Full text: Proceedings, XXth congress (Part 5)

    
  
  
International Archive 
coordinates (Xi V, ) of other control points on the chip 
can be obtained. The difference between image point 
measured coordinate (xdi) and their ideal coordinates 
(x; V.) is the total of all system errors caused by 
imaging. Proved by several experiments, the distribution 
regularity of the systematic error including its value and 
direction, is in accordance with the distribution regularity of 
optical radial distortion. 
The surface degree of CCD chips may maintain at the level 
of +1 iM . The object space plane control network is 
produced on the AVIO TAB plotting table by Leica 
Company, Swiss. Aided by the computer plotting functions, 
a 1.1mx0.9m control grid is created on the 0.1mm thick 
polystyrene film. The interval of the grid is Sem. There are 
399 black solid round symbols with a diameter of 15mm on 
the intersection of grids. Thus, the network might have 
enough intensity, as Figure 1 shows. With regards to 
experiences, the central error of each symbol is no more 
than +0.05mm. With the similar method, similar plane 
control networks can be created on the 8mm thick adhesive 
backing glass. 
  
Fig.1 X 2D control field 
The Nikon E2 (1280x1000 pixels, sizel0 £71 ) and Kodak 
DCS 4800 (2160x1440 pixels) digital cameras are used as 
the test camera to implement “vertical photography” for the 
control network. Select four points that are close to the 
s of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004 
image edge as the control points. Establish the projection 
relationship between the oblique image and plane control 
network and work out the coefficients 1. 
Set the measured coordinates of symbols as (x, 2 ). The 
ideal coordinates would be (xi, vj) aided by factor I. 
Their difference (Ax, j Ay,) is caused by various optical 
distortion errors (Ax; ’ Ay, ). 
AUG mx jc 
=a (6) 
Av: y: Yu 
Where (x.y ;) is the theoretical value, while 
(Ax,, Ay;) is the distortion error. 
Presented method does not touch any relationship between 
the physical factors of the error. The physical factors include 
radial distortion and decentring distortion in objective-lens 
optics, radial distortion coefficients changing with the 
focuses, radial distortion change of the object point that is 
not on the focus, and other geometric imaging errors caused 
by electronic and optical factors, and etc. Take the test done 
with Kodak DCS 4800 as an example. Rectified according 
to four corner angle points (001,015,136,150), the image 
point distortion error Ar. of each control point is shown as 
Figure 3. Its size and direction is in accordance with the 
general deduction of distortion error of radial distortion. The 
error vector directs to the center of the image frame (it 
should directs to PPA: Principal point of autocollimation). 
Moreover, it is easy to analyze that the distortion of the 
whole image presents as from the image edge to the center, 
when the distortion of the four corner points is not zero. 
A primary model of distortion error Ar can be created 
with these data, as Figure 4 shows. The model is created by 
the distortion error Ar of the 150 points on the digital 
camera chip. The unit of each direction (x, y, Ar ) is pixel. 
The primary model on direction x is shown as Figure 3. 
while the model on direction y is shown as Figure 6. 
     
   
  
   
  
   
   
   
   
   
  
   
  
   
  
    
    
   
  
   
   
   
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
    
 
	        
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.