Advanced search
Start date

Imagens de grande alcance dinâmico aplicadas ao mapeamento da distribuição de luminâncias da abóbada celeste

Dennis Flores de Souza
Total Authors: 1
Document type: Doctoral Thesis
Institution: Universidade Estadual de Campinas. Faculdade de Engenharia Civil, Arquitetura e Urbanismo
Defense date:
Examining board members:
Daniel de Carvalho Moreira; André Santanchè; Ricardo Carvalho Cabús; Fernando Oscar Ruttkay Pereira
Advisor: Paulo Sergio Scarazzato

Lately, the high dynamic range images (HDR) have experienced a significant growth in their usage in lighting studies, due to their capacity to store data of luminance distribution in a scene. Various studies have attested, for instance, the possibilities of using digital images in the register of daylighting, since the features of HDR images could enhance the results. Among different applications, the record of the light on the sky vault is one that can benefit most from HDR techniques, because this procedure is simpler than those performed by luminance meters or sky scanners measurements. Besides, the identification and classification of sky types are still done mostly by subjective methods. This can be explained by the unavailability of a methodology able to compare HDR images with mathematical models, although there are unidimensional methods that focus on one or another aspect of digital images. This research aimed at the development of a multidimensional method of identification, classification and extraction of daylight data from HDR images of the sky vault. The images registered by the camera were calibrated using available methods to establish the reliability of the analysis and interpretation of data. They were then obtained on a site with minimal obstruction to the vision of the sky vault. The multidimensional analysis method was developed in conjunction with a routine in MATLAB, which served the purpose of verifying its feasibility and accuracy. The extracted data were tested in Flash platform using ActionScript 3 language to briefly demonstrate the usage possibilities. This method relies on a classification system based on the relevance of the features identified in the image, such as cloud covering and luminance distribution, to choose the most appropriate sky type according to ISO 15469:2004 (e) / CIE S 011/E:2003 Standard. The results demonstrate the feasibility of this method in choosing the most relevant sky type according to the data extracted from the HDR image. The proposition of this multidimensional analysis method may contribute to the creation of a classification system and a digital database useful for future simulation software, providing input data from measurements of a physical reality, easily recorded with accuracy and confidence by photographic images. (AU)

FAPESP's process: 11/10314-7 - Use of High Dynamic Range images in mapping sky vault luminance distribution
Grantee:Dennis Flores de Souza
Support type: Scholarships in Brazil - Doctorate