International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004 
sons 
{ E — Eu 
Stimulus | Experience & le 
es 3 i { Expectation —— | 
  
     
am. 
Spatial Grouping 
   
1 
Detection | Planar Grouping 
(Visual) 
Recognition 
—t. 
| Identification —————— —————— ——— 
  
(Technical 
Analysis) 
Figure 1: The Image Interpretation Process (Albertz, 1970) 
Ground pixel size of SAR images is considered to be of 
greater importance than different polarisation shown as col- 
our. The pixel size differs for the test areas from 1.5 to 4.0m 
(Table 1). In the Trudering area no polarisation has been 
used, leading only to black and white SAR images. Some au- 
thors (Ohlhof et al, 2000) refer to the NATO Standard 
(STANAG 3769), recommending the appropriate ground 
pixel size for the detection, recognition, identification in 
some cases also technical analysis of image objects (Table 3), 
whereby the following definitions are used: 
  
  
  
  
  
  
  
Object Detec- Recogni- Identi- | Technical 
tion tion fication | Analysis 
Terrain - 800m 90m 3m 0.75m 
Features 
Urban 60m 15m 3m 0.75m 
Areas 
Roads 6m 4.5m 1.5m 0.38m 
Railroad 15m 4.5m 1.5m 0.38m 
Bridges 6m 4.5m sm. | 0.3m 
Airfield 6m 4,5m 3m 0,15m 
Fascility 
  
  
  
  
  
  
Table 3: Required ground pixel size for optical image in- 
terpretation (STANAG 3769) 
  
  
  
Detection: In image interpretation, the discovering of the ex- 
istence of an object without its recognition. 
Recognition: The ability to fix the identity of a feature or ob- 
ject in images within a group. 
Identification: The ability to place the identity of a feature 
or object on imagery as a precise type. 
Technical Analysis: The ability to describe precisely a fea- 
ture, object or component imaged. 
Although STANAG 3769 does not discriminate colour or 
black and white imagery as well as contrast, it gives an indi- 
cation of the expected details which may be extracted from 
imagery. The work being carried out in this investigation can 
be allocated to the detection, recognition and sometimes also 
identification tasks in image interpretation, while the techni- 
cal analysis is something of more military nature. 
536 
  
Considering these facts and the experiences of the interpret- 
ers a list of common objects was set up prior to the interpre- 
tation in order to ensure comparable results. Altogether 45 
object types have been defined as common features although 
each interpreter also had some personal features. Some of 
these features have been selected for a detailed analysis sepa- 
rated for linear and planar (area) object types. 
At first the interpretation was made just with the SAR images 
to avoid an influence of the higher information contents of 
the optical images. Then the optical imagery was interpreted 
using the same object types as for SAR. In some cases this 
was leading to totally different results. For example the radar 
image of Trudering showed a historical site which was con- 
sidered a specified object type but could not be detected at all 
within the optical imagery (Figure 2). The reason is not 
known but might be caused by the different acquisition times. 
Figure 2: Comparison of optical and SAR images 
The analysis was carried out for all 4 test sites and resulted in 
a data set for each of the operators as shown in Figure 3. 
In 
ilk | 
mi [ 
  
  
R 
di 
( à 
tir 
D 
tif 
or 
to 
ar 
cu 
Fi 
tic 
te] 
ge 
It 
th 
fir 
art 
re 
pk 
Sh 
ria 
op 
tic 
tic 
TI 
Wi 
ad