Page 21 - Digital Analysis of Remotely Sensed Imagery

xx Pr ef a c e

Parallel to the advances in sensing technology is progress in
pertinent geocomputational fields such as positioning systems and
geographic information systems. The geospatial data from these
sources not only enrich the sources of data in digital image analysis,
but also broaden the avenue to which digitally processed results are
exported. Increasingly, the final products of digital image analysis are
not an end in themselves, but a part of a much larger database. The
prerequisite for integrating these data from diverse sources is
compatibility in accuracy. This demands that results derived from
digital analysis of remotely sensed data be assessed for their thematic
accuracy.
Reliable and efficient processing of these data faces challenges
that can no longer be met by the traditional, well-established per-
pixel classifiers, owing to increased spatial heterogeneity observable
in the imagery. In response to these challenges, efforts have gone to
developing new image processing techniques, making use of additional
image elements, and incorporating nonremote sensing data into
image analysis in an attempt to improve the accuracy and reliability
of the obtained results. In the meantime, image analysis has evolved
from one-time to long-term dynamic monitoring via analysis of
multitemporal satellite data.
A new book is required to introduce these recent innovative
image classification methods designed to overcome the limitations of
per-pixel classifiers, and to capture these new trends in image analysis.
Contained in this book is a comprehensive and systematic examination
of topics in digital image analysis, ranging from data input to data
output and result presentation, under a few themes. The first is how
to generate geometrically reliable imagery (Chap. 5). The second
theme is how to produce thematically reliable maps (Chaps. 6 to 11).
The third theme of the book centers around the provision of accuracy
indicators for the results produced (Chap. 12). The last theme is about
integration of digital image analysis with pertinent geospatial techniques
such as global positioning system and geographic information system
(GIS) (Chaps. 13 and 14).
This book differs from existing books of a similar topic in three
areas. First, unlike those books written by engineers for engineering
students, this book does not lean heavily toward image processing
algorithms. Wherever necessary, mathematical formulas behind certain
processing are provided to ensure a solid theoretical understanding.
Nevertheless, the reader is left with the discretion to decide on the level
of comprehension. Those who are mathematically challenged may
wish to skip the mathematical equations. Instead, they can concentrate
on the examples provided and on the interpretation of processed
output. In this way the fundamental concepts in image analysis are not
lost. Second, the book features the geometric component of digital
image analysis, a topic that is treated rather superficially or in a
fragmented manner by authors with little background in geography,