|
A completely different line of research from the use of impervious
surface coefficients is our work on modeling impervious
surface using “subpixel” analysis
of Landsat data. The ERDAS Imagine SubPixel ClassifierTM, engineered
by Applied Analysis Inc., is a supervised classifier that enables
the detection of materials of interest (MOIs) as whole or fractional
pixel composition, with a minimum detectable threshold of 20
percent and in increments of 10 percent (i.e., 20-30%, 30-40%, …90-100%).
Exploratory work conducted in Massachusetts showed promising results
on the accuracy of the subpixel estimates. An impervious surface
layer for the entire state of Connecticut has been produced from
spring 1995 Landsat TM. A close-up of this data layer for the town
of Simsbury CT is shown here. During 2004, CLEAR researchers will
be using this technique to chart the changes in IS coverage in
the Connecticut and New York areas of the Long Island Sound drainage
basin, for the period 1985 to 2002. For more information please
see our reference papers.
|
