Visualização e reconstrução de ambientes utilizando técnicas de estereoscopia

• Main Author: Araujo, Rodrigo Catão
• Publication Date: 2016
• Format: Bachelor thesis
• Language: por
• Source: Repositório Institucional da UTFPR (da Universidade Tecnológica Federal do Paraná (RIUT))
• Download full: http://repositorio.utfpr.edu.br/jspui/handle/1/12493
• topic: Robótica; Inteligência computacional; Aprendizado do computador; Robotics; Computational intelligence; Machine learning; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO

Computer vision and robotics are studied for decades, however, the development/enhancement of computing devices made possible to use these technologies together. Such technologies can be applied in many areas, such as the use in military or professional for applications of Search and Rescue (SAR). This situation has been extensively studied in robotics and computer vision, because it approaches a scenario where the conditions are not suitable for human beings, in other words, it is inaccessible or too dangerous. Although, in a more complex situations of SAR, the use of autonomous robots is not possible, for this reason most parts of robots sent to SAR environments are controlled remotely. In a scenario of SAR, the operator must have a heightened awareness of their surroundings. Therefore there is a need to develop a technology to improve this control, i.e. improve the operator efficiency. The technology most used in these situations are GUIs, however, these GUIs have many spaces that are useless or with almost no use. A solution for this problem is the use of a head mounted display (HMD). Based on this, the operator can see the video as a 3D world, with the overall sense of depth, something that would not be possible using a common system. The purpose of this study is to use the HMD technologies in conjunction with stereoscopy techniques in order to allow better viewing experience for the operator and also to make a 3D environment reconstruction. At the end of this work it was developed a remotely controlled robotic device that allows to reconstruct the environment in three dimensions, and that allows the operator an immersion in the environment by using stereoscopy techniques.