3 Processing Remotely Sensed Data: Hardware and Software Considerations
57
3.2.6 Image and Map Cartographic Composition
Popular image processing software (e.g. Photoshop) and graphics programs (e.g.
Freehand) can be used to produce useful unrectified images and diagrams. However, if scaled images and/or thematic maps in a map projection are required, then
a full-function digital image processing system or GIS must be utilized that produce quality cartographic products that can be printed to Postscript level II output
devices.
3.2.7 Geographic Information Systems (GIS)
Information derived from remote sensor data often fulfills its promise best when
used in conjunction with other ancillary data (e.g., soils, elevation, slope, aspect,
depth-to-ground-water) stored in a GIS (Lunetta et at, 1991). Therefore, the ideal
integrated system performs both digital image processing and GIS spatial modeling and considers map data as image data (and vice-versa) (Cowen et al., 1995).
The GIS analytical capabilities are hopefully based on 'map algebra' logic that can
easily perform linear combinations of GIS operations to model the desired process. The GIS must also be able to perform raster-to-vector and vector-to-raster
conversion accurately and efficiently.
3.2.8 Utilities
The digital image processing system should have the ability to network not only
with colleagues and computer databases in the building but with those throughout
the world. Therefore, efficient transmission lines and communication software
(protocol) must be available. Much information is now routinely served on the
world-wide-web (WWW) (Ubois, 1993). Some have suggested that the Internet is
the sales channel of the future for imagery (Thorpe, 1996).
The type of data compression algorithm used to store the image data can have a
serious impact on the amount of mass storage required. The basic idea of image
compression is to remove redundancy from the image data, hopefully, without
sacrificing valuable information. This is usually done by mapping the image to a
set of coefficients. The resulting set is then quantized to a number of possible
values that are recorded by an appropriate coding method. Most commonly used
image compression methods are based on the discrete cosine transform such as the
JPEG algorithm (Joint Photographic Experts Group), on vector quantization, on
differential pulse code modulation, and on the use of image pyramids (Lammi and
Sarjakoski, 1995).
One must decide on whether to use a loss-less or lossy data compression algorithm (Sayood, 1996). When an image is compressed using lossy logic and then
uncompressed, it may appear similar to the original image but it does not contain
all of the subtle multispectral brightness value differences present in the original
(Bryan, 1995; Nelson, 1996). Imagery that has been compressed using a lossy
algorithm may be suitable as an illustrative image where cursory visual photointerpretation is all that is required. If lossy compression is absolutely necessary,
the algorithm of choice at the present time appears to be JPEG which has a com-
57
3.2.6 Image and Map Cartographic Composition
Popular image processing software (e.g. Photoshop) and graphics programs (e.g.
Freehand) can be used to produce useful unrectified images and diagrams. However, if scaled images and/or thematic maps in a map projection are required, then
a full-function digital image processing system or GIS must be utilized that produce quality cartographic products that can be printed to Postscript level II output
devices.
3.2.7 Geographic Information Systems (GIS)
Information derived from remote sensor data often fulfills its promise best when
used in conjunction with other ancillary data (e.g., soils, elevation, slope, aspect,
depth-to-ground-water) stored in a GIS (Lunetta et at, 1991). Therefore, the ideal
integrated system performs both digital image processing and GIS spatial modeling and considers map data as image data (and vice-versa) (Cowen et al., 1995).
The GIS analytical capabilities are hopefully based on 'map algebra' logic that can
easily perform linear combinations of GIS operations to model the desired process. The GIS must also be able to perform raster-to-vector and vector-to-raster
conversion accurately and efficiently.
3.2.8 Utilities
The digital image processing system should have the ability to network not only
with colleagues and computer databases in the building but with those throughout
the world. Therefore, efficient transmission lines and communication software
(protocol) must be available. Much information is now routinely served on the
world-wide-web (WWW) (Ubois, 1993). Some have suggested that the Internet is
the sales channel of the future for imagery (Thorpe, 1996).
The type of data compression algorithm used to store the image data can have a
serious impact on the amount of mass storage required. The basic idea of image
compression is to remove redundancy from the image data, hopefully, without
sacrificing valuable information. This is usually done by mapping the image to a
set of coefficients. The resulting set is then quantized to a number of possible
values that are recorded by an appropriate coding method. Most commonly used
image compression methods are based on the discrete cosine transform such as the
JPEG algorithm (Joint Photographic Experts Group), on vector quantization, on
differential pulse code modulation, and on the use of image pyramids (Lammi and
Sarjakoski, 1995).
One must decide on whether to use a loss-less or lossy data compression algorithm (Sayood, 1996). When an image is compressed using lossy logic and then
uncompressed, it may appear similar to the original image but it does not contain
all of the subtle multispectral brightness value differences present in the original
(Bryan, 1995; Nelson, 1996). Imagery that has been compressed using a lossy
algorithm may be suitable as an illustrative image where cursory visual photointerpretation is all that is required. If lossy compression is absolutely necessary,
the algorithm of choice at the present time appears to be JPEG which has a com-