Desarrollo de un prototipo de seguimiento corporal utilizando visión artificial para análisis de marcha.

Abstract

The objective of the study was to develop a body tracking prototype using machine vision for gait analysis. Initially, the problem examined around the subject of study was raised. Sufficient information and documentation were collected for the analysis of the factors and parameters required in the processing of information. The project was developed in three stages; The first stage consisted of studying the various body tracking technologies and selecting the technology and tools that use neural networks to process the images. The second stage consisted of the design and development of the necessary algorithms in the body tracking prototype for gait analysis focused on the gait cycle. A scenario was proposed in which a starting point and an end point were established by means of cones, the distance between the safety cones is 1.90 m, the distances between the center of the stage and camera 1 of the side view is of 1.90 m, the distance between camera 2 of the rear view is 2 m. The height of the cameras must be 75 cm. The third stage was based on the implementation of the system through an executable program that processes the acquired images provided by mobile devices connected wirelessly through a wireless local area network (WLAN), in order to perform performance tests. It is concluded that the prototype will serve as a support tool for gait analysis for operators in the field, the project opens a viable approach with low-cost hardware, thus providing a support solution for economical and efficient gait analysis, in contrast to specialized laboratories, focused on capturing human movement and locomotion. It is recommended in body tracking studies to study and develop more utilities for gait analysis and human locomotion through computer vision.