CBL - Campus del Baix Llobregat

Projecte llegit

Títol: Análisis de imagen en competiciones de petanca


Director/a: TARRÉS RUIZ, FRANCESC

Departament: TSC

Títol: Análisis de imagen en competiciones de petanca

Data inici oferta: 11-07-2017     Data finalització oferta: 11-03-2018



Estudis d'assignació del projecte:
    GR ENG SIS TELECOMUN
Tipus: Individual
 
Lloc de realització: EETAC
 
Paraules clau:
procesado de imagen, medidas,
 
Descripció del contingut i pla d'activitats:
Se trata de diseñar un sistema de cámaras que permita registrar y transmitir por
streaming los vídeos capturados durante una competición de petanca. El sistema
deberá
ser portable y facilmente instalable para la captura de las imágenes, la
instalación de
servidores de streaming, etc. Además, se considerará la posibilidad de instalar
módulos
independientes que permitan realizar algunos análisis automáticos del contenido
como
por ejemplo el cálculo de distancias entre bolas, la posibilidad de realizar
resúmenes
automáticos, etc.
 
Overview (resum en anglès):
It is about designing a system that allows you to measure balls from a reference, without having to bend over. To do this, a smartphone is used, which will be the device by which an intuitive image can be obtained for the user so that he knows which ball wins. The best way to make feasible the idea of measuring the distance between balls and mini ball is to implement it on the Android platform. Each user uses an app of his petanque club in which he has integrated a module capable of capturing videos and images through the camera of the mobile phone and that has implemented enough algorithms to reach the final image.

What has really been achieved as the project has been developed is, an algorithm that under certain conditions, is able to detect what is the bottom of the objects that you want to work in question and then is able to label circular objects of different sizes of template to finally take the centers of mass of each position and calculate the distances using a programming environment designed for MATLAB and not for Android as had been thought at first. Broadly speaking, it is sought to develop a module that will be implemented in a mobile app capable of extracting the characteristics in an environment from a situation where it is required to proceed to measure the balls with our app, as a reference support of ' eye of falcon 'to decide correctly if by the vision of the human eye is not able to appreciate which are closer and which farther from the bowling alley.

The methods that will be carried out for the study of the project are: point transformation for binarization, estimation of threshold for the detection of contours, suppression of false detections applied to the detection of shadows and labeling.

As we will see later, the results obtained are sufficiently satisfactory for the system to work in a real case, but under certain conditions. In this chapter you will see the whole process developed through images. The final code used is in the annexes.


© CBLTIC Campus del Baix Llobregat - UPC