IDENTIFICAÇÃO DE OBJETOS DO FUTEBOL DE ROBÔS UTILIZANDO ALGORITMO DE DESCRIÇÃO DE PONTOS CHAVE

• Main Authors: Ricardo da Silva Barros, Francisco Assis da Silva, Mário Augusto Pazoti, Danillo Roberto Pereira, Leandro Luiz de Almeida, Almir Olivette Artero
• Format: Article
• Language: Portuguese
• Published: Universidade do Oeste Paulista 2015-06-01
• Series: Colloquium Exactarum
• Subjects: Keypoint Descriptors; Robot Soccer; Computer Vision; SIFT; GPU
• Online Access: http://revistas.unoeste.br/revistas/ojs/index.php/ce/article/view/1427/1463

This work proposes a method implemented in computer vision part required to build a robot soccer, which provides the position and orientation of the ball and robots, using an algorithm description of keypoint. SIFT algorithm was chosen because it provided better results than other similar algorithms. To achieve adequate time for processing to robot soccer, we used the video card processor (GPU) with the algorithm, making obtaining location information and orientation closer to the real-time. The view is just one of the parts that make up the construction of a robot soccer. Object classes were defined and shaped in the C ++ programming language to the minimum necessary tasks of this part of view, in order to facilitate and to allow the development of other parts (mechanical construction, robots control and game strategy) to be a full robot football.