RPAS: A Refined Positioning and Analysis System Based on Aerial Platform for Swimming Scenes

In modern competitive swimming, the evaluation of position, speed and technique of athletes in swimming is of significant importance for improving grades. In past years, there are plenty of methods to evaluate the position and speed of swimmers, in-cluding speedometer, IMU(Inertial Measurement Unit), GPS(Global Positioning System)+accelerometer. To overcome some of their disadvantages, i.e., inconvenient installation, complicated speed calculation, and limited scenarios, we propose a refined positioning and analysis system based on aerial platform for swimming scene, which includes the unmanned aircraft system, the aruco positioning system, and the siamban tracking system. UAV hovers at a fixed height to capture videos of swimmers. The siamban tracking system tracks the position of swimming cap, mapping the the position to the real world through the aruco positioning system. In static experiment, the minimum position error is 0.002m, when the ground truth is 2.1m. The MAE is 0.0231m, which has declined 74.3% compared to IMU, showing high precision of our method. In addition, we have implemented applications in dynamic swimming scenes, demonstrating great potential in some non-contact measurement scenes of our method.