Utilize este identificador para referenciar este registo: http://hdl.handle.net/10400.13/877
Título: Color measurement using a smartphone applied to Madeira wines
Autor: Silva, José Carlos Vieira da
Orientador: Marques, José Carlos
Chen, Mon-Chu
Palavras-chave: Computer-vision
Madeira Wine
Color
Smartphone
3D printing
Histograms
Informatics Engineering
.
Centro de Ciências Exatas e da Engenharia
Data de Defesa: 2015
Resumo: Computer vision is a field that uses techniques to acquire, process, analyze and understand images from the real world in order to produce numeric or symbolic information in the form of decisions [1]. This project aims to use computer vision to prepare an app to analyze a Madeira Wine and characterize it (identify its variety) by its color. Dry or sweet wines, young or old wines have a specific color. It uses techniques to compare histograms in order to analyze the images taken from a test sample inside a special container designed for this purpose. The color analysis from a wine sample using an image captured by a smartphone can be difficult. Many factors affect the captured image such as, light conditions, the background of the sample container due to the many positions the photo can be taken (different to capture facing a white wall or facing the floor for example). Using new technologies such as 3D printing it was possible to create a prototype that aims to control the effect of those external factors on the captured image. The results for this experiment are good indicators for future works. Although it’s necessary to do more tests, the first tests had a success rate of 80% to 90% of correct results. This report documents the development of this project and all the techniques and steps required to execute the tests.
URI: http://hdl.handle.net/10400.13/877
Designação: Master in Informatics Engineering
Aparece nas colecções:Dissertações de Mestrado

Ficheiros deste registo:
Ficheiro Descrição TamanhoFormato 
MestradoJoseCarlos.pdf4,26 MBAdobe PDFVer/Abrir


FacebookTwitterDeliciousLinkedInDiggGoogle BookmarksMySpace
Formato BibTex MendeleyEndnote Degois 

Todos os registos no repositório estão protegidos por leis de copyright, com todos os direitos reservados.