The new GUIs allow more efficient and accurate human-computer 
interaction. The history of GUIs is chronicled by Seymore (1989) 
who describes their evolution from the first Xerox ‘Star 
workstation’ to Open Software Foundation’s ‘Motif’. GUIs use 
window-icon-mouse-pointer (WIMP) theory, where icons of 
familiar objects represent operational functions (e.g. a magnifying 
glass is used to zoom in). The advantage of such systems lies in 
the ease with which a user can initiate commands and manipulate 
the operating environment. In addition, ‘multi-tasking’ is 
facilitated, allowing the user to operate within two or more 
applications simultaneously. 
Even the new WIMP user interfaces, however, suffer from the 
problems of overfilling the screen with icons, the creation of very 
long menus, the use of inappropriate metaphors, and the lack of 
‘activity indicators’ on the status of an operation (Raper and 
Bundock, 1991). Also, the interface must progress and allow the 
user to manipulate objects that are meaningful in terms of the 
application, such as "sub-divide a parcel" instead of "split a 
polygon". 
There are several new WIMP based GUIs in the marketplace 
including ERDAS 8.0, Arc-Info 6.0, Intergraph, and ER Mapper 
3.0. Figure 1 depicts the ER Mapper 3.0 GUI for analyzing 
standard 1-8 band digital remote sensor data. It uses the WIMP 
point and click technology plus ‘Macintosh’ like "pull down 
menus” (ER Mapper, 1992). The GUI of the Spectral Image 
Processing System (SIPS) developed by the Center for the Study 
of Earth from Space (CSES) at the University of Colorado, 
Boulder is shown in Figure 2. This unique interface is designed 
to analyze "hyper-spectral" remote sensor data composed of up to 
192 spectral bands. It is anticipated that such data will be 
common place in the Earth Observation System (EOS) era of the 
21st century and will require such a user interface (Wickland, 
1991). SIPS uses menus, buttons, and slider-bars along with a 
mouse and keyboard input to create a user-friendly interface 
(Kruse et al., 1992). Basically, the user can move the cursor to 
any x, y coordinate in the scene and plot on the bottom graph the 
complete spectral signatures (e.g. .4-2.5 pum) for that pixel. This 
signature can be compared to a library of ‘saved’ spectra in an 
adjacent graph. Therefore, this system represents the first truly 
hyper-spectral, graphical user interface. 
CALIBRATED GRAY-SCALE STEP-WEDGES 
COLOR-WHEELS, AND HISTOGRAM INFORMATION 
One of the simplest and most useful tools for interpreting digitally 
processed images or image maps is the presence of calibrated 
gray-scale step-wedges for black and white images or a color 
wheel or color bar for color images. The basic function of such 
annotation on the screen or hard copy is to ensure correct 
exposure and correct visual presentation, and interactive color 
selection. 
The concept of incorporating gray-scale step-wedges can be found 
in one of the earliest image processing systems -- the Video 
Information Communication and Retrieval (VICAR) system 
(Castleman, 1979). Figure 3 depicts a typical VICAR black and 
white ‘mask’ composed of systematic gray-scale step-wedges. The 
wedge is applied to all black and white images on the CRT and 
hard copy. If the exposure is correct all shades of gray level will 
be interpretable. However, if the image is over or underexposed, 
only a portion of the wedge will appear correctly and the user 
knows that some adjustments are required. Also present is the 
histogram which can be useful for communicating ‘before’ and 
‘after’ image enhancement operations. 
When working with color images and image maps, there are 
standardized color specifications which can be used to depict the 
126 
exact nature of the colors used. For example, one digital image 
processing system allows the user to easily switch between any of 
the following color specification systems (Duotone, Indexed, RGB, 
CMYK, HSL, HSB, Multichannel) which may be displayed in a 
color wheel diagram (Adobe Systems Inc., 1991). However, the 
‘best’ color wheel legend has yet to be determined. 
IMAGE AND/OR MAP COORDINATES & ANNOTATION 
Maps and image maps must be geo-referenced to a standard 
coordinate system and map projection to be truly useful. It is 
quite common for a final image map to be composed of data from 
various remote sensing systems (e.g. SPOT, Landsat TM merge) 
or for a final GIS map to be the product of data from very diverse 
source materials. Therefore, it is necessary for the data to be 
transformed to a single coordinate system, most commonly the 
Universal Transverse Mercator (UTM) or the State Plane 
Coordinate System (in the U.S.). The process involves the 
application of one of several rectification algorithms to transform 
the image to standardized planimetric basemap (Jensen, 1986). 
Once rectified, the image file contains both image coordinates 
(row and column) and map coordinates (e.g. UTM), and can be 
merged with other similarly geo-referenced GIS data. Figure 3 
has an ‘image space’ grid superimposed on it. In addition, users 
of remotely sensed or GIS data must be provided with maps 
containing accurate map graticules whenever possible. 
Another very important annotation which is often overlooked is 
the ‘sector location diagram’. When a map sheet being displayed 
(e.g. sheet 3E) is but one of several other sheets in a region, a 
location diagram will allow the user to correctly identify which 
map or map image is currently being studied. Additional research 
on the design of these diagrams is required. 
LEGENDS FOR STATIC AND DYNAMIC MAPS 
Cartographers and remote sensing professionals have become quite 
adept at creating ‘static’ thematic map legends which depict the 
condition of the earth at a static instant in time. Ideally, good 
cartographic practice is followed using a relatively small number 
of classes (e.g. <16), logical class intervals, and cartographically 
correct use of colors or shades of gray. Plumb (1988) suggests 
that the class intervals for static maps be selected using a 
‘goodness of fit’ index which will more accurately depict the data. 
Many new products are based on the analysis of multiple dates of 
imagery. These ‘dynamic’ maps are very powerful but require 
new legends in order to communicate effectively. The ubiquitous 
change detection map is a good example of a ‘dynamic’ map. 
Monmonier (1992) suggests that an animated sequence of maps 
and their related statistical graphics could be used to study these 
"spatial-temporal" data. These methods would be useful for maps 
shown on CRTs, however, for ‘hard-copy’ output there is a need 
for more carefully designed dynamic map legends which depict 
change. New legends are required which depict the "from-to" 
information more efficiently and accurately. 
SPATIAL (GEOMETRIC AND THEMATIC) 
RELIABILITY DIAGRAMS 
Cartographers often use manually drawn ‘reliability diagrams’ to 
communicate the geometric and thematic reliability of their 
products and the source materials used (Robinson et al., 1984). 
This tradition should be continued in products derived from 
remote sensing and GIS technologies. Information on source 
material used and the accuracy of the material should be 
represented by digital geometric and thematic reliability diagrams 
(Lunetta et al., 1991). 
pr 
It 
en 
"I;