joint of 
me and 
ame are 
Ithough 
rge, as 
'de and 
tion of 
stortion 
| when 
hat the 
ifferent 
ndicate 
r radial 
can be 
| error 
> edge, 
is not 
there is 
always 
ere the 
ed than 
  
24 25 26 
1 
ferent 
]ters 
| point 
placed 
a new 
ns and 
i| point 
hen no 
> in the 
which 
cedure 
replace 
stem in 
ended, 
in the 
  
inbul 2004 
Generally speaking, the terrestrially calibrated parameters can 
be directly used for map production. This approach is efficient 
and provides adequate accuracy for rapid response applications. 
However, for precision mapping applications, another more 
precise alternative is used; where a part of the airborne survey 
data is used to refine the calibrated parameters. This approach is 
called Calibration and Quality Control. This approach has been 
experimented earlier and showed viable (c.f, Mostafa and 
Schwarz, 2001) where a number of well distributed tie points 
are automatically or semi/automatically generated and used 
together with the GPS/INS exterior orientation parameters and 
the terrestrially calibrated parameters to compute a refined 
system calibration parameters under the flight conditions at 
hand. 
3. THE DSS TEST FLIGHTS 
The DSS is frequently being test flown for performance 
evaluation purposes. The results of a number of DSS test 
flights are presented in this paper. This section is dedicated to 
briefly introduce the flight configuration of each of these 
flights. 
In December, 2002, the DSS was flown over a test field in 
Lakeland, Florida, USA. This flight will be referred to as 
Lakeland Dec02. The Lakeland test field has a total of 37 
ground control points (GCPs). The flight altitude is about 2000 
m which when coupled with the 55 mm focal length of the lens 
in use, results in a Ground Sample Distance (GSD) of about 0.3 
m. Six strips of images were flown where a total of 65 images 
were collected. Two more flights were flown over the same test 
field in January 2004 and March 2004, respectively. These two 
flights will be referred to as Lakeland Jan04 and Lakeland 
Mar04, respectively. 
  
  
  
  
  
  
  
  
  
Flight ig 
Flight E GSD #Strips/ # Check 
ID Aktitudesdote i-e retos ^ Points 
AGL (m) his 
Lakeland 
f 
get 1800 0.3 6/65 37 
1200 0.3 
EORUM & & 6/30 30 
e 1800 0.5 
1200 0.2 
ann & & 5/30 32 
1800 0.3 
NASA 1800 0.3 12/242 9 
Stennis 
E 1200 0.2 
ER & & 5/41 60 
s 1800 0.3 
Ajax 1200 0.2 9/165 46 
03 
PASCO 300 0.05 6/150 16 
Toyonaka 
  
  
  
  
  
  
Table 1: Configuration Parameters of Different DSS Flights 
NASA Stennis Space Centre has independently evaluated a test 
flight data set especially acquired for that purpose. The same 
data set has been evaluated independently at Applanix and the 
results are presented here. In February 2003, a set of data was 
collected in Japan where a total of 41 images over 5 strips were 
collected over a test field which has about 60 GCPs. In August, 
2003 a test flight was flown over Southern Ontario, Canada, 
where a part of this test flight data was flown over Ajax, 
  
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004 
Ontario and will be referred to as Ajax 03. Note that the test 
flights discussed so far have been flown to attain a GSD of 
about 0.2 m, 0.3 m, and 0.5 m. PASCO Corporation flew a 
rather high resolution test flight over Toyonaka City in Japan, 
where the flight altitude is 300 m which resulted in a GSD of 
about 0.05 m. This flight will be referred to as PASCO 
Toyonaka flight. Table 1 lists a summary of the configuration 
parameters of the aforementioned flights. An example of the 
flight trajectories of these flights is shown in Figure 7, where 
the Southern Ontario flight trajectory is plotted and the Ajax 
part of this flight is highlighted near the upper right corner of 
Figure 7. 
  
     
   
  
  
  
  
1 
4885 P oen: | 
Vivian Forest 7 E $2753 | 
ns Aura t Ostrava 
486% | f 
4855 ie e 
E 4848 
2 Don Valtey Parkway - } ^ Seaborough 
- Area 7} Four Area 
E 4835 x 
= 
t rn 
$e 825 z 
z 4825 7 
P Downtown Toronto 
p Area 
4815 7 
£ 
/ 
4505 LS 
479% 
XU 
oF Hamilton Afea 
4785 : i 
£80 606 629 640 664 650 “ao 
Easting (Km) 
+ GPS Basestation 
&  Alrpont 
GPS Basestation lecated at Appadx Cerperation 
Average Flying Altitnde: 1206 m AGL 
DSS Focal Lengthc £2 pun 
Figure 7: Southern Ontario Test Flight Trajectory 
4. TEST FLIGHT RESULTS AND ANALYSIS 
As mentioned above, the DSS test flights were flown for 
performance evaluation purposes. The most important 
parameter in the performance evaluation of a survey system is 
the absolute accuracy of the system when examined against a 
known reference. Therefore, the efficient way of analyzing the 
performance of the DSS is to compare the DSS-derived 
checkpoint coordinates to those surveyed independently by land 
surveying techniques. The statistics of checkpoint residuals 
simply depict the final absolute accuracy of the checkpoints. 
Moreover, the minimum and maximum values of the residuals 
show the error bounds for individual points. In addition, the 
mean value of the residuals show whether or not there is 
remaining biases in the system. 
In the following discussion, the results from a number of the 
DSS test flights are presented using the aforementioned 
checkpoint residuals as an evaluation criterion. Some of the 
data results have been previously presented in some detail in 
Mostafa (2003), Tachibana et al (1004) and Ip et al (2004). The 
presentation herein is, therefore, rather concise and focused on 
the repeatability of the DSS system performance in the different 
test flights discussed in Section 3. 
Note that all the results presented here are after the DSS data 
has gone through the airborne calibration and quality control 
procedure shown in Figure 2.