3D modeling of large structures in augmented reality

• Ano de defesa: 2019
• Autor(a) principal: Soares, Leonardo Pavanatto
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Pontifícia Universidade Católica do Rio Grande do Sul; Escola Politécnica; Brasil; PUCRS; Programa de Pós-Graduação em Ciência da Computação
• Palavras-chave em Português: Augmented Reality; 3D Modeling; Point Marking Techniques; Massing Studies; CIENCIA DA COMPUTACAO::TEORIA DA COMPUTACAO
• Link de acesso: http://tede2.pucrs.br/tede2/handle/tede/8715

Three-dimensional modeling in augmented reality allows the user to create or modify the geometry of virtual content while it is registered to the real world, enabling the verification of the correspondence between the model and the real world on-the-fly. We propose a new approach for modeling in augmented reality on environments with unknown geometry that uses point marking techniques to define the position of features present in the real world, such as neighboring buildings and pathways. Some tools were designed to increase the usefulness of these features and to allow the act of modeling, that usually takes place over empty terrains but must consider the surrounding structures. We developed a use case application for the architecture domain, aimed at allowing the design of a massing model in-situ. A massing model is a simplified representation of the architect's idea for the building, defining its general size and location in the terrain, while in-situ refers to being at the place where the building will be made. It aims at understanding the building, the characteristics of the environment, and how they related. Our application allows the architect to create this model while they can visualize the real world and make modifications accordingly. We analyze the problems regarding the approach and propose solutions and evaluated the application under formal user studies. Results indicated it is a reasonable approach for modeling a mass in-situ. We also evaluated point marking techniques with different levels of precision and the use of perception to align lines, and concluded that low precision marking techniques and the use of perception to perform alignments impact the accuracy of the model, and the ease of use, and usefulness of the application.