Mapping the visual perception of travelers in a city using google street view as sensor

Have you ever been impressed by the colorfulness of some Tokyo or New york streets during the nights? A traveler wandering in a city got some inner impressions generated by some visual characteristics of the location e.g. its openness, its colorfulness, its visual complexity or even some higher visual features e.g. its architecture style). Can we create map at a city level representating such perception?
I propose to use the massive amount of geolocated images of the Google Street View service as a network of visual sensors. Using computer vision technics I analysed the 100K of images of Paris and created new kinds of perceptual maps.

Psychogeography

A large body of work emerged at the beginning of the XX century with the field of psychogeography defined by Guy Debord as “the study of the precise laws and specific effects of the geographical environment, consciously organized or not, on the emotions and behavior of individuals” in 1955. By wandering, letting onself float or drift (dériver is the French word used) each person can discover his or her own ambient unities of a specific city.


In the above image illustrating the book “Psychogeographic guide of Paris” Guy Debord, 1955 represented a map of Paris cut in different areas that are experienced by some people as distinct unties (neighbourhoods), ‘unity of atmosphere’.The mentally felt distance between these areas are visualized by spreading out the pieces of the cut up map. The red arrows indicate the most frequent used crossings between the islands of the urban archipel.

However applying quantitative methods to the analysis and characterization of the visual perception of travelers at a city-wide scale remains unclear so far.

Quantitative Approach

I propose to use the massive amount of geolocated images of the Google Street View service as a network of visual sensors. By Identifying and quantifying certain visual features on each image I am able to generate new “perceptual” city maps and conduct further analysis e.g. geographical distribution of some visual features at a district and city level or make comparison between differents cities.

digital humanities

it’s a bit like the article “Quantitative Analysis of Culture Using Millions of Digitized Books” . you change the corpus from digitized books to digitized streets. You change temporal analysis to spatial analysis and you could publish “Quantitative Analysis of Culture Using Millions of Digitized Streets”

(personal work started in 2010 but never achieved due to a lack of support)


Comments are closed.