Detection system for humanoid robots in semi-structured environments based on stereoscopic vision
Sistema de detección de robots humanoides en ambientes semiestructurados basado en visión estereoscópica
| 2021 | |
|
Visión artificial Video estéreo Red neuronal convolucional Robots humanoides Artificial vision Stereo video Convolutional neural network Humanoid robots |
|
| Español | |
| Universidad de Montevideo | |
| REDUM | |
| http://revistas.um.edu.uy/index.php/ingenieria/article/view/855 | |
| Acceso abierto | |
| Atribución 4.0 Internacional |
| Sumario: | This work proposes the design, development and implementation of an artificial vision system based on stereo video, which is executed in an embedded system, to identify humanoid robots within a semi-structured area. The embedded system uses an Intel RealSense camera that, in addition to being able to obtain distances to objects due to its stereo vision, is capable of discriminating information after a distance threshold determined by the user, eliminating objects in the background of the scene and, through Image recognition based on a convolutional neural network recognizes the humanoid robots within it. The application of the system is based on the Robocup Humanoid League contest, where two teams of robots play soccer, so in addition to recognizing humanoid robots at different angles and distances, the system can classify recognized robots as companions or opponents (depending on the recognition of color marks they carry), emulating that, in the future, the proposed system will be mounted on another humanoid robot. |
|---|
