No description available.

Ce mémoire s'intéresse à la reconstruction d'un modèle 3D à partir de plusieurs images. Le modèle 3D est élaboré avec une représentation hiérarchique de voxels sous la forme d'un octree. Un cube englobant le modèle 3D est calculé à partir de la position des caméras. Ce cube contient les voxels et il définit la position de caméras virtuelles. Le modèle 3D est initialisé par une enveloppe convexe basée sur la couleur uniforme du fond des images. Cette enveloppe permet de creuser la périphérie du modèle 3D. Ensuite un coût pondéré est calculé pour évaluer la qualité de chaque voxel à faire partie de la surface de l'objet. Ce coût tient compte de la similarité des pixels provenant de chaque image associée à la caméra virtuelle. Finalement et pour chacune des caméras virtuelles, une surface est calculée basée sur le coût en utilisant la méthode de SGM. La méthode SGM tient compte du voisinage lors du calcul de profondeur et ce mémoire présente une variation de la méthode pour tenir compte des voxels précédemment exclus du modèle par l'étape d'initialisation ou de creusage par une autre surface. Par la suite, les surfaces calculées sont utilisées pour creuser et finaliser le modèle 3D. Ce mémoire présente une combinaison innovante d'étapes permettant de créer un modèle 3D basé sur un ensemble d'images existant ou encore sur une suite d'images capturées en série pouvant mener à la création d'un modèle 3D en temps réel. / This master concentrates on the reconstruction of a 3D model from multiple images. The 3D model is built with a hierarchical representation of voxels using an octree. A cube surrounding the object is calculated from the camera's positions. This cube contains all the voxels and it defines the position of the virtual cameras. The 3d model is initialized by a visual hull that is based on the uniform color of the images’ background. This visual hull is used to pre-carve the 3D model. Then a cost is calculated to evaluate the quality of each voxel as being on the surface of the object. This cost takes into account the similarity of the pixels from each images associated to a virtual camera. Finally a surface is calculated for each virtual camera using the SGM method that is based on the voxel cost. The SGM method takes the surrounding voxels into account when calculating the depth and this master presents a variation to this method where we take the previously excluded voxels into account. The excluded voxels coming from the initialization step or from the carving done by another virtual camera. The resulting surface is used to carve the voxel representation. This master presents an innovative combination of steps leading to the creation of a 3D model from a set of existing images or from a series of images capture one after another leading to a real-time creation of a 3D model.

Syfte: Att ta fram en metod för att hjälpa mindre sågverk att bättre tillvarata mesta möjliga virke från en timmerstock. Metod: En kvantitativ studie där tre iterationer genomförts enligt Design Science. Resultat: För att skapa en effektiv algoritm som ska utföra volymberäkningar i ett punktmoln som består av cirka två miljoner punkter i ett industriellt syfte ligger fokus i att algoritmen är snabb och visar rätt data. Det primära målet för att göra algoritmen snabb är att bearbeta punktmolnet ett minimalt antal gånger. Den algoritm som uppfyller delmålen i denna studie är Algoritm C. Algoritmen är både snabb och har en låg standardavvikelse på mätfelen. Algoritm C har komplexiteten O(n) vid analys av delpunktmoln. Implikationer: Med utgångspunkt från denna studies algoritm skulle det vara möjligt att använda stereokamerateknik för att hjälpa mindre sågverk att bättre tillvarata mesta möjliga virke från en timmerstock. Begränsningar: Studiens algoritm har utgått från att inga punkter har skapats inuti stocken vilket skulle kunna leda till felplacerade punkter. Om en stock skulle vara krokig överensstämmer inte stockens centrum med z-axelns placering. Detta är något som skulle kunna innebära att z-värdet hamnar utanför stocken, i extremfall, vilket algoritmen inte kan hantera. / Purpose: To develop a method that can help smaller sawmills to better utilize the greatest possible amount of wood from a log. Method: A quantitative study where three iterations has been made using Design Science. Findings: To create an effective algorithm that will perform volume calculations in a point cloud consisting of about two million points for an industrial purpose, the focus is on the algorithm being fast and that it shows the correct data. The primary goal of making the algorithm quick is to process the point cloud a minimum number of times. The algorithm that meets the goals in this study is Algorithm C. The algorithm is both fast and has a low standard deviation of the measurement errors. Algorithm C has the complexity O(n) in the analysis of sub-point clouds. Implications: Based on this study’s algorithm, it would be possible to use stereo camera technology to help smaller sawmills to better utilize the most possible amount of wood from a log. Limitations: The study’s algorithm assumes that no points have been created inside the log, which could lead to misplaced points. If a log would be crooked, the center of the log would not match the z-axis position. This is something that could mean that the z-value is outside of the log, in extreme cases, which the algorithm cannot handle.

Level of detail (LOD) systems has long been a central concept within the field of computer graphics. By selectively allocating detail only to where it is necessary, the cost of rendering virtual objects can be greatly decreased. The usefulness of an LOD system is closely tied to the metrics used to evaluate how to distribute detail, as getting these wrong can lead to visual artifacts or rendering detail which is not visible to an observer. To offset this, researchers have investigated how we can optimize these systems by leveraging human perception to allocate detail based on if the detail is perceivable. This paper explores the use of LOD systems in the context of virtual terrains, a common use case were LOD systems have seen much prior use, and delves into how such systems can be optimized further by leveraging human perception. The paper presents a method centered around the use of octrees, a data structure used to partition and compress volumetric data, alongside spatial frequency, a concept related to human perception which provides insight into where visual stimulus borders on being percievable, to dynamically generate transition distances for an LOD system. The presented method is evaluated using a terrain generator created in the Unity engine alongside Nvidia ꟻLIP, a tool for evaluating differences in images based on human perception metrics. Results indicate that the presented method did lead to a decrease in resource cost when generating terrain when compared to some other methods while the ability to optimize LOD transition distances based on silhouette-background contrast remains uncertain. / Nivå av detalj (LOD) system har länge varit ett centralt begrepp inom forskningsområdet för dator-grafik. Genom att selektivt allokera detalj så kan kostnaden för att rendera virituella föremål minskas drastiskt. Användbarheten av ett LOD-system är nära kopplat till de specifika värden som används för att utvärdera hur man fördelar detalj efterssom man annars riskerar att få virtuella artefakter eller renderar detaljer som inte är synliga för en observatör. För att motverka detta har forskare undersökt hur vi kan optimisera dessa system genom att utnyttja visuell uppfattning hos människor för att allokera detalj baserat på var den är synlig för användaren. Denna text utforskar användningen av LOD-system i samband med virtuella terränger, ett användningsområde där LOD-system har använts tidigare och går in på hur denna typ av system kan optimiseras bättre genom att utnyttja mänsklig perception. Texten presenterar en metod centrerad runt användandet av octrees, en datastruktur som används för att partitionera och komprimera volumetrisk data, tillsamans med spatiell frekvens, ett koncept relaterat till mänsklig synförmåga som ger inblick i var visuell stimulans gränsar till att vara synlig, för att dynamiskt generera övergångsavstånd för ett LOD-system. Den presenterade metoden utvärderas med en terränggenerator skapad i Unity spel-motorn tilsamans med Nvidia ꟻLIP, ett verktyg för att utvärdera skillnader i bilder baserat på mänskliga perceptionsmått. Resultaten indikerar att den presenterade metoden ledde till en minskning av resurskostnaden vid generation av terräng jämfört med vissa andra metoder medan möjligheten att optimera övergångsavstånd för LOD-systemet baserat på kontrast mellan bakgrunden och silouetten för terrängen kvarstår som osäker.

Three-dimensional scans and other volumetric data sources often result in representations that are more complex topologically than the original model. The extraneous critical points, handles, and components are called topological noise. Many algorithms in computer graphics require simple topology in order to work optimally, including texture mapping, surface parameterization, flows on surfaces, and conformal mappings. The topological noise disrupts these procedures by requiring each small handle to be dealt with individually. Furthermore, topological descriptions of volumetric data are useful for visualization and data queries. One such description is the contour tree (or Reeb graph), which depicts when the isosurfaces split and merge as the isovalue changes. In the presence of topological noise, the contour tree can be too large to be useful. For these reasons, an important goal in computer graphics is simplification of the topology of volumetric data. The key to this thesis is that the global topology of volumetric data sets is determined by local changes at individual points. Therefore, we march through the data one grid cell at a time, and for each cell, we use a local check to determine if the topology of an isosurface is changing. If so, we change the value of the cell so that the topology change is prevented. In this thesis we describe variations on the local topology check for use in different settings. We use the topology simplification procedure to extract a single component with controlled topology from an isosurface in volume data sets and partially-defined volume data sets. We also use it to remove critical points from three-dimensional volumes, as well as time-varying volumes. We have applied the technique to two-dimensional (plus time) data sets and three dimensional (plus time) data sets.

The first component of this work is a parallel algorithm for constructing non-uniform octree meshes for finite element computations. Prior to octree meshing, the linear octree data structure must be constructed and a constraint known as "2:1 balancing" must be enforced; parallel algorithms for these two subproblems are also presented. The second component of this work is a parallel matrix-free geometric multigrid algorithm for solving elliptic partial differential equations (PDEs) using these octree meshes. The last component of this work is a parallel multiscale Gauss Newton optimization algorithm for solving the elastic image registration problem. The registration problem is discretized using finite elements on octree meshes and the parallel geometric multigrid algorithm is used as a preconditioner in the Conjugate Gradient (CG) algorithm to solve the linear system of equations formed in each Gauss Newton iteration. Several ideas were used to reduce the overhead for constructing the octree meshes. These include (a) a way to lower communication costs by reducing the number of synchronizations and reducing the communication message size, (b) a way to reduce the number of searches required to build element-to-vertex mappings, and (c) a compression scheme to reduce the memory footprint of the entire data structure. To our knowledge, the multigrid algorithm presented in this work is the only matrix-free multiplicative geometric multigrid implementation for solving finite element equations on octree meshes using thousands of processors. The proposed registration algorithm is also unique; it is a combination of many different ideas: adaptivity, parallelism, fast optimization algorithms, and fast linear solvers. All the algorithms were implemented in C++ using the Message Passing Interface (MPI) standard and were built on top of the PETSc library from Argonne National Laboratory. The multigrid implementation has been released as an open source software: Dendro. Several numerical experiments were performed to test the performance of the algorithms. These experiments were performed on a variety of NSF TeraGrid platforms. Our largest run was a highly-nonuniform, 8-billion-unknown, elasticity calculation on 32,000 processors.

The dissertation deals with utilisation of multispectral optical measurement for data fusion that may be used for visual telepresence and indoor/outdoor mapping by heterogeneous mobile robotic system. Optical proximity sensors, thermal imagers, and tricolour cameras are used for the fusion. The described algorithms are optimised to work in real-time and implemented on CASSANDRA robotic system made by our robotic research group.

This thesis deals with the global illumination in the scene by using Voxel Cone Tracing method. It is based on the voxelization of a triangle mesh scene. The voxels can be stored to a full regular 3D grid (texture) or to the hierarchic Sparse Voxel Octree for saving of the memory space. This voxel representation is further used for computations of the global indirect illumination in real time within normal triangle mesh scenes for more realistic final image. Values from the voxels are obtained by tracing cones from the pixels which we want to get illumination for.

No description available.

This thesis examines an automated method to detect surface deformations on automotive panels for the purpose of quality control along a manufacturing assembly line. Automation in the automotive manufacturing industry is becoming more prominent, but quality control is still largely performed by human workers. Quality control is important in the context of automotive body panels as deformations can occur along the assembly line such as inadequate handling of parts or tools around a vehicle during assembly, rack storage, and shipping from subcontractors. These defects are currently identified and marked, before panels are either rectified or discarded. This work attempts to develop an automated system to detect deformations to alleviate the dependence on human workers in quality control and improve performance by increasing speed and accuracy. Some techniques make use of an ideal CAD model behaving as a master work, and panels scanned on the assembly line are compared to this model to determine the location of deformations. This thesis presents a solution for detecting deformations of various scales without a master work. It also focuses on automated analysis requiring minimal intuitive operator-set parameters and provides the ability to classify the deformations as dings, which are deformations that protrude from the surface, or dents, which are depressions into the surface. A complete automated deformation detection system is proposed, comprised of a feature extraction module, segmentation module, and classification module, which outputs the locations of deformations when provided with the 3D mesh of an automotive panel. Two feature extraction techniques are proposed. The first is a general feature extraction technique for 3D meshes using octrees for multi-resolution analysis and evaluates the amount of surface variation to locate deformations. The second is specifically designed for the purpose of deformation detection, and analyzes multi-resolution cross-sections of a 3D mesh to locate deformations based on their estimated size. The performance of the proposed automated deformation detection system, and all of its sub-modules, is tested on a set of meshes which represent differing characteristics of deformations in surface panels, including deformations of different scales. Noisy, low resolution meshes are captured from a 3D acquisition, while artificial meshes are generated to simulate ideal acquisition conditions. The proposed system shows accurate results in both ideal situations as well as non-ideal situations under the condition of noise and complex surface curvature by extracting only the deformations of interest and accurately classifying them as dings or dents.