ФОТОГРАММЕТРИЧНА ПІДСИСТЕМА ЗАХОПЛЕННЯ РУХУ ДЛЯ АНАЛІЗУ ЗМІН ПОЛОЖЕННЯ ТІЛА ВІДНОСНО ФРОНТАЛЬНОЇ ТА САГІТАЛЬНОЇ ПЛОЩИН

Context. The increase in other orthopedic injuries, particularly among military personnel, requires new innovative solutions to assess posture changes and monitor rehabilitation effectiveness. Existing systems have limitations in terms of portability, cost, and flexibility of use, which necessitates the development of hybrid systems that combine computer vision and sensory analysis methods.Objective. To assess the effectiveness of combining non-contact computer vision and accelerometric sensors for detecting changes in human posture under different lighting, background, and movement speeds.Method. The study implemented a photogrammetric subsystem that includes MediaPipe Holistic for markerless tracking of key body points and WitMotion WT9011DCL-BT50 accelerometers for analyzing inertial motion parameters. The system model was built in IDEF0 notation. The accuracy was assessed by comparing the obtained values of the blade inclination angle and the asymmetry coefficient with the specified norms.Results. The combined use of visual and sensory data made it possible to reduce the error to 5.05% under normal conditions and ensure the stability of the results under conditions of changes in the external environment. Image modification (contrast, noise filtering) increased the accuracy of computer vision. Threshold values of the asymmetry coefficient corresponding to normal, mild and severe postural disorders were determined.Conclusions. The proposed system demonstrates high potential effectiveness in telemedical rehabilitation support for patients with musculoskeletal disorders. Its practical significance lies in the creation of an affordable, portable, and accurate diagnostic and monitoring tool suitable for further integration into personalized medicine systems with built-in artificial intelligence modules.