il 2004
| infor-
image
size of
r crys-
>
"D
rture À
tion of
sure in
> in the
es the
;ase of
E is to
/ small
| deter-
is also
output
just as
evices.
possi-
chnol-
device,
signal,
SITY
similar
parent
red. In
energy
mostly
'e;p is
),) and
: trans-
€ sam-
s half
e
rve
ince
led
as
icity
ot cor-
propor-
ing the
(or of
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B1. Istanbul 2004
The proportionality of thickness and density can be seen in the
Table 2.
Table 2. Thickness of material, transmittance, opacity and
density.
d T O D
1 0,5 2 3
2 0,25 4 0,6
3 0,125 8 0,9
4 0,0625 16 1.2
S 0,03125 32 1,5
6 0,015625 64 1,8
7 0,0078125 128 =
8 0,00390625 256 2,4
9 0,001953125 512 27
| 10 0,0009765625 1024 3,0
Reflection density is derived analogously as
D = log(1/p).
Assuming that a monitor is able to present linearly a density
range of 3,0, i.e. that realized by means of a step wedge, the
normal l-byte or 8-bit presentation spreads the density range
over the values 0 to 255; in increment of 1 corresponds there-
fore to a density change of — 0.0118. The 8-bit density values
presented by a step wedge with density increments of 0.15, a
clear material of density 0.5 and a maximum density of 3.05 are
shown in the following Table 3; also shown the corresponding
transmittance values assuming 8-bit, 9-bit, 10-bit, 11-bit, 12-bit,
13-bit and 14-bit presentation.
Table 3. Transmittance in several quantizations and 8-bit
density values for a steptablet.
D D (8) | t(8) | 1(9) | z(10 | c (11) | 02) | x03) ] 1 (104)
3.05 0 0 0 l 2 4 7 15
2.90 12 0 1 l 3 5 10 21
2.75 25 0 | 2 4 7 15 29
2.60 38 | 1 3 5 10 2] 4]
2.45 50 1 2 4 7 15 29 58
23 63 ] 3 5 10 21 4] 82
2:15 75 2 4 7 14 29 56 116
2.00 88 3 5 10 20 4] 82 164
1.85 100 4 7 14 29 58 116 231
1.70 113 S 10 20 41 82 163 327
1.55 125 7 14 29 58 145 231 462
1.40 138 10 20 41 81 163 326 652
1.25 150 14 29 58 nas 230 461 92]
j 10 163 20 41 81 163 325 651 | 1301
0.95 176 29 57 115 23 459 919 | 1838
0.80 188 40 81 162 324 649 | 1298 | 2597
0.65 201 57 114 229 458 917 | 1834 | 3668
0.50 213 81 162 324 647 | 1295 | 2590 | 5181
0.35 226 114 228 457 914 | 1829 | 3656 | 7318
0.20 238 161 322 645 | 1292 | 2584 | 5168 [10337
0.05 251 227 455 912 | 1824 | 3650 | 7300 |14601
0 255 255 511 1023 | 2047 | 4095 | 8191 |16383
While the density values are evenly spread over the available
range, the transmittance values are clearly not. In order to
achieve similar resolution in the densest part of the step wedge,
13 to 14 bits of transmittance are needed.
Density requirements vary to some extent from country to
country depending upon the type of terrain, the type of used
film and the processing facilities available. They are deter-
mined by two main considerations: (1) best resolution is
achieved at densities in the order of 0.8 to 1.0, and (2) the
illumination used in the plotter should not generate heat, i.e. be
not intense. Analogue monochrome contact or contrast-adjusted
diapositives are most commonly used in plotters, while
negatives or contrast-adjusted diapositives are scanned to obtain
digital image data for the use in digital evaluation equipment.
Exposure and processing of monochrome aerial film shall be
such that, using density values above base plus fog and read on
land detail at least 5 mm in extent, the minimum density is not
below 0.2 or over 0.6 within a 10 cm radius from the image
centre. The density shall not be less than 0.1 anywhere within
the image outside of this area. The maximum density shall not
exceed a value of (minimum density + 1.0); in very-high-
brightness-range scenes such as mountain snowfields it may
reach a value of 2.0, and it may exceed this value for images of
extremely bright spots such as specular reflections of the sun.
An average gradient of development shall be chosen such that
the negative density range is as close to 1.0 as possible. This
aim density rànge can be obtained with low-contrast processing
for high-brightness range terrain and with high-contrast proc-
essing for low-brightness range terrain. The requirements are
considered not to be met if the density range of a roll of film is
less than 0.7 and its average gradient less than 1.3, or if the
density range of a roll of film is more than 1.4 and its average
gradient less than 1.0.
If densities exceeding 2.0 are disregarded, only — 11 bits are
needed for the transmittance values to achieve a separation
similar to that of the density values for the allowed maximum
density of 2.0, and the 10 bits permit a clear separation. It is
noteworthy at this point, that the scanner available to the au-
thors uses 10 bits and can only meaningfully scan a density
range of ~ 1.9. Therefore, all photographs are measured in a
densitometer before they are to be scanned, and the scanner is
than set such that the available dynamic range is applied to the
density range to be determined.
5. EXPERIMENTS
In order to demonstrate the dynamic range of photographic
emulsions compared to that of a CCD used in a digital camera,
a digital colour image (RGB) available with 12-bit radiometric
resolution was converted to an intensity-hue-saturation image
(HIS). The black-and-white intensity image was then converted
from recorded (reflectance) values to densities to obtain logH
subject values and transformed using a characteristic film curve
to reflect the process of photographically recording the image.
The latter process simulated several exposures to explore the
possible loss of detail as a result of underexposure. The
resulting images were visually investigated in regard to
information loss within deep-shadow areas. All computations
were carried out using PCI's Geomatica software.
The investigation was based on a section (Figure 3) of a single
DMC image available on a demo CD "Digital Mapping Camera
(DMC) Operational Flight Data", Version B, from September
2003. The image was taken on 2003-MAI-24 from 1700m and
shows castle Neuschwanstein. Sunny skies causing shadows
and forest areas helped generate an image with a large subject
range.
