A digital image processing method for urban scenes brightness assessment

Abstract

The appearance of the city is strongly linked to many aspects such as climate, culture and sociological background. In cities with a prevalence of clear skies, one of the determining aspects is the climatic condition, which defines a luminous context where all elements appear significantly bright. In order to establish the luminosity of the urban scene and its components, an analysis of pictures taken in real environments is proposed, carried out through a digital image processing software developed by the authors. The information contained in each pixel of a digital image is used to assess the distribution of light in the picture. The pictures selected for assessment represent an urban scene as seen by a standing observer. In this situation the scene can be divided in three zones with similar features: pavement, sky and façades. The developed software provides output information for the aforementioned zones and the whole image consisting on light data (average, maximum, minimum and standard deviation) convertible to luminance (cd/m²) if reference values are provided. The validity of the procedure is tested in this work and implemented in real pictures. After testing it, the software has proven to be easy to use and the results obtained in the examples have shown a trend in their values. The implementation of the software suggests that a wider number of cases of study would make possible to establish a pattern in the daylight distribution of urban scenes conceived as an addition of zones with different luminous appearance.