|
Descripción
|
|
|---|---|
| This article focuses its objective in presenting the results obtained in the research done for surface detection using non-invasive vision-based tech-niques. Two different algorithms based on color and depth information, respectively, are compared to obtain the most optimal and accurate solu-tion for the detection of a table surface. This detection will act as first stage to create a workspace for the monitoring of hands movement and object manipulation during handmade daily activities. The research has been carried out at the Centre for Automation and Robotics by using OpenCV libraries for image processing and sensor KinectTM, as a camera device. The main idea is to evaluate how a table can be properly detected and positioned in order to provide reliable information about the relative position of hands while preparing drinks or even food. Continuous and robust data of hands attitude is essential to analyze trajectories, even more considering cognitive rehabilitation techniques of those people who suffer from brain injuries. First of all, a short introduction to other works related in the field is done and the importance of hand kinematics in cognitive rehabilitation is also highlighted. Later, the current solution adopted is described to finally present the methodology followed, results and final conclusions. | |
|
Internacional
|
No |
|
Entidad
|
Centro de Automática y Robótica CAR (UPM-CSIC) |
|
Lugar
|
Madrid |
|
Páginas
|
|
|
Referencia/URL
|
http://www.car.upm-csic.es/events/robocity16/ |
|
Tipo de publicación
|
Ponencia en Workshop RoboCity16 Open Conference on Future Trends in Robotics, RoboCity2030 Madrid's Robotics Hub |