G.P. and R.J. Pike, 1991, Pike R.J. and G.P. Thelin,
1992). Digital shading using DEMs, gives the possibility
of creating realistic, detailed and accurate images of the
relief (Burrough 1986, Thelin G.P. and R.J. Pike, 1991).
The brightness of each point or pixel in the shaded relief
image, depends on its elevation relative to the elevation
of its neighbors, its slope and aspect, and the simulated
position of the source of light; direction (azimuth) and
elevation (angle above the horizon).
Shades are a primary clue to depth perception. But true
3D peception involves the existence of two images of the
same place or feature seen from different angles (Toutin,
P. And B. Rivard, 1995). A virtual 3D image of the
shaded relief can be generated by applying the DEM to
the original shaded relief image to obtain a second image
with parallax. For viewing, one stereoscopic image is
made assigning the original image with the blue and
green colors, and the image with parallax is assigned the
red color. When seen through anaglyphs a 3D image
"jumps". Other techniques include the use of polarized
glasses, stereoscopes, etc.
In a new technique, Toutin, T. and B. Rivard, 1995, use a
single color image taking advantage of the effect of
chromostereoscopy. In this case, depth is coded in color,
so that blue hues are seen farther, and red objects look
closer. A hypsographic map, when seen through special
lenses, the areas in hues of blue and green will seem to
be at lower elevations, and those with red will appear
higher.
DATA AND METHODS.
Digital elevation models and vectorial data.
The 255 DEMs that cover continental Mexico and most of
its islands were used. Taking as a base the 1:250,000
scale topographic map series from INEGI, these models
were generated by the United States Defense Mapping
Agency (US-DMA), and currently are available to the
general public. Vectors for international borders and
coastlines were obtained from the 1:4,000,000 scale
topographic map from DGG-INEGI.
Digital elevation model processing.
The individual one by one degree models were combined
first to form larger models (6 degrees of longitude by 4
degrees of latitude), keeping the resolution (3 seconds)
and geographic projection. A single elevation model was
designed to cover all the Country, with elevation values
every 500 meters and in a Lambert Conical Conformal
projection. Using the EASI/PACE* program REGPRO
running in VAX workstations, all the mosaics were
registered to this single model. The registration process
involved a resampling of the elevation data and a
projection change. No estimation of the accuracy of the
resulting model was made.
Shaded relief image.
The shaded relief image was obtained from the DEM
using the program REL, with 315° (NW) of azimuth and
* Mention of brands or companies is for information purposes only.
630
an elevation of 30°. A mask was created for all the zero
elevations, corresponding to the sea, and to portions of
the model with no data (other countries). The image
under this mask was set to a gray value of zero to
separate them from flat areas above sea level. The
shaded image was transformed through a piece-wise
transformation to enhance the relief perception in
relatively smooth areas.
Stereo shaded relief image.
Taking the first image and combining it with the DEM, a
second image with parallax was generated using the
program STER. The factor for parallax was 100. This
means that the displacement in number of pixels for a
given point between the original image and the second is
the elevation value from the DEM divided by 100.
Shaded hypsographic image.
The elevations were divided in 11 ranges, most of them in
500m intervals. For each range, a mask was created
from the DEM. Next, one base color was selected for
each interval, going from light blue in the sea (elevation
zero), and blue-green in the lowest parts to dark brown
and gray in the highest points. The colors were first
selected in a Red Green Blue (RGB) color space, and
then were mapped to the masks corresponding to each
elevation range. The three band RGB image was
transformed to an Intensity Hue Saturation (IHS) image.
A lookup table for transforming the shaded relief gray
image, was prepared for each elevation range. In the
lookup tables the value 128 (mean value in the gray
image) was substituted by the intensity value of the color
of the corresponding interval. Besides, the lookup tables
included a piece-wise stretching to enhance relief
perception. The lookup tables were then applied to the
shaded relief image to derive a new image, which was
then used as the intensity channel. The IHS image was
transformed back to RGB. This resulted in an image in
which each elevation range has a single base color
(same hue and saturation) but different shades (intensity)
of the same color, lighter of darker, according to the
shaded relief. A series of tests with different colors was
conducted until a satisfactory color combination was
achieved. In the last trials the colors were chosen directly
in the IHS color space; keeping constant the saturation
level, and varying the intensity and hue. This was also a
better way to make sure that a gradient from colors with a
major blue component to those with red as the main
component was produced.
Map production.
The previous paragraphs describe the DEM and image
processing. For the final map production, the images
were exported to the Laser - Scan (LAMPS) environment,
for adding a frame, international borders, names,
graticule of meridians, parallels, and legend. The final
steps were made with Barco Graphics software, and
included the transformation of the RGB images to CMYK
(for the 3D and hypsographic maps) the integration of
images, with the cartographic frame, preparation of hard
copy tests and print positives, with color separates in the
case of the 3D and the hypsographic maps.
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996