Image Processing Reference
In-Depth Information
urban areas have been conducted (Oke 1987 ; Voogt and Oke 2003 ). From
such studies, urban heat island and any amelioration methods could be studied
and monitored continuously and linked to land use patterns. For example, city
planners would like to know the effect of temperature and air pollution from
the layout of parks, water bodies, industrial and commercial areas. The tech-
nology of using airborne and spaceborne thermal infrared sensors to measure
surface temperature of urban areas has been available since the 1970s. Such
studies entail the conversion of thermal infrared data into surface temperature.
A major obstacle encountered with the use of satellite imagery in environ-
mental studies has been the relatively coarse spatial resolution (see Chapter
5). However, as discussed in Chapter 7, the spatial resolution of satellite
imagery has increasingly become finer over the last three decades. Starting
from Landsat Multispectral Scanner (MSS) with a resolution of 79 m per
pixel in the early 1970s, presently QuickBird satellite has a spatial resolution
of 0.61 m. In parallel with the improved spatial resolution is the improved
spectral resolution. Landsat MSS has five bands in the visible and near-
infrared regions, but presently the Hyperion satellite has 220 bands between
0.4 and 2.5 µm. However, the improvement in the spatial resolution of ther-
mal bands do not parallel the advances in the reflective bands and has been
relatively lower than other regions - due to different technologies used for the
visible, near-infrared, shortwave and thermal infrared regions. At the moment,
the highest spatial resolution for satellite thermal infrared imagery is 60 m
per pixel from Landsat Enhanced Thematic Mapper Plus (ETM+). The
Advanced Very High Resolution Radiometer (AVHRR) has a spatial resolu-
tion of 1.1 km per pixel (at nadir) and has been utilized in several large-scale
urban analyses (Roth et al. 1989 ; Owen et al. 1998 ).
14.3
Urban Vegetation Mapping
Although vegetation mapping is one of the primary uses of optical remote sensing
data, relatively little work has focused on urban vegetation. In part, this is a result
of the greater diversity of natural vegetation mapping applications but it is also a
result of the characteristic scales of urban vegetation relative to the spatial resolu-
tion of operational sensors. Recent comparative analysis of IKONOS 1 m resolu-
tion imagery in several urban areas worldwide indicates that the characteristic scale
of individual features in urban mosaics is between 10 and 20 m (Small 2004 ). Until
relatively recently, SPOT and Landsat with 20 and 30 m multispectral spatial reso-
lutions, respectively, served as the primary tools for fine scale vegetation mapping.
As a result of the low resolutions, SPOT and Landsat image urban areas primarily
as spectrally mixed pixels. Although vegetation has a strong, and measurable influ-
ence on the mixed spectra that these sensors image, it is not amenable to the thematic
 
Search WWH ::




Custom Search