Elizabeth A. Wentz (Arizona State Univ.)
Urban remote sensing offers the promise of frequent and detailed monitoring of urban areas to support environmental assessment and urban
planning. Nevertheless, classifying urban area images into usable information remains challenging because of the heterogeneous nature of the
urban landscape. The goal of my presentation is to share recently developed and tested methods for analyzing the urban environment through
remote sensing technology. Initially I provide an overview of topics I have studied including Phoenix area air quality, modelling residential
water use, and urban open space analysis. I will then present details of two specific projects. The first one assesses the accuracy of an expert
system image classification technique to classify land cover compared to on the ground observations and aerial photography classification. The
second detailed study examines the portability of techniques from the Phoenix study site to Delhi India. In image classification for urban
areas, that each pixel represents a mixture of classes with potentially highly variable spectral ranges. Land cover classification approaches
using ancillary data, such as knowledge based or expert systems, have shown to improve the classification accuracy in urban areas, particularly
with medium or low-resolution imagery. This is because information other than the spectral signature is used to assign pixels to classes.
Acquisition of appropriate ancillary data, however, may not always be available. I conclude the presentation with an overview of the outcomes
from an international workshop on urban remote sensing that was held at Arizona State University in April 2011.