and T.F. Wood, 
listic measures of 
od classification, 
Sensing, 58, pp. 
» sub-pixel woody 
ificial inteligance 
g and Remote 
ration of fraction 
g model”. Journal 
p. 23 — 27. 
ing and the 
national Journal 
1 Methods for 
ondon. 
ication of remote 
cience and Remote 
yy supervised 
:motely sensed 
ork approaches, 
pp. 615-628. 
IAPRS & SIS, Vol.34, Part 7, “Resource and Environmental Monitoring”, Hyderabad, India,2002 
COLOR IN PHOTOGRAMMETRIC REMOTE SENSING 
F. Leberl. R. Perko 
Institute for Computer Graphics and Vision 
Graz University of Technology, Inffeldgasse 16, A-8010 Graz, Austria 
M. Gruber 
Vexcel Imaging Austria, Miinzgrabenstr. 11, A-8010 Graz, Austria 
KEYWORDS: Color imagery, coloration, pan-sharpening, fusion, photogrammetry, remote sensing 
ABSTRACT: 
A photogrammetric process using “color” as a physical measurement of topographic objects of interest may be denoted as 
“photogrammetric remote sensing”, where the photogrammetric determination of the “where” and “how large” merges with the 
remote sensing ideology of automatically classifying the “what” of topographic objects. 
As digital cameras are beginning to enter into photogrammetric practice, a dimension of complexity is being added to 
photogrammetry that has been a main concept in remote sensing for a long time. While color aerial photography simply is being 
produced with three perfectly co-registered emulsion layers, digital images can be obtained in a variety of different approaches. 
High-resolution panchromatic imagery may be combined with lower resolution color images. Color may also get created by 
sequentially producing its components, thereby introducing a need to eliminate geometric differences. The use of color in analyzing 
terrain has been a traditional topic in remote sensing. In photogrammetry the digital sensors present a new need to understand the 
coloration alternatives. We therefore present the various alternatives to sense color at the intersection of photogrammetry and remote 
sensing. We then proceed to address the different “coloration” schemes combining high resolution black-and-white pixels with the 
lower resolution color pixels and we take a look at the differences between the color images obtained from various approaches. The 
geometric resolution of the resulting “pan-sharpened” color images is of interest. 
1. COLOR SENSING IN REMOTE SENSING AND 
PHOTOGRAMMETRY 
Color has great importance in remote sensing and is considered 
the primary object of the physical sensor measurements made 
about the terrain surface and its cover. Remote sensing has 
traditionally been interested in objects with an extended surface 
such as agricultural or other geo-science entities. In 
photogrammetry, the focus is on geometric accuracy and detail 
as a feature of point- and line-like objects, which is reflected in 
the panchromatic information of an image. As a result, color 
has been traditionally of lesser importance. 
However, color is of growing interest in photogrammetry as 
well, as automation of the image interpretation function 
evolves. In many regions of the world, photogrammetry has 
begun to employ color film as standard source material, as 
reflected by the fact that film manufacturers over the last 
decennium report a doubling of the sales of color film at the 
expense of the use of black & white film. At the same time 
photogrammetric film scanners have become challenged with 
the need to resolve the subtleties in color of the film sources. In 
the transition from analog film imagery to digital sensing, 
however, color is becoming a growing challenge, since the 
focus on geometric accuracy and detail continues to exist, but 
color information is also of interest. How can high geometric 
accuracy and resolution be combined with color information in 
an affordable manner? 
We discuss in this contribution various technologies to create 
digital color images. Most prominently are procedures where a 
high-resolution panchromatic image is being “colorized” by 
lower resolution color data. A factor 3 to 5 between the 
geometric resolutions of panchromatic and color sensing may 
be acceptable. 
2. SENSING COLOR AND COLORATION SCHEMES 
2.1 Color Sensing 
Various approaches to color sensing have been proposed. 
Recent satellite missions and aerial digital camera solutions 
have demonstrated that color can be obtained by: 
Y Simultaneous collection of multiple colors per pixels using 
linear push-broom detector arrays, all at one geometric 
resolution (as in the Landsat-approach and in the Leica 
ADSA0 aerial camera, Leica, 2002); 
Y Simultaneous collection of panchromatic images, each 
with a color filter, as presented by Positive Systems (2001) 
¥ Simultaneous collection of panchromatic and color data 
with linear arrays CCDs, but at different geometric 
resolutions (the standard satellite approach in the IRS-, 
Ikonos-, Digital-Globe- and other systems); 
¥ High geometric resolution panchromatic area CCD sensing 
in combination with lower resolution area CCD color 
sensing using the Bayer pattern (no commercial solution 
currently being offered); 
¥ High geometric panchromatic area CCD sensing in 
combination with lower resolution area CCD sensors using 
different filters (the Z/I Inc DMC aerial camera, Z/I Inc, 
2002); 
Y Color area CCD using the Bayer pattern, and no separate 
higher resolution panchromatic sensor (the aerial camera 
solutions by Emerge, 2002 and by Enerquest, 2002). 
The new aerial digital cameras by Leica and Z/I are proposed as 
remote sensing sensors separate from the traditional aerial film 
cameras. As a result these cameras offer 4 spectral bands, 
including an infrared channel. They are considered to address