ı using 
tivities 
sing in 
| radar 
ncy of 
r. This 
s hilly, 
t, with 
)image 
6 km’. 
sed on 
f time, 
led the 
ensing 
ber 20, 
cessed 
the X- 
ta with 
ited on 
> radar 
| of the 
ds to a 
current 
of the 
10st no 
from a 
/s have 
Roman Arbiol 
  
visibility to the aircraft. This was achieved by computing a DTM from the contour lines of existing 1:500,000 scale 
maps and by performing a visibility analysis. GPS data at the reference stations was recorded for post-processing. 
3 MISSION PREPARATION 
The project area was divided into strips 130 km. long and 14 km. wide, and flown at 26,000 feet. Depending on the 
terrain type - flat, hilly or mountainous - the side overlap was 50%, 67% and 75%, in order to guarantee that enough 
images would avoid the shadowing and layover effects. Consecutive tracks were flown in opposite directions. 
The airfields were selected from the few available in the area. The selection criterion was suitability for flight 
operations: availability of supplies, accessibility and minimization of ferries. The selected airfields were Puerto 
Ayacucho, La Esmeralda and Santa Elena de Uairén. Fuel and food were carried by river and by air. 
The AeS-1 system relies on GPS and INS data for georeferencing, so ground control is used for checking purposes only. 
The checkpoints can be made visible on the radar image by installing radar reflectors (corner reflectors) at known 
positions. Therefore, a survey campaign was performed together with the Geodetic Department of the National Mapping 
Agency of Venezuela to measure 31 positions where the corner reflector would be installed. None was installed at that 
time, due to the risk of being displaced, destroyed or even covered by the vegetation. 
4 FLIGHT CAMPAIGN 
Before starting the mission, AeroSensing flew a calibration flight, in order to ensure the absence of any systematic error 
and to measure different parameters: for example, the corner reflector shape and width in the image, a precise 
measurement of the baseline, radiometric parameters for compensating the different radiometry in the far and near range 
parts of the swath, and the effective number of /ooks required to maintain a low enough noise level. For this project the 
number of /ooks was set to 7. 
4.1 DATA CAPTURE 
As mentioned earlier, the installation of the corner reflectors was performed simultaneously with the flight. Two 
reflectors were installed with two different orientations, so that they could be imaged on two opposite tracks. Some had 
to be installed on top of the “tepuys” and required the use of a helicopter to transport the people, the reflectors and the 
GPS equipment. 
The flight campaign lasted from October 28, 1998 to February 3, 1999, including repairs, mandatory revision of the 
airplane, re-flights, change of airfields, etc. There were 67 effective flying days, with around 10 tracks per day. While 
cloud cover does not affect the radar images, dense clouds with high water content and turbulence do prevent 
operations, since they directly affect the quality of the image. 
4.2 ON-SITE QUALITY CONTROL 
After landing, the on-board disks were removed and the contents copied to DLT 7000 tapes. On-site quality control 
followed, in order to detect any anomaly: 
e Radar checking: analysis of parameters such as raw signal variations, power variations, etc., in order to 
ensure that the radar had operated correctly during data acquisition 
e Movements of the aircraft: attitude variations of less than 3° in roll, 1° in pitch and 2? in heading, and 
variations of less than 30 m. in height and 100 m. in position 
e [mage quality: a 0.8 x 14 km. segment of each track was fully processed and checked 
If the control was passed, a copy of the DLT was sent to the ICC (Barcelona). The quality control detected some bad 
tracks, which were re-flown immediately. The amount of data collected was around 16 Tb. 
  
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part Bl. Amsterdam 2000. 25