Procedural Urban Modelling
Whilst being far from a solved problem, techniques for algorithmically generating convincing computer graphics models of landscapes, sea scapes, cloudscapes, forests etc. are reasonably well known. However, convincing urban environment generation is still a hot topic.
Space Syntax methods might offer useful metric in evaluating generated spaces as functional urban environments. Indeed, in a recent Guardian article, Professor Bill Hillier of UCL's Space Syntax Laboratory is quoted as saying "I wouldn't design a city … I'd grow one."
Over at Northwestern University, there's a Procedural Modelling of Cities project to do just that, with some interesting results neatly visualised using Sim City 2000. Also of note is CityBuilder, an open source project to generate urban forms (links via Mike Davis's lightcycle).
At a planning/expansion level, there's a lot of research into how urban forms take shape across the landscape. In CASA (the Centre for Advanced Spatial Analysis) at UCL, Mike Batty and his team have looked long and hard at the fractal qualities of city growth, resulting in the publication of their 1994 book Fractal Cities and the forthcoming Cities and Complexity (Understanding Cities with Cellular Automata, Agent-Based Models, and Fractals). Also of interest is this online book on Fractals and Fractal Architecture hosted at Tuwien which includes a chapter on City Planning.
I can't write this many links without referring to a relevant project built with Processing, and Jared Tarbell's stunning Substrate project is just the thing. Hugely reminiscent of urban landscapes, and lovingly rendered (in a palette Jared concedes is borrowed from Jackson Pollock). I'd love to see a 3D environment generated in this way!
Space Syntax methods might offer useful metric in evaluating generated spaces as functional urban environments. Indeed, in a recent Guardian article, Professor Bill Hillier of UCL's Space Syntax Laboratory is quoted as saying "I wouldn't design a city … I'd grow one."
Over at Northwestern University, there's a Procedural Modelling of Cities project to do just that, with some interesting results neatly visualised using Sim City 2000. Also of note is CityBuilder, an open source project to generate urban forms (links via Mike Davis's lightcycle).
At a planning/expansion level, there's a lot of research into how urban forms take shape across the landscape. In CASA (the Centre for Advanced Spatial Analysis) at UCL, Mike Batty and his team have looked long and hard at the fractal qualities of city growth, resulting in the publication of their 1994 book Fractal Cities and the forthcoming Cities and Complexity (Understanding Cities with Cellular Automata, Agent-Based Models, and Fractals). Also of interest is this online book on Fractals and Fractal Architecture hosted at Tuwien which includes a chapter on City Planning.
I can't write this many links without referring to a relevant project built with Processing, and Jared Tarbell's stunning Substrate project is just the thing. Hugely reminiscent of urban landscapes, and lovingly rendered (in a palette Jared concedes is borrowed from Jackson Pollock). I'd love to see a 3D environment generated in this way!
+44 20 7679 2000 - Copyright © 2006 UCL
1 Comments:
Jared Tarbell's Substrate project is indeed absolutely stunning. He mentions the fractals pattern of his lines. What is remarkable is that the
"axial lines" of space syntax -- long straight lines drawn through actual urban space -- also follow a fractal pattern (http://arxiv.org/abs/cond-mat/0305164).
Space syntax researchers are usually more interested in graph network properties of these lines, and it would be extremely interesting to conduct a "space syntax" analysis of these substrate lines.
By Alasdair, at 3:59 PM
Post a Comment
<< Home