Application of Remote Sensing and Cellular Automata Model to Analyze and Simulate Urban Density Changes

LINARES, SANTIAGO; PICONE, NATASHA

Urban Remote Sensing

Taylor and Francis

Lugar: Florida; Año: 2018; p. 213 - 231

The used of remote sensing and GIS technology to the study of urban growth and changes is relevant, particularly to understand these processes in urban area of the developing world where there is little information about it. Furthermore, knowing which are the possible ways how urban areas are going to growth and change is important information for urban planners to take knowledge based decision to generate sustainable cities. The objective of these chapter is to extract urban densities using spectral mixture signatures methodology based on Landsat images and then to model urban sprawl and density changes in middle cities of Argentina by 2030.The methodology used is divided in two part. The first one is based on Landsat 5 TM satellite images which have world-wide coverage during a long period of time, this insures the replication process. The procedure to generate the signature of each urban density class started with a radiometric calibration and then a non-supervised classification was carried out, to identify four well defined classes. Using the real signature for each class a reconstruction was made applying spectral mixture procedure and the specific signatures class where obtain . The construction density classes obtained before are the basic data for modelling urban sprawl and density changes. We used the software LanduseSim (http://www.landusesim.com/) which is based on cellular automata model. The processes involve to run the simulations are: first, the most important driving factor for urban growth were selected (eg. primary road network, CBD or slope); then the suitability maps were build (initial transition potential maps); after, the transitions rules and neighborhood filter were set. Finally, with all these the possible urban density simulation results can be obtained. Both parts involved quantitative evaluations using confusion matrixes. Three scenarios where generated to showed the variations in urban sprawl and density changes according to different degrees of sustainable. In the first case, a predictive simulation was made in which all the rules are the ones that correlated with the process that had occurred. Then an explorative case, in which a less sustainable scenario was simulated in order to learn about the consequences of it. Finally, a normative simulation was made where a sustainable urban growth and change was modulated.