SEPTEMBER 2001
8
Chromo-Stereoscopy: 3D Stereo
-Use the 3D ChromaDepth Viewing Spectacle Included in
A very modern trend in photogrammetry is the automated production of DEM data from
overlapping stereo-pairs of remotely sensed image data, followed by the generation of
an orthoimage or an orthoimage mosaic. These operations may be carried out as a
single integrated process using suitable imagery acquired from a spaceborne platform.
By Gordon Petrie, Thierry Toutin, Hasin Rammali & Christophe Lanchon
Examples are (i) the overlapping images
acquired cross-track from the SPOT or IRS
satellites on two different orbits; and (ii)
those acquired along-track from a single
pass of the MOMS, IKONOS or EROS-A1
satellites. (iii) A further possibility is to use
either stereo- or interferometric radar (SAR)
imagery to generate the DEM and orthoimage
data. Even more common is the use of
imagery taken from an airborne platform.
Within this latter context, the process of
DEM and orthoimage production is commonly
carried out by mainstream photogrammetric
service providers using stereopairs
of aerial frame photographs, whether
collected in analogue or digital form. Now
the method is being implemented with the
digital imagery acquired by the new generation
of airborne pushbroom scanners, e.g.
the DLR HRSC-A and LH Systems ADS40
imagers. The great attraction of the process
from the photogrammetric production point
of view is the highly automated procedure
that can be implemented with a minimum
of human input after the identification and
measurement of the ground control points
(GCPs). While this is extremely satisfactory
from the point of view of the photogrammetric
service providers, the situation may
be viewed in a quite different way by the
actual clients or end-users.
Need for 3D Stereo-Viewing
On the one hand, the client or end-user
does receive an ortho-rectified image having
a good map geometry, together with
the corresponding DEM data. On the other
hand, delivery of an orthoimage throws the
task of carrying out the interpretation and
extraction of the point, line and area data
on to the end-user. Previously, when the
traditional type of vector line map with contours
was being supplied to the client, the
detection, identification and extraction of
the topographic features was carried out as
a prior operation by the photogrammetrist
and the specialist image-interpreter using
3D stereo-viewing. Nowadays the end-user
may have to execute this task himself using
the orthoimage without the benefits of having
a 3D stereo-model available for the purpose.
Besides which, 3D stereo-viewing is
also essential for the execution of the quality
control that is a necessity with DEMs
produced using automated image matching
methods. This quality control operation
takes the form of editing and removing mismatches
and ensuring that the DEM data
and the lakes, rivers, mountain crests, passes
and ridge lines occurring in the area of
interest are all consistent with one another
and make sense from the topographic point
of view. Quite apart from feature extraction
and DEM quality control, the simple stereoviewing
of an area in 3D is often highly
beneficial to an understanding of the terrain
and its landforms and hydrology by field
scientists, engineers, planners and military
personnel. However there is no possibility
of them doing so using only the DEM and
orthoimage data.
3D Chromo-Stereoscopic Images
One solution to this particular problem is
to generate a 3D chromo-stereoscopic
image from the orthoimage and DEM data.
This method of 3D stereo-viewing has been
developed originally by Steenblink in the
late 1980s. During the 1990s, it then
became available from the Chromatek
Technologies company in the U.S.A. as a
commercial product, called ChromaDepth.
This has been further developed and
refined specifically for use with remote
sensing data by Dr. Toutin of the Canada
Centre for Remote Sensing (CCRS), which
forms part of Natural Resources Canada
(NRCan). Details of the underlying theory
and the actual process of generating chromo-stereoscopic
images are given in two of
his papers (Toutin 1995, 1997).
(a) The refraction of white light through a glass prism; (b) The superchromatic double prism; (c) ChromaDepth
viewing spectacles using a pair of double prisms (drawn by Mike Shand)
Chromo-Stereoscopy: 3D Stereo
SEPTEMBER 2001Latest News? Visit www.geoinformatics.com
The Basis of Chromo-Stereoscopy
The basic process is relatively straightforward
- at least in principle! The individual
elevation values from the DEM are encoded
systematically into the orthoimage on a
pixel-by-pixel basis in the form of different
colours. This allows an observer to exploit
the well-known phenomenon that objects
lying in the same plane - e.g. on the screen
of a display monitor or on a hard-copy print
- but in a different colour, appear to an
observer to lie at different depths. Thus a
red object appears to the observer to lie
above the level of a blue object. While a
yellow object appears to lie at an intermediate
level or depth between the red and
blue objects. To further enhance this effect,
use is made of suitably designed spectacles
that utilize the differential refraction of light
that is produced by prisms. These prisms
cause differential refraction of the various
components of white light. These spread
out to produce the classic "rainbow" effect
in which the light occurring at different
wavelengths is spread out and seen as contiguous
bands of different colours. Utilizing
this effect, the ChromaDepth spectacles that
appearing to lie nearer (i.e. higher) to the
observer, while those encoded in blue (with
the lowest elevation values) appear to lie
further away. Terrain objects having intermediate
elevation values between these two
extremes (which have been encoded in
orange, yellow, green, etc.) will appear at
the appropriate intermediate depths or levels
between these highest and lowest elevation
values.
Examples
The chromo-stereoscopic technique can be
used with a wide range of imagery
acquired from a variety of spaceborne and
airborne platforms; using different types of
imaging sensors; and with images having
widely differing ground resolution values.
Three examples covering quite different
types of terrain at a variety of image scales
and ground resolutions are given below.
-Montreal, Quebec, Canada
This image is a space radar orthoimage
mosaic covering the city of Montreal and
are worn by the users are equipped with
pairs of specially manufactured prisms.
These are used to enhance the apparent
depth of the differently coloured parts of
the image for each eye and so produce a
3D stereoscopic image.
ChromaDepth 3D Stereo-Viewing
The ChromaDepth spectacles utilize a pair
of very thin prisms made of a completely
transparent plastic material. In practice,
these act like thick glass prisms. To achieve
this effect, each prism comprises two component
thin prisms cemented together - the
one using a low dispersion material, the
other a high dispersion material. The combined
effect of the pair of these double
prisms when used in the spectacles is to
decode the colour-encoded orthoimage as
seen by each eye. This allows the observer
to experience an increased perception of
depth corresponding to the elevation values
that are present in the DEM. Thus a strong
stereoscopic impression of the terrain is
obtained with those objects having the
highest elevation values (encoded in red)
with Orthoimages & DEM Data
this Magazine!with
Orthoimages & DEM Data
Figure 1: Montreal, Quebec Canada (copyright: Canadian space Agency)
9
SEPTEMBER 2001
10
its surroundings. It has been produced by
CCRS/NRCan from two Radarsat-1 SAR fine
mode images having a 6m ground pixel
size. The chromo-stereoscopic mosaic has
been generated using a two-stage process.
(i) First of all, each of the radar images has
been fitted to GCPs extracted from the
standard 1:50,000 scale topographic maps
of the area produced originally by the
Canada Centre for Topographic Information
(CCTI/NRCan). This was followed by the
actual ortho-rectification process using the
DEM data generated by CCRS/NRCan from
the digitized 10m contour lines taken from
the topographic maps. The merging of the
two individual ortho-rectified images produced
the required radar orthoimage mosaic.
(ii) The second stage in the process was
to integrate this radar orthoimage data
with the DEM data to generate the final
chromo-stereoscopic image. This was carried
out via a radiometric process involving
the use of the IHS (intensity-hue-saturation)
transformation to define and generate
the appropriate colours for each individual
pixel in the radar orthoimage mosaic. This
was done on the basis of the corresponding
elevation values derived from the DEM.
The final 3D chromo-stereoscopic mosaic
was published originally by CCRS/NRCan in
the form of a poster, with the actual image
measuring 53 x 35cm. However it has of
course been reduced greatly in size and
resolution for inclusion in GeoInformatics.
The image shows the St. Lawrence River
crossing the area from the lower left corner
to the top right corner of the image. The
Ottawa River flows into it as a tributary
starting midway up the left side of the
image. The city of Montreal occupies the
central part of the image with the Mont
Royal featuring prominently above the St.
Lawrence River.
-Misratah, Tripolitania, Libya
The second example forms part of the
coastal area around the city of Misratah
located 150km east of Tripoli, the capital of
Libya. The area has been used for the production
of a series of experimental space
image maps intended to explore the cartographic
design aspects of this type of map
at 1:50,000 and 1:250,000 scales. The
underlying objective behind this project was
to see if this type of image map would be a
suitable replacement for (i) the existing
series of line maps @ 1:50,000 scale that
had been produced for the more developed
areas of Libya; and (ii) for the series of
space image mosaics @ 1:250,000 scale that
provide complete national coverage of the
country. Both series are now completely outof-date.
As part of this experimental work
carried out at the University of Glasgow,
chromo-stereoscopic images of the test area
were produced with the help of Dr. Toutin.
On this occasion, the data that was used
was SPOT Pan monoscopic imagery with a
10m ground pixel size. As with the previous
Canadian example, the required DEM was
formed from the digitized contours taken
from the existing 1:50,000 scale topographic
map. The procedure used for the generation
of both the SPOT orthoimage and the final
chromo-stereoscopic image was basically the
same as that used with the Montreal example.
The final image shows clearly the
strongly incised wadis (valleys) cut into the
semi-arid plateau and the slopes running
down to the flat coastal plain where
Misratah city and the developed areas of
agricultural land are located.
-Sophia Antipolis, Southern France
The third example shows part of the "science
city" of Sophia Antipolis which lies a short
distance inland from the Cote D'Azur and just
north-west of the coastal city of Nice in the
south of France. This image has been generated
by the ISTAR company, whose headquarters
and main processing facilities are actually
located in Sophia Antipolis. The 3D
chromo-stereoscopic image has been produced
from overlapping high-resolution
images (with 0.25 ground pixel size) of the
area acquired by DLR's HRSC-A airborne
pushbroom scanner. This features a three-line
pushbroom arrangement with the overlapping
images generated from the forward, nadir
and backward pointing linear arrays. This
allows stereo-coverage to be collected alongtrack
during a single flight. The HRSC-A
imager has its own integrated DGPS/INS system
which provides continuous positional
and attitude information for each line of the
image data. In this particular case, first of all,
the elevation data was generated from the
overlapping HRSC-A stereo-imagery using
automatic image matching techniques in conjunction
with the in-flight position and attitude
data given by the DGPS/INS unit. This
took the form of a digital surface model
(DSM) including both building and tree
canopy elevation data. Thereafter the orthoimage
was produced using the DSM data as
the basis for the ortho-rectification. This was
Figure 2: Misratah, Tripolitania, Libya (copyright: University of Glasgow)
SEPTEMBER 2001Latest News? Visit www.geoinformatics.com
followed by the production of the final chromo-stereoscopic
image (called a
"Hypsoimage" by ISTAR) using the same
technique as described for the two previous
examples. Once again, the pronounced depth
of the 3D image at the large scale of the
image shows clearly the rolling wooded character
of the area and the heights of the many
large buildings that have been constructed by
the industrial, commercial and scientific
organisations that have established themselves
in Sophia Antipolis.
Conclusion
The combination of orthoimage and DEM
data - when allied to the complementary
Canada, 588 Booth Street, Ottawa, Ontario,
K1A 0Y7, Canada
Dr. H. Rammali (hrammali@hotmail.com) Surveying
Department of Libya, Algeria Square, Tripoli. Libya
Mr. C. Lanchon (lanchon@istar.fr) Marketing
Department, ISTAR, 2600 Route des Cretes, 06905
Sophia Antipolis, France
References
Chromatek Inc., Alpharetta, Georgia, U.S.A. - Web
Site:- http://www.chromatek.com/
Th. Toutin & B. Rivard, 1995 - A New Tool for Depth
Perception of Multi-Source Data, PE&RS, Vol. 61, No. 2:
1209-1211.
Th. Toutin, 1997 - Qualitative Aspects of ChromoStereoscopy
for Depth Perception, PE&RS, Vol.63, No.
2: 193-203. s
chromo-stereoscopic image that can be
derived from these two data sets - is well
suited to a wide range of topographic mapping,
GIS and thematic mapping applications
where 3D stereo-viewing would be
helpful, indeed often invaluable, to the
users. As the examples show, the technique
can be applied to a wide range of spaceborne
and airborne imagery having quite
different ground resolutions.
Professor G. Petrie (g.petrie@geog.gla.ac.uk)
Department of Geography & Topographic Science,
University of Glasgow, Glasgow, G12 8QQ,
Scotland, U.K.
Dr. Th. Toutin (thierry.toutin@ccrs.nrcan.gc.ca)
Canada Centre for Remote Sensing, Natural Resources
Figure 3: Sophia Antipolis, Southern France (copyright: ISTAR)
11