A. Rothkirch
Figure 1: EGO Goniometer at the Joint Research Centre, Ispra, Italy. In the sketch of the instrument on the left the
numbers indicate: 1) horizontal rail, 2 + 4) vertical quarter-arc, 3) light source, 5) detector, 6) target support. On the right
you can see the author preparing the sample ’roof tile’. In the back of the right image you can see the two quarter-arcs
with mounted detector (left, 0,, = 70°) and source (right, 0; = 30°).
2.2 Measured BRDF values
In this section we present the measurements made in the principal plane’, this means measurements with a relative azimut
angle between source and detector of 0° or 180°. While the detector zenit angle was varied across the principal plane,
source zenith angles! were choosen to 8; = 30°,45°, 55°, 65°. Measurements were done for roof covering materials (a
roof tile made of baked clay and a sample of roofpaper made of sand and bitumen) as well as for a Spectralon (Labsphere
Inc.) sample of 50 % reflectance (a description of the Spectralon can be found in (Meister et al., 1996)). BRDF values f,
were calculated from the ratio of the reflected L, radiance and the incident irradiance E;. The accuracy of the measured
values was determined to of, / fr = 4.3 %, a detailed desription can be found in (Rothkirch et al., 1999).
Fig. 2 shows the measured BRDF values of the different surface materials across the principal plane. Each column gives
the measured values of a surface depending on viewing zenith angle 0, at different illumination angles 0; = 30°, 45°, 55°
and 65°. The solid lines indicate measurements at polarization state ss (sample illuminated by s polarized light and
detection of reflected light at polarization state s), the dashed lines indicate sp, the crosses show pp-measurements and the
plus-signs show ps-measurements. Negative viewing zenith angles correspond to backward scattering direction.
It can be seen that all sample surfaces show a specular peak at the ss-measurements which increases with increasing inci
dent zenith angle 0;. For example the ratio q,, between maximum and minimum BRDF value of the sample "Red roof tile
rises from q,, = 3 at 0; — 30? up to q,, = 12 at 0; = 65°. The specular peak is shifted towards greater viewing zenith
angles 60, (to 0, 70? for all ss measurements). The pp-measurements also show a specular peak which inreases with
increasing incident zenith angle 0;, but much weaker as compared to the ss-measurements. Similar results were found
for a Spectralon panel with 100 % reflectance by (Haner et al., 1999). For the sample Red roof tile’ its maximum ratio
is Gpp ~~ 3 at 0; = 65°. At the pp-measurements of the sample "Red roof tile’ there is an increased reflection around
viewing angle of 0, = 25°. While it is dominating at an incident zenith angle 6; = 30°, it seems to vanish at 0; z 59?
(unfortunately the field of view of the detector was too small to give reasonable results for measurements at 0; = 65° and
low viewing zenith angles ).
It also can be seen that there is a significant difference of the absolute height of the BRDF values between the measure
ments at different polarization states. Comparing the ss- and pp-polarization measurements of the sample ' Red roof tile,
one can see that while in backward scattering direction the BRDF values are = 0.04 [1 / sr] for ss-measurements, the
values for pp-measurements are greater by = 0.015 [1 / sr] ( see also fig. 3). Every sample surface also shows different
absolute BRDF values at cross polarization measurements (sp resp. ps).
While the samples 'Spectralon' and "Red roof tile' only show significantly increasing specular reflection in forward
scattering direction, a noticable increase in backscatter direction can be seen for the sample 'sanded roof paper . This
is due to the sample structure: the sample consists of bitumen sparsely covered by sand. Hence if one looks from nadir
onto the sample one can see the low reflective bitumen as well as the brighter sand. With increasing viewing zenith angle
masking effects decrease the visible portion of bitumen until only sand can be seen. Assuming rotational symmetry of the
sample surface one expects that a portion of the overall BRDF rises with increasing viewing zenith angle. Although this
P c . . - A o
1 A zenith angle 0 — 0? corresponds to nadir. In the sketch of the instrument (see fig. 1), detector angle 0, and source zenit angle 0; are = 60°.
778 International Archives of Photogrammetry and Remote Sensing. Vol, XXXIII, Part B3. Amsterdam 2000.