)0K DIRECTION 
7 8 9 
* \ / 
7~7—r-?- r -r~r-? 
/ / 
LOUGHED STRIP 
1 2 
grass surface 
>rs in the 
er barbed wire 
ew metres wide 
iivide. Farther 
s grass cover, 
on on 23 June 
lar in X and C 
over the test 
rocessed (high 
obtained over 
able in X- and 
2° towards the 
only in C-band 
and during the 
respectively, 
if drum reflec- 
ght patches by 
film has been 
3x enlargement 
:he transparen- 
lears as a dark 
s could be ex- 
ction. On both 
as occur repre 
shallow depres- 
his is followed 
(urface texture, 
which separates 
ivide. The fire 
s of the rough- 
OK DIRECTION 
3 LOUGHED STRIP 
'—' 1 2(?) 
ilectors on 
tted) 
ness of the ploughed bare soil surface. The back 
ground tone of the grass field surrounding the drum 
reflectors is uniform and relatively dark as a 
result of the low scattering. 
On the radar strip 77 C-band, most drum reflec 
tors can be differentiated under the microscope, as 
shown in fig. 3. Their size and separation were 
measured with a micrometer. 
Drum reflectors 1 and 2 are 4 m apart and can be 
seen as two different point targets. The same is 
true for 2 and 3. Target points 3 and 4 are separa 
ted only 2 m in the field but appear on the image 
as two separate targets roughly 3 to 4 m apart, 
whereas drum reflectors 4 and 5, which are also 2 m 
apart, are hardly separable on the image. All five 
targets have the same orientation, horizontal, and 
not inclined towards the incoming radar beam. There 
is no indication for the reason why the first two 
targets gave less backscatter than the third one. 
Drum reflectors 6 and 7 are only 1.5 m apart and 
could not be seen as separate targets, but appear 
as one large blob. Reflectors 8 and 9, on the other 
hand, which are 2 m apart, can be seen as two dif 
ferent targets. Reflectors 10 and 11 also gave high 
returns, but 12 is not discernible. The approximate 
2 m azimuth resolution for these targets confirms 
the data given by Intertech. 
Fig. 4 Radar response drum-reflector field as seen 
under a 63X microscopic enlargement (3rd 
generation optically correlated imagery): 
a. Strip 78X under 32° depression angle 
b. " 77C " 42° 
c. " 76C " 70° 
Strip 76 C-band gives a distribution of point 
targets as sketched in fig. 4c. Most of the drum 
reflectors seem to be present in the image under 
the microscope and their site and separation is 
more or less according to scale. The alignment of 
the drums, however, is absent, and the image seems 
to suffer some distortion. 
On strip 78 X-band, the drums are not well defined 
point targets, but appear to be smeared out in range 
direction over distances of approximately 10 m. This 
seems to occur over the entire image. In azimuth 
directions, a number of the aligned point targets 
can be separated, although it is difficult to 
assign numbers to the blobs (fig. 4a). 
Of reflectors 1-5, at least 4 seem to be present 
and separable, and of 6-9, at least 2. This agrees 
wit an azimuth separation of about 2 metres. The 
range resolution, however, is far less for the X- 
band (about 10 m) than for the C-band (about 2m). 
As no more X-band images are available over the 
site, no conclusion is warranted if range resolu 
tion is far below specification on all X-band 
images. 
The following is a summary of the results yielded 
by the four radar strips. 
Strip 78 X-band, depression angle 32°: Drums not 
well-defined point targets, range resolution ± 10 m 
azimuth resolution ± 2 m. 
Strip 78 C-band: Drums reasonably defined point 
targets, range resolution ± 2 m, azimuth resolution 
± 2 m. 
Strip 77 C-band, depression angle 42°: Drums 
well-defined point targets, range and azimuth reso 
lution better than 2 m. 
Strip 76 C-band, depression angle 70°: Drums 
reasonably defined point targets, range distortion: 
no proper alignment, azimuth resolution ± 2 m. 
CONCLUSION 
As a conclusion, it can be stated that oil drums 
cut in half can be used as point targets for 
assessing spatial resolution. They are detectable 
on the radar image. Care must be taken that a uni 
form low-response background is present. A smooth 
grass or sand surface is good for this. A number of 
these drums, oriented with respect to radar look 
direction, should be placed in a recognisable con 
figuration so that it is possible to locate them on 
the radar images. Working with digital data makes 
quantification easier: nevertheless, optically 
processed data studied under high magnification can 
give good quantitative insight into the ground 
resolution and image quality of the system. Such a 
test site can be cheaply set-up, even in developing 
countries, and can serve as a test site at the 
beginning and end of each flight mission during a 
radar survey campaign. 
REFERENCES 
Forshaw, M.R., A. Haskell, P.F. Miller, D.J. 
Stanley and J.R. Townshend 1980. A review paper: 
Spatial resolution of Remotely Sensed Imagery. 
U.S. Committee Peaceful Uses of Outer Space, 53 
pp. 
Intertech Remote Sensing Ltd., 1981. ESA/JRC SAR 
580 campaign system specification document. 
Intertech report 951-81-1. 
Koopmans, B.N., C. Anton Pacheco, T. Woldai and A. 
Payas, 1985. SARTHI: A side-looking radar survey 
over the Iberian Pyrite Belt. Investigators final 
report Vol. 2. The European SAR-580 Experiment 
JRC/ESA publ. SA.1.04.E2.85.12/2, p. 125-149. 
Smith, D.J. Quality assessment of SAR-580 pro 
ducts. Investigators final report Vol. 1. The 
European SAR-580 Experiment JRC/ESA publication 
SA.1.04.E2.85.12/1, p. 63-69.