Avaliação de métodos de rastreamento de marcadores para um sistema óptico de captura de movimento

Abstract

There are many applications for a motion capture system. Systems that record human motion can be employed in several fields, such as medicine, sports, security and entertainment. Nevertheless the usefulness and benefits provided by these systems are many, its use is restrictedly disseminated, mainly in Brazil. This work is part of a larger effort to create the \textit{OpenMoCap}, a Brazilian and open source software for motion capture, which uses cameras and passive markers. In this work, three tracking methods are used in the task of tracking passive markers for optical motion capture. The methods evaluated and compared are: the Kalman filter, the Alpha-Beta filter and the CONDENSATION algorithm. To evaluate the performance of the methods the robustness of the tracking was tested and the computational cost measured. Considering the overall performance, the Kalman filter was chosen to be employed in the capture system.