Reconocimiento automático de cartas de baraja ubicadas sobre una mesa
This work presents the recognition of playing cards on a table as part of a larger project wich purpose is to assist blind people in playing card games, specifically the Ecuadorian game called "Cuarenta". The solution proposed in the prototype takes into account the specific rules of the &...
| Autor Principal: | Cisneros Navarrete, Soraya Karina |
|---|---|
| Otros Autores: | Medina Encalada, Leonardo Patricio |
| Formato: | bachelorThesis |
| Idioma: | spa |
| Publicado: |
2017
|
| Materias: | |
| Acceso en línea: |
http://dspace.ups.edu.ec/handle/123456789/14564 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| Sumario: |
This work presents the recognition of playing cards on a table as part of a larger project wich purpose is to assist blind people in playing card games, specifically the Ecuadorian game called "Cuarenta". The solution proposed in the prototype takes into account the specific rules of the "Cuarenta" game; that is to say, it recognizes the number of the 40 decks in play (the symbol, color and the cards with numbers 8, 9 and 10 are discarded). In addition, it solves difficulties that could present environmental aspects like perspective illumination, etc.
The development of the prototype was based on the standard procedure for the digital image processing, so for the pre-processing, image enhancement methods such as: Binarization, Adaptive Umbralización, Dilation and Erosion were used. For Image Segmentation, the Connected Component Tagging method was applied and finally for image detection and recognition, we worked with the algorithm of Backpropagation of Artificial Neural Networks.
In the detection phase 98% accuracy is obtained, while for the recognition phase the accuracy is 99% |
|---|