Desenvolvimento de um sistema de posicionamento local para monitoramento da natação

Abstract

With advances in technology and miniaturization of electronic devices, such as GPS and accelerometers, and means for wireless data transmission, it has become common the use of these systems for telemetry training in dry sports with long distances in open or closed facilities. However for training in the swimming pool, or aquatic field, most systems used for analysis of swimming technique are based on visual supervision (by coach), or sensor-based systems, or yet using techniques in which athletes are linked directly to the measuring apparatus by means of wire. The solution proposed in this work, the Sistema de Posicionamento Local usando Imagens - SPLI (Local Positioning System using Images), is a system aimed to measure the kinematic variables without physical connection to the swimmer using regular USB camera footage, in real time. Using real-time processing, the results can be analyzed immediately beside the pool, or saved for later analysis. The proposed system was implemented in hardware and software through the LabView development platform from National Instruments and the evaluation of system performance was conducted through laboratory experiments (miniature pool) and in real situations (25m pool). Several displacements, around some centimeters, were evaluated in the miniature pool. We calculated the lowest circunference which can be identified by the system and this value was 4 pixels radius corresponding to a marker of 0.72 cm in diameter in the real world simulation. For the real pool of 25 meters tests, which was filmed in perspective, the system calculated the displacement of the athlete during his trajectory. The search for the swimmers position was done using the change of luminance and the system calculated, correctly, a displacement of 25 m. The measurement of the kinematic variables through a system without connection has the advantage of allowing the athlete to move freely without interference. Another advantage would be the easiness to follow long term training or tests, an important feature since the athlete's performance during execution may vary, being different from the initial period of measurement which is used by current evaluating systems. The results of both experiments were satisfactory, showing that the design, implementation and testing of Local Positioning System (LPS) using Images were conducted successfully.