Global urban extent from Landsat uising multi-level object-based texture features

Panshi Wang, University of Maryland, pswang@umd.edu (Presenter)
Chengquan Huang, University of Maryland, cqhuang@umd.edu
James Tilton, NASA GSFC, james.c.tilton@nasa.gov
Bin Tan, NASA GSFC, bin.tan@nasa.gov
Eric Brown de Colstoun, NASA GSFC, eric.c.browndecolsto@nasa.gov (Presenter)

More than half of the world’s population lives in the urban areas and urban population is still growing rapidly at an unprecedented rate. It is important to monitor, understand, and model the growth of urban land and population. One of the prerequisites of further study into urbanization is a detailed urban extent map. At global scale, Landsat is an ideal data source for urban extent mapping. However, it is difficult to produce such a map using spectral data from Landsat, because of spectral similarity between some urban and non-urban objects and spectral variability within the urban and non-urban class. Here we present an approach for mapping urban extent at global scale using multi-level object-based texture features. Preprocessed Global Land Survey (GLS) 2010 Landsat surface reflectance images were segmented using a hierarchical segmentation software package and texture features were extracted at multiple levels of the segmentation hierarchy. Random forest classifiers for different continents were trained to classify objects into urban/nonurban categories. This method is being applied to the entire GLS 2010 Landsat data collection to produce a fine resolution global urban extent map.

Presentation Type:  Poster

Session:  Theme 4: Human influence on global ecosystems   (Mon 4:30 PM)

Associated Project(s): 

• Brown de Colstoun, Eric: Using Landsat Global Land Survey Data to Measure and Monitor Worldwide Urbanization ...details
• Huang, Cheng: Using Landsat Global Land Survey Data to Measure and Monitor Worldwide Urbanization ...details

Poster Location ID: 36