|
|
Visual Localization in Natural Environments
Brejcha, Jan ; Sattler, Torsten (oponent) ; Matas, Jiří (oponent) ; Čadík, Martin (vedoucí práce)
We focus our work on camera position and orientation estimation given a query photograph; we call this problem visual geo-localization. Specifically, we focus on photographs captured in natural, mountainous environments. We introduce a thorough review of state-of-the-art computer vision methods, datasets, and evaluation practices for visual geo-localization problems. The survey revealed that researchers usually cast visual geo-localization in natural environments as a similarity or a correspondence search between an input photograph and a terrain model; we call this problem the cross-domain matching. We identified three main goals to improve over the state of the art in visual geo-localization in mountainous environments using cross-domain matching: (I) the need for new datasets for training, validation, and evaluation of cross-domain visual geo-localization algorithms, (II) the need to verify whether the cross-domain matching algorithms may benefit from using different features-horizon lines, edge maps, semantic segmentation, and satellite imagery, (III) the need to illustrate the usefulness of visual geo-localization methods by developing novel applications. In this thesis, we thoroughly describe our research studies to illustrate how we examined particular goals. We introduce several novel datasets for evaluation and training of cross-domain matching methods. These novel datasets allowed us to propose a novel method for cross-domain photo-to-terrain matching using a combination of semantic segments and classic edge-based features. We illustrate the benefits of our novel approach over the state of the art on camera orientation estimation. Furthermore, we propose a meta-algorithm based on a cross-domain Structure from Motion for a weakly supervised acquisition of cameras aligned with the synthetic terrain. This novel cross-domain data acquisition scheme allowed us to train a compact cross-domain keypoint descriptor. We illustrate the descriptor performance by estimating full camera pose by matching the query photograph to the rendered terrain model. Finally, we demonstrate a practical usability of outdoor visual geo-localization by designing a novel application of photography presentation on a computer screen or in virtual reality. Moreover, we illustrate that our novel presentation method helps the user with complex outdoor scene understanding and improves self-localization in unvisited outdoor environments.
|
|
|
Implementace algoritmu Seamless Patches for GPU-Based Terrain Rendering
Jozefov, David ; Polok, Lukáš (oponent) ; Bartoň, Radek (vedoucí práce)
Tato diplomová práce se zabývá vykreslováním terénu s využitím moderního algoritmu pro adaptivní úroveň detailů. Popisuje dvě v současnosti nejpoužívanější grafická aplikační rozhraní a jejich high-level nadstavby a shrnuje princip a vlastnosti několika používaných level-of-detail algoritmů pro zobrazování terénu. Podrobněji pak popisuje implementaci algoritmu Seamless patches for GPU-based terrain rendering.
|
|
|
Mobilní aplikace zobrazující model terénu v reálném čase
Hejl, Radim ; Tomešek, Jan (oponent) ; Čadík, Martin (vedoucí práce)
Tato bakalářská práce se zabývá vykreslováním terénu na mobilním zařízení v reálném čase. Zaměřuje se na aktuální vývoj aplikací v operačním systému Android, zejména vykreslování grafiky pomocí OpenGL ES. Na základě těchto informací a aktuálního stavu již existujících řešení je navrhnuta, implementována a testována výsledná aplikace, která vykresluje okolní terén a za pomoci senzorů zařízení v něm naviguje.
|
|
|
Visual Localization in Natural Environments
Brejcha, Jan ; Sattler, Torsten (oponent) ; Matas, Jiří (oponent) ; Čadík, Martin (vedoucí práce)
We focus our work on camera position and orientation estimation given a query photograph; we call this problem visual geo-localization. Specifically, we focus on photographs captured in natural, mountainous environments. We introduce a thorough review of state-of-the-art computer vision methods, datasets, and evaluation practices for visual geo-localization problems. The survey revealed that researchers usually cast visual geo-localization in natural environments as a similarity or a correspondence search between an input photograph and a terrain model; we call this problem the cross-domain matching. We identified three main goals to improve over the state of the art in visual geo-localization in mountainous environments using cross-domain matching: (I) the need for new datasets for training, validation, and evaluation of cross-domain visual geo-localization algorithms, (II) the need to verify whether the cross-domain matching algorithms may benefit from using different features-horizon lines, edge maps, semantic segmentation, and satellite imagery, (III) the need to illustrate the usefulness of visual geo-localization methods by developing novel applications. In this thesis, we thoroughly describe our research studies to illustrate how we examined particular goals. We introduce several novel datasets for evaluation and training of cross-domain matching methods. These novel datasets allowed us to propose a novel method for cross-domain photo-to-terrain matching using a combination of semantic segments and classic edge-based features. We illustrate the benefits of our novel approach over the state of the art on camera orientation estimation. Furthermore, we propose a meta-algorithm based on a cross-domain Structure from Motion for a weakly supervised acquisition of cameras aligned with the synthetic terrain. This novel cross-domain data acquisition scheme allowed us to train a compact cross-domain keypoint descriptor. We illustrate the descriptor performance by estimating full camera pose by matching the query photograph to the rendered terrain model. Finally, we demonstrate a practical usability of outdoor visual geo-localization by designing a novel application of photography presentation on a computer screen or in virtual reality. Moreover, we illustrate that our novel presentation method helps the user with complex outdoor scene understanding and improves self-localization in unvisited outdoor environments.
|
|
|
Deriving suitable surface shader and displacement map information from terrain erosion simulations
Lanza, Dario ; Wilkie, Alexander (vedoucí práce) ; Mirbauer, Martin (oponent)
Realistic models of landscapes are frequently needed for 3D renderings, VFX work or video games. However, modelling landscapes can be a complicated and labour-intensive task, and for this reason many algorithms have been proposed to automate the process. Among the many possible ways to create a synthetic landscape, the most common one is to simulate the various types of erosions (e.g. erosion caused by glaciers and rivers) that create real eroded landscapes, like the Grand Canyon. Many solutions have been published to simulate such terrain erosion processes in computer graphics. However the authors usually only focus on recreating a landscape at geometry level, and ignore the shading level. But surface colours and textures that match the coarse geometric features created by the erosion simulation are also essential ingredients for a believable result. And obtaining detailed surface textures by running a simulation that is able to catch all the micro-details involved is usually technically infeasible due to the involved complexity. The method that we propose attempts to get around this barrier by applying suitable detail shaders to the results of a coarse-grid erosion simulation. Specifically, we will work with a dictionary of pre-generated shaders for landscape appearance: these will be both "plain" colour...
|
|
|
Mobilní aplikace zobrazující model terénu v reálném čase
Hejl, Radim ; Tomešek, Jan (oponent) ; Čadík, Martin (vedoucí práce)
Tato bakalářská práce se zabývá vykreslováním terénu na mobilním zařízení v reálném čase. Zaměřuje se na aktuální vývoj aplikací v operačním systému Android, zejména vykreslování grafiky pomocí OpenGL ES. Na základě těchto informací a aktuálního stavu již existujících řešení je navrhnuta, implementována a testována výsledná aplikace, která vykresluje okolní terén a za pomoci senzorů zařízení v něm naviguje.
|
|
|
Implementace algoritmu Seamless Patches for GPU-Based Terrain Rendering
Jozefov, David ; Polok, Lukáš (oponent) ; Bartoň, Radek (vedoucí práce)
Tato diplomová práce se zabývá vykreslováním terénu s využitím moderního algoritmu pro adaptivní úroveň detailů. Popisuje dvě v současnosti nejpoužívanější grafická aplikační rozhraní a jejich high-level nadstavby a shrnuje princip a vlastnosti několika používaných level-of-detail algoritmů pro zobrazování terénu. Podrobněji pak popisuje implementaci algoritmu Seamless patches for GPU-based terrain rendering.
|
host ::
RSS.