Desarrollo de un sistema de captura de movimiento abierto para análisis biomecánico

Abstract

This work presents the development of an open motion capture system based on OptiTrack Flex 3 cameras, aiming to reduce proprietary software dependency through an accessible, reproducible, and verifiable workflow. The study was conducted in a controlled environment at the Girón Campus under a quantitative–experimental approach structured into modules: (i) low-level camera communication using the Camera SDK and MEX compilation to initialize the SDK, detect devices, and acquire synchronized infrared frames; (ii) reflective marker acquisition and assessment, identifying practical limitations due to clothing, specular reflections, and occlusions; (iii) intrinsic calibration; (iv) extrinsic parameter estimation and validation to unify the multicamera setup into a common reference frame defined by the calibration board; and (v) 3D reconstruction via triangulation and implementation of measurement modules (2D/3D) to estimate joint angles from three markers. During testing, 3D reconstruction performance was observed to degrade when extrinsic calibration was constrained to a nearly planar motion of the board (e.g., board placed on a table); therefore, a complementary 2D validation strategy was implemented, including angle estimation using a goniometer reference and additional trials for relative distance measurement after planar experimental scale calibration. Preliminary results confirm the technical feasibility of the proposed system for IR acquisition, marker detection, and geometric reconstruction under proper calibration, and provide a solid foundation for future accuracy validation and standardized data export toward environments such as MATLAB.