Visão estéreo para auxílio à navegação autônoma em ambientes agrícolas

• Autor(a) principal: Trentini, Vinicius Bergoli [UNESP]
• Data de Publicação: 2015
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://hdl.handle.net/11449/127694; http://www.athena.biblioteca.unesp.br/exlibris/bd/cathedra/11-09-2015/000845659.pdf
• topic: Computação; Processamento de imagens; Técnicas digitais; Visão por computador; Robos moveis; Agricultura de precisão

efficient equipments on its specific functions. Considering the context of computer vision, the development of an autonomous navigation system aimed to external rural farming environments, considers, besides the difficulties faced on closed and controlled environments, which can be targeted as irregular lands and problems with not controlled lighting. By using the stereo vision to perform the navigation in an external farming environment is possible, through the creation of the disparity map, and the v-disparity matrix, verify through what depth an area is possible to be navigated with just one scene and still map the obstacles and deflections along the way. Therefore, calculating the disparity map of the observed scene, it is possible, using the v-disparity matrix, map the objects in the scene according to their depth. Based on the mapping of the objects it is possible to perform a obstructions free navigation along all the way. The aim of this study is, considering an external agricultural environment, map the objects in the scene according to their depth making it possible to identify an obstruction free navigation path. Experiments were performed using an own database, composed of 5.379 pictures in video sequence form, all representing the test's environment. The developed methodology showed an accuracy rate of approximately 92%, with results calculated in real time, proving itself, as an efficient technique for autonomous navigation in agricultural and external areas