mobile Robotics. Having only sensors, an autonomous robot needs to explore unknown environments and, simultaneously, build a reliable map in order to get its own location and perform the task. In the presence of odometry errors, the robot is not capable of establish its own position on its internal map and ends up building a deformed map that does not reflect reality. This paper presents a solution for the problem of self-localization of autonomous mobile robots. This solution uses a linear method for calculating estimatives called Kalman Filter to correct the robot's position on its internal mapping of the environment while exploring. The proposal considers that any being that moves in an environment always counts on having some reference points to establish its own position. This points are implemented as special objects called Kalman landmarks. In simulation, the recognition of such landmarks can be done in two different ways: through its position on the map or through its identity. In the experiments performed in simulations, the method is tested for different errors in the robot's inclination angle. The results are compared considering the deformations on the generated map, with and without the Kalman landmarks, and the average error of the robot's position during the exploratory process.

Construir um robô capaz de realizar tarefas sem qualquer interferência humana é um dos maiores desafios da Robótica Move!. Dispondo apenas de sensores, um robô autônomo precisa explorar ambientes desconhecidos e, simultaneamente, construir um mapa confiável a fim de se localizar e realizar a tarefa. Na presença de erros de odometria, o robô não consegue se auto-localizar corretamente em seu mapa interno e acaba por construir um mapa deformado e não condizente com a realidade. Este trabalho apresenta uma solução para o problema da auto-localização de robô moveis autônomos. Esta solução faz use de um método linear de calculo de estimativas chamado Filtro de Kalman para corrigir a posição do robô em seu mapa intern° do ambiente enquanto realiza a exploração. A proposta leva em consideração que toda entidade que se movimenta em um ambiente conta sempre com alguns pontos de referencia para se localizar. Estes pontos são implementados como objetos especiais chamados marcas de Kalman. Em simulação, o reconhecimento das marcas pode ser feito de duas maneiras: através de sua posição no mapa ou através de sua identidade. Nos experimentos realizados em simulação, o método é testado para diferentes erros no angulo de orientação do robô. Os resultados são comparados levando em consideração as deformações no mapa gerado, com e sem marcas de Kalman, e o erro médio da posição do robô durante todo o processo exploratório. / Build a robot capable of performing tasks without any human interference is one of the biggest challenges of the Mobile Robotics. Having only sensors, an autonomous robot needs to explore unknown environments and, simultaneously, build a reliable map in order to get its own location and perform the task. In the presence of odometry errors, the robot is not capable of establish its own position on its internal map and ends up building a deformed map that does not reflect reality. This paper presents a solution for the problem of self-localization of autonomous mobile robots. This solution uses a linear method for calculating estimatives called Kalman Filter to correct the robot's position on its internal mapping of the environment while exploring. The proposal considers that any being that moves in an environment always counts on having some reference points to establish its own position. This points are implemented as special objects called Kalman landmarks. In simulation, the recognition of such landmarks can be done in two different ways: through its position on the map or through its identity. In the experiments performed in simulations, the method is tested for different errors in the robot's inclination angle. The results are compared considering the deformations on the generated map, with and without the Kalman landmarks, and the average error of the robot's position during the exploratory process.

A maior parte dos mapas empregados em tarefas de navegação por robôs móveis representam apenas informações espaciais do ambiente. Outros tipos de informações, que poderiam ser obtidos dos sensores do robô e incorporados à representação, são desprezados. Hoje em dia é comum um robô móvel conter sensores de distância e um sistema de visão, o que permitiria a princípio usá-lo na realização de tarefas complexas e gerais de maneira autônoma, dada uma representação adequada e um meio de extrair diretamente dos sensores o conhecimento necessário. Uma representação possível nesse contexto consiste no acréscimo de informação semântica aos mapas métricos, como por exemplo a segmentação do ambiente seguida da rotulação de cada uma de suas partes. O presente trabalho propõe uma maneira de estruturar a informação espacial criando um mapa semântico do ambiente que representa, além de obstáculos, um vínculo entre estes e as imagens segmentadas correspondentes obtidas por um sistema de visão omnidirecional. A representação é implementada por uma descrição relacional do domínio, que quando instanciada gera um campo aleatório condicionado, onde são realizadas as inferências. Modelos que combinam probabilidade e lógica de primeira ordem são mais expressivos e adequados para estruturar informações espaciais em semânticas. / Most maps used in navigational tasks by mobile robots represent only environmental spatial information. Other kinds of information, that might be obtained from the sensors of the robot and incorporated in the representation, are negleted. Nowadays it is common for mobile robots to have distance sensors and a vision system, which could in principle be used to accomplish complex and general tasks in an autonomously manner, given an adequate representation and a way to extract directly from the sensors the necessary knowledge. A possible representation in this context consists of the addition of semantic information to metric maps, as for example the environment segmentation followed by an attribution of labels to them. This work proposes a way to structure the spatial information in order to create a semantic map representing, beyond obstacles, an anchoring between them and the correspondent segmented images obtained by an omnidirectional vision system. The representation is implemented by a domains relational description that, when instantiated, produces a conditional random field, which supports the inferences. Models that combine probability and firstorder logic are more expressive and adequate to structure spatial in semantic information.