Design of Low-Cost High Accuracy Motion Capture System

In this study, a low-cost passive marker-based optical motion capture system was developed and experimentally validated as an alternative to high-cost commercial systems that are widely regarded as the industry standard. The proposed system consists of multiple IP cameras operating in IR mode, passive reflective markers, and computer vision-based image processing algorithms. Camera calibration was completed with reprojection errors ranging from 0.02 to 0.04 pixels. The markers were detected with sub-pixel accuracy, and their three-dimensional positions were calculated using triangulation. To evaluate the system, four passive markers with 10 cm spacing were mounted on a platform. Experiments were conducted while the platform remained stationary and while it was moved by hand along the X, Y and Z axes. The three-dimensional distances between the markers were analysed using RMSE, MAE, signed mean error (bias), and standard deviation metrics. Under static conditions, the frame-to-frame standard deviation of the calculated distances ranged from 0.086 to 0.204 mm, while the RMSE values ranged from 0.222 to 3.360 mm. The dominant source of error was found to be fixed offsets caused by the manual placement of the markers and the tolerance of the reference measurement device, rather than measurement noise. The findings demonstrate that, despite its low cost, the proposed system provides sub-millimeter repeatability and millimeter-level accuracy, making it a feasible alternative to industrial motion capture systems for static and low-speed applications.