ing
ng.
ely
n a
'eas
ther
ted
olar
es a
| the
tical
tive
‘gets
obal
lite
atial
hour
be
er to
pass
gular
f the
s an
sional
an be
)s are
track
time
didate
o the
y and
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004
history of the elements, and the expected cloud cover situation
for them. Different formulas for computation of the weight
factor can be specified. If Cloud Avoidance Scheduling is
enabled, then target elements with predicted cloud cover are not
taken into account.
The scheduler starts with the selection of the strip with the
highest priority that does not violate the imaging constraints.
Subsequently strips with the next highest priority are selected,
etc. Note that during the selection process the imaging
constraints are becoming stricter due to the increasing amount
of time needed for imaging and slewing to already selected
strips. Although this procedure does not necessarily result in the
most optimal selection, it leads to a rather efficient task list.
Especially the adoption of a fixed strip length is not optimal.
However the simulator can easily be extended with alternative
scheduling algorithms due to its modular architecture.
The capturing of a target element is recorded to be successful if
the element appears to be not cloud covered at the acquisition
moment. Of each element, the co-ordinates, the imaging times
and the imaging results are stored in the observation results
database.
In figure-3 an overview of the graphical user interface of the
CLIMAS simulator is shown.
{ Image information
+ GeollFF ft dui ne 3m, lation Browse | Lat (deg)
TARGET XML: f Upper lei: 15
Lower aght: i41
Ec
i
| c Missions |
i Dimensions (455 852
{
i
w Use cloud coves
tage Resolution: 130 20 km
Cloud de jE
Î Satelite file: [Satelite Piisence3455. Browse |
Maximum cloud percentage | 90 % e
Browse |
i Fist orbit {1
£V Use cloud coverage when scheduling
fis Last orbit. [100
For schaduing: take clouds [0 bout: ago disi
* Maximal lime difference scheduling Taurer [30 km
scheduling
Alfares: 12.0 km
Miranum [7 day: uo Tue Feb 01 10 30 13 19341
I
i
Í
i Maxmal number of tecordings
i
i # saps: (10
i Mavimun f21 day
í Sobt [5 Hae [2
Output le; [X Browser
Mininum solar elevation. (10 deg
| Next | Abort |
Iv. Pause between orbits o i i
Figure-3: Overview of the CLIMAS GUI
i
5. SIMULATION RESULTS
Simulations have been made for several satellite configurations
with varying number of platforms and sensor parameters. First
the simulation of a defined ‘standard’ satellite configuration is
discussed, after which the effects of variations of several of the
parameters are described.
The features of the standard satellite configuration are described
in table-1. In general this configuration consists of 4 high
resolution optical satellites as currently operational. The area
for which the simulations are made covers 3400x2150km of the
European continent. The simulation is run on a grid of 2x3km
elements (1700x750=1.275.000 pixels).
First an ideal situation is simulated in which no constraints due
to cloud coverage or minimum sun elevation are taken into
account. A scheduling strategy is applied for monitoring of the
network with a frequency of 14 days. This is filled in by
weighting the pipeline elements with factor 0 to 7, depending
on the number of days since the last observation (minus 7 days
and with a maximum of 14 days).
Table-1: Used ‘standard’ satellite constellation parameters
Parameter Value
Platforms
Nr satellites 4
Altitude 500 km
Inclination 97.3785 degrees
Orbits/day 15.225
Ascending node crossing time 94013 12:23:00.0 (sat 1)
Ascending node crossing longitude 0.0 (sat 1)
Track direction descending
Agility:
Max pointing angle along track 33 degrees
Max pointing angle across track 33 degrees
Slew speed 2.0 degrees/sec
Stabilization time 2.0 sec
Scheduling:
Nr of sub-strips 10
Monitoring frequency: 14 days 14 days
Tasking parameters LSETmin, LSETmax 7,14
Observation strategy max monitoring days
Sensor:
FOV: 10km 1.4 degrees (10km)
Atmosphere conditions:
Cloud period 1994/1995
Use of cloud information for scheduling yes
Cloud info time delay 0 hours
Use of cloud information for collection yes
Sunlight elevation constraint > 15 degrees |
The simulation resulted in the observation of 932.209 network
elements, or 36.3 times the network. The total area of all these
network element observations is 9.7% of the maximal system
observation capacity. This means that the inefficiency due to the
line structure of the network is more than 90%.
In a next step the constraints of sun elevation and clouds are
introduced. The results of the simulations are shown in table-2.
Table-2: Simulation results for basic cases
observations total | observ. 7-14 days | monitoring days
number times| number times number % full
case elements network} elements network monit.
No clouds, sun>0 932.209 36.3] 607.521 23.7% 5.346.013" 57.0%
No clouds, sun>15 911.206 35.5] 589.409 22.9]. 5.183.592. 55.3%
Clouds, sun>15 582.860 22,7) 274.909 10,7| 2.717.495 29.0%
For describing the effectiveness of the observation system
several parameters have been defined, as shown in table-2. First
one can look at the total number of observed elements. This
number is shown in the first column, with next to it the times
the network can be covered by this number of elements. Not all
observations are relevant however, for the required two weekly
monitoring frequency only the observations taken after 7 to 14
days after the last observation of the element are taken into
account. The number of these relevant observations is listed in
the second column, also accompanied with the times the
network can be covered by this. A factor of 26 times the
network covered by relevant observations does not mean that
10096. monitoring takes place however, because many of the
observations does not take place after 14 days, but after a
shorter period of up to 7 days. Therefor a third parameter is
defined: the number of monitoring days. This means the sum of
each relevant observation multiplied by the number of days
after the last observation of this element. In fact the last
parameter most correctly denotes the effectiveness of the
system.
From the table it can be seen that about 30 — 50% of the
observations done are not relevant, not within 7-14 days after
the last observation. Further that the influence of the sun
elevation constraint of 15 degrees is very limited for the total
system effectiveness. It locally may have large impacts however
for the northern regions. The influence of the cloud conditions
above Europe is significant, as may be expected. The
monitoring capacity is reduced from 55.3 to 29.0%. In general
this means that with the defined constellation of 4 high
1063