COMPUTER PROCESSING OF REMOTELY-SENSED IMAGES
THIRD EDITION
Paul M. Mather
Chichester: John Wiley &
Sons Ltd.
Publication date: 28 April 2004
Exercises contributed by
Dr Magaly Koch,
Center for Remote Sensing,
Boston University, Boston, MA
The CD accompanying Computer Processing of Remotely-Sensed
Images contains four advanced exercises, contributed by Dr Magaly Koch,
Center for Remote Sensing, Boston University, Boston, MA. These exercises
can be found in the EXAMPLES folder. Each exercise has:
-
a PDF file containing the text of the exercise
-
a DATA folder containing the full datasets used in the exercise,
and
-
an IMAGES folder, containing the illustrations as separate
JPEG/TIF files (for easy use on data projectors).
The four exercises are:
1. Geological Reconnaissance
of the Red Sea Coast in Egypt.
The aims of this exercise are
to illustrate:
-
Generating true and false colour composites;
-
Image filtering for sharpening and smoothing;
-
Using the decorrelation stretch transform and the principal
component analysis (PCA) for lithological mapping.
using TM and ASTER imagery.
2. Agricultural Land Use Change
in Los Monegros, a Semiarid Mediterranean Environment.
The aims of this exercise are:
-
To enhance the TM images dating from 1984 and 1997 for visual
analysis;
-
To convert raw digital numbers (DNs) into reflectance values
for spectral curve extraction;
-
To perform a tasselled cap transform using two methods: (a)
based on pre-determined, and (b) user-determined transformation coefficients;
-
To compare the tasselled cap results obtained from two images
13 years apart and evaluate changes observed in agricultural land use.
Two Landsat TM images (for 1984 and 1997) are provided.
3. Monitoring Forest Clearing
in Holmul, an Archaeological Maya Site in Guatemala.
The aims of this exercise are:
-
To match the contrast of two images
of the same area but of different dates;
-
To generate forest change maps using
change detection techniques;
-
To categorize and analyse the type
of change due to landscape variability and human induced forest clearing
(by using classification methods).
Landsat TM and ETM+ images of
the area are provided.
4. Mapping Wetland Features
in La Mancha, Central Spain.
The aims of this exercise are:
-
To visually compare multi-source and
multi-date images;
-
To detect changes in vegetation distribution
using the NDVI;
-
To classify wetland features such as
soils types (IsoData, Spectral Angle Mapper);
-
To identify spectral characteristics
of wetland components (Spectral Curve analysis)
ASTER, DAIS and ETM+ images are provided.
Return to main page
May 2004