A method for the visual analysis of the streetscape

This paper outlines an interdisciplinary approach, utilising architectural knowledge and computer imaging, to develop an analytical tool that describes the physical characteristics of a streetscape. Techniques for connecting the urban texture at the scale of the individual, with the urban character of precincts within the city are difficult to find. One theory that does appear to be able to make such a connection is space syntax. By considering the open spaces generated by the existence of an interdependent built boundary extending in scale from the individual house through to the streets that form cities, space syntax attempts to explain human behaviour as it occurs in those spaces. Describing the visual character of a streetscape through analysis of its spatial configuration might then provide an objective measure within a planning field that is dominated by qualitative methods. A method of analysis is developed using an algorithm, based on the Hough transform, which provides a global measure of the geometry within the convex space of the streetscape. This image processing technique segments an image at a fine scale on the basis of discontinuity and similarity, allowing the edges that define features within the built surface to be detected. The density of edges within the textured surface then provides a measure of how visual detail is distributed throughout an image at different scales, a concept that is closely related to fractal geometry. The approach is conceptually related to the theory of description retrieval, facade configuration and the facade isovist discussed by Hillier (Hillier 1996:p238; Hillier 2003). The streetscape is considered as a convex space where the visual field is constructed of elements that can be studied as an objective reality. The interrelationship of the elements or as Hillier describes, the way they are synchronised, might then provide an insight into how the street is understood at an experiential level.