Full text: Commission VI (Part B6)

0 toolboxes 
AL software 
oxes, icons, 
0SS OR 
EARCH MATRIX 
XEL OVER 
CIENT (r) IS 
OF r IS 
ee 
HELP 
NU STOP 
  
AL program 
he text lines 
arget matrix 
aatrix at the 
8. 
d CD-ROM 
ly stored on 
age data is 
Irive or from 
osses and 
re displayed 
ed figure. In 
the search 
yed at each 
two images 
mbers or as 
he formulas 
in addition. 
hen explain- 
carrying out 
se, the field 
1e best fit of 
to be found, 
search area. 
or "LEFT 
direction. At 
t has to be 
the compu- 
the compu- 
OWLEDGE" 
age with the 
1e keyboard 
It is correct, 
JORRECT!" 
answer are 
SK" or "RE- 
ceed. 2.4.3 
Correlation in the subpixel range: The basic information 
about this theme is presented by means of four text lines and 
two figures (see fig. 3). The sequence of the text lines can be 
repeated or stopped. Additional information such as formulas 
and explanation of the parameters is displayed when clicking 
on the field "INFORMATION". The exercise is started by 
pointing to "TASK". The contents of the theme covers the 
correlation within a profile, which means the target matrix is 
shifted in one direction only. This topic is easier to understand 
than correlation in two dimensions. It serves here as an 
introduction, but has in itself practical application for height 
determinations. 
  
THEME 2: CORRELATION IN THE SUBPIXEL RANGE RP 
b * gl 
a 
“> 
X - THE TARGET AREA HAS TO BE SHIFTED A SMALL AMOUNT 
[ FROM AN APPROXIMATE POSITION 
& 
Target area - THE TARGET AREA HAS TO CHANGE IN DENSITY AS WELL 
92 - THE SOLUTION IS FOUND BY A LEAST - SQUARE ADJUSTMENT 
- THE OBSERVATION EQUATION CAN BE FORMULATED AS: 
v2gi(g*a*b*gi(e) *c- ga(s) 
Information 
  
Search area 
  
  
  
INFORMATION 
  
Fig. 3. Theme 'Correlation in the subpixel range' from the CAL 
program "LDIP". At this moment of the sequence all text lines 
are displayed as well as the additional information (in smaller 
print). The sequence of the text lines can be restarted by 
activating the field 'REPEAT'. 
2.4.4 Correlation in the subpixel range - task 2.1: This 
exercise requires to find the coefficients of the observation 
equations which have to be read from the attached figure and 
set into the matrix and the vector areas. Thereafter the 
calculation of the normal equations is started and can be 
watched. The precise position of the target area and its 
accuracy have to be calculated by the student again. He or 
she uses the integrated pocket calculator for this purpose 
(see fig. 4). The results have to be entered into the keyboard 
and they will be displayed at the premarked result windows. 
All input values are analysed and commented by the program. 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
THEME 2: CORRELATION IN THE SUBPIXEL RANGE RP 
COMPLETE OBSERVATION AND COMPUTE UNKNOWNS! 
TASK 2.1: Observations: v = A * x - 1 e 
= 
Fill in A-matrix and 1-vector 
$11.92 0.00 1.00 1.00 1.00 
i319 BÍ m. 
10 040 9. ! 
= |0.00 9.00 1.00 8.00 
t Ll EE cj gm 
7 (_n Colculator — ) 000 100 1.00 1.00 
5 t ladjustment: viv = Minimum 
| LL 29% 90. [f] - [4 
LJ j 4 - 
2 rit x: 0.920 0.00 3100 700 e 27.00 
1 H 
pram [Solution of normal equations : 
EE EE 
i is 
081 ; 
m ENTER OFF §| « 
2 e 
=. Jde 
EL) 
rod i 
ji Caci) 5) 
m 
KNOWLEDGE RETURN TO MAIN MENU 
  
  
  
  
CALCULATOR EXE ER ER | 
Fig. 4. Exercise to the theme 'Correlation in the subpixel range 
- one dimension'. The unknowns of the correlation are 
determined by an integrated calculator. 
  
63 
2.4.5 Automatic measurement of real réseau images: The 
theoretical treatment is now supplemented with practical 
réseau images. The presented knowledge contains again a 
sequence of lines of text, dynamic figures, formulas and 
recommendation of literature to read (compare fig. 5). They 
are presented in various levels. 
  
THEME 4: 
AUTOMATIC MEASUREMENT RP" 
of Real Réseau Images 
x :;^$ has fine crosses in the image 
plane, which are exposed onto the film 
  
£$ show the crosses 
  
  
   
  
2* can be carried out 
and more accurate 
- The trensformation of the measured crosses to the 
coordinate values of the :z XH 
reveals the film deformatio 
  
  
anner deficencies 
  
à zoom factor: [+] - Imaged object points can be corrected by means of an 
affi i: where the sorrounding 
ntrol points E 
     
  
Select réseau image 
INFORMATION TASK REPEAT RETURN TO MAIN MENU STOP 
  
  
  
Fig. 5. Screen image for the theme 'Automatic Measurement 
of Real Réseau Images’. All the text lines are displayed by 
now; additional information will be shown when ‘clicking’ on 
the underlined key-words. A hypertext window will then 
display more text, formulas, figures or recommended litera- 
ture. The image at the left side of the screen can be inter- 
changed and zoomed. The selectable images are taken from 
project work. 
Additional information is received by activating either the 
keywords in the text or an icon. A hypertext window appears 
which can be shifted to another place, if it hides information. 
The exercise with real réseau images starts with adjusting the 
template in size, rotation and geometric resolution (pixel size) 
to the crosses of the réseau image. The first two crosses are 
measured by pointing at the approximate location. By means 
of the derived transformation parameters the measuring can 
now occur automatically. The speed of measurement and the 
setting accuracy is displayed. Various sets of transformation 
(affine or projective, with all crosses or with four surrounding 
crosses) are carried out, and the residuals at the crosses are 
displayed. The results demonstrate the advantages of the 
réseau camera. 
2.5 Some experiences from the development 
The development of "LDIP" occurred in the usual steps 
design, programming, testing, and documentation. 
The design of the program comprises the screen images for 
'Knowledge' and 'Task, the user interface and the program 
structure. The screen images are designed according to the 
technical contents, here "Automatic measurement ". The 
practical exercises also determine the amount of information 
which has to be presented in 'knowledge', for example the 
formulas, the figures, their expressions, etc. Computational 
tasks should not be too lengthy. Parts of the computations will 
run automatically and intermediate results are displayed. The 
user interface was designed according to standards estab- 
lished in previous CAL programs by the author. For example, 
the operation has to be as simple as possible. The sequence 
at the normal use of the program as well as all possible 
manipulations had to be specified for the programmers. 
Programming of the "LDIP" software was split into 'user 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B6. Vienna 1996 
  
  
 
	        
Waiting...

Note to user

Dear user,

In response to current developments in the web technology used by the Goobi viewer, the software no longer supports your browser.

Please use one of the following browsers to display this page correctly.

Thank you.