Use of archival Landsat imagery to monitor urban spatial growth - Urban Remote Sensing: Monitoring, Synthesis and Modeling in the Urban Environment - page 16

FIGURE 2.2 Working procedural route adopted, with several major components: primary and secondary data acquisition,
image preprocessing, image classification and thematic accuracy assessment, and change detection and analysis.

growing in the fashion of Los Angeles. Atlanta is a good example
of such a city because it exhibits all the characteristics of the
postmodern city described above.
Research on Atlanta's internal structure has led to the for-
mulation of the urban realms model to depict the multi-nuclei
nature of the city in contrast to the conventional single-core
urban form (Hartshorn and Muller, 1989). Thus, Atlanta is an
ideal city to study postmodern urban dynamics and environ-
mental consequences of accelerating urban growth. During the
past 15 years, the author has been involved various research
projects aiming to understand the dynamics of change in Atlanta
by using remote sensing. The case study examines urban spa-
tial growth and landscape change along the outskirts of the
Atlanta metropolitan area by using a time series of Landsat data
covering a period of nearly four decades. The research methodol-
ogy identified here includes several major components: primary
and secondary data acquisition, image processing of remote
sensor data, change detection, and interpretation and analysis
(Fig. 2.2).

at 6 - 8-year intervals, beginning in 1973 when Landsat MSS
data became available. This series comprises 13 predominantly
cloud-free images acquired by the three Landsat sensors for
Atlanta during 1973 - 2007 (Table 2.1). Note that because of the
permanent SLC failure in the ETM + instrument on 31 May 2003,
two cloud-free TM scenes were acquired for 2007. Most of the
scenes were acquired during the spring or earlier summer seasons,
when vegetation is in the stage of vigorous growth. The 1998 and
1999 scenes are the two exceptions. The 1998 TM scene, acquired
in the winter season, was used to improve vegetation mapping.
The ETM + scenes were acquired in late summer because they
are the only scenes free from clouds available between April
and September 1999. Because of good image quality, the 1988
MSS scene was mainly used as the reference image for relative
radiometric normalization of other MSS images.
To facilitate satellite image-based change mapping, a vari-
ety of ancillary data have been collected, including digital
images of Advanced Thermal and Land Applications Sensor
(ATLAS) acquired on 11 May 1997, contact prints of aerial
photographs for 1986 - 1988, 1993 USGS digital orthophotos,
the 1988 - 1990 land-cover classification from Georgia Depart-
ment of Natural Resources, and high-resolution satellite imagery
from Google Earth. In addition, a GPS-guided field survey was
conducted to help establish the relationship between image sig-
nals and ground conditions. Fieldwork also helped obtain first
hand information about suburban sprawl throughout the study
area, which can be useful for understanding the dynamics of
change.

2.3.2 Data acquisition and land
classification scheme

A time series of Landsat images was used as the primary data to
detect and measure the spatio-temporal urban growth pattern