ı order to 
:ollected 
sensito- 
jates the 
jhts and 
/ studies 
NE 
.1.Letus 
he top of 
s of facts 
3l driving 
k trough 
lected in 
! general 
or to con- 
EU 
TO- 
ZE 
LIGHT 
:NSI- 
ITO- 
  
  
stem 
Some parts of the inference engine like backtracking 
are given within the Prolog library while the others are 
written by the "knowledge engineer" who was identical 
with the human expert in this project. 
We can imagine the work of inference engine as it 
searches the solution relentnessly from top to the 
bottomin both sides of Fig.1. 
3. THE DEVELOMPMENT OF AIR PHOTOGRAPHIC 
KNOWLEDGE BASE 
The ES written under Borland's Turbo Prolog uses a 
.DBA file which consists.of facts while the rules are 
imbedded into the AIRPHOT.EXE file. The data used for 
investigation are collected in an internal knowledge base 
which is called onto the main program at the start of 
executing. The internal data base has its own advant- 
ages and restrictions. Opposed to external data 
handling the acces times may be shorter but the main 
program cannot handle more than one thousand 
Missions taking into account 640 KB memory.The pro- 
gram sends message about the free memory existing 
forfurther data input. The Prolog itself allowes develop- 
ing external data base as well but it needs longer 
acces times. The facts are organized into groups con- 
taining dataoffilmtest sensitometry, description of me- 
teorological situations,on-board data of photography, 
laboratory data of photographic development,control 
sensitometry, image densitometry and optional data of 
microimages. 
3.1. Preliminary Test Sensitometry. 
When the aerial film material arrives into the stock, the 
sensitometric data are investigated and compared to 
the data sent by the film manufacturer. The values 
Stored in facts are:No of emulsion, Type of 
emulsion,Dfog, Gamma and Speed. The facts may be 
referred 5y the 
No. of emulsion. 
3.2. Descriptionof Meteorological Condition and 
Visibility. 
The photographic staff characterize the haze and 
smoke conditions in situ which influence the image 
quality . These are judged value of horizontal sight, the 
type of clouds and cloud cover if they are present, 
the verbai description of visibility. Most important is the 
judgement of horizontal sight, while the other para- 
meters are used in the descriptive part of the report. 
3.3. On-board Photographic Data, 
The on-board photographic data are collected in se- 
parate facts. These consist of the most important 
photographic data which are used for description of 
the mission or partly for qualification. The input data 
arethe No. of mission, No.of emulsion, date and time of 
the mission,type of field, flight height, data of exposure 
and the names of the flying staff. The data will appear 
in the descriptive part of the Report, only the flight 
height related to the horizontal sight will be evaluated. 
The on-board photographic data may be achieved by 
the No. of Mission. 
623 
3.4. Data of Photog raphic Development. 
The data of photographic developments are stored in 
development facts. The characteristic data are the 
No. of Mission, Type of processor, type of developer 
solution, time and temperature and the name of the 
responsible laboratory personnel. The dataare usedin 
the descriptive part of the Report. 
3.5. Data of Control Sensitometry and Image 
Densitometry 
The data of control sensitometry and image densito- 
metry are collected in the after development clauses, 
whose are identified also withthe No. of Mission. These 
are the control gamma, control speed,Dfog, Dmin, 
Dmax. These data together with the test sensitometry 
play an important role in this program so that these 
dataare essential for the evaluation. (Boberg,A.,1988.) 
3.6. Datalnput from Edge Gradient Analysis. 
The ES is prepared to use data optionally collected out 
of a separate program protocol of EGA on line measur- 
ing program developed for aerial photographs before. 
The parameters are the height/diameter ratio of Line 
Spread funtion, and the factor of the MTF which is 
calculated by linear regression, before the MTF is line- 
arized withtwo fold logaritmisation. 
3.7. Some Remarks to the Knowledge Base 
Development. 
The ES deals with the knowledge base development 
as well. The knowledge base has to content reliable 
data, otherwise the decisions will fail. The data collec- 
tion has to accomodate to the on-board protocol 
usual in different countries and organisations, so that 
slight modification of the program may be necessary. 
But the fundamental data of test and control sensito- 
metry and image densitometry may not be absent. 
Themain programis preparedfor correcting and delet- 
ing of data inthe knowledge base as well. 
Now we have already understood that the ES needs 
some slight modifications before it is transported, be- 
causeithas noShelllike adaptability. 
4. RULES FOR QUALITY EVALUATION 
The other part of the System is the collection of the 
expert rules summarized in the left side of Fig.1. The 
rules follow the thoughts of a "tame" expert when he 
studies the on-board and laboratory protocols in order 
to investigate the film quality and rank it into classes. 
The computer makes the same: the facts placed in 
the knowledge base are investigated with the rules 
built into the main program. Conclusions are drown 
and appear in the Report. 
The expert rules may be divided into two main groups. 
In the first one the data are interpreted and explained 
a bit more freely in order to stimulate the personnel for 
thinking continually in technological processes refining 
them with "tiny steps". 
The second part of rules contains the national regula- 
tionforrankingfilms into one of four quality groups.