ng this 
"he test 
amera. 
aneous 
linates 
üch is 
m the 
ection 
Jeyapalan, Kandiah 
  
  
  
  
  
  
  
  
  
Calibration of Mandli CCD Camera Using Test Range 
Std Dev of Residuals Calibrated Principal Standard Error 
At Control Points (ft) Distance of unit weight 
X Y Z (pixel) 
Photo 1 0.268 0.301 0.041 -165 5.09E-01 
Photo 2 0.689 0.508 0.088 -878 2.33E+03 
Photo 3 0.057 0.120 0.025 -868 4.61E+00 
All Photo 0.438 0.315 1.531 -774 2.77E+04 
  
  
  
  
  
  
  
  
Table 3. Results of Calibration 
Table 3 also shows that (X,Y,Z) of feature locations using the multiple video imageries (non stereo) can be obtained to 
an accuracy less than +0.5 ft; however the accuracy in Z is about +1.5 ft. The error of +0.5 ft in (X,Y) is acceptable for 
mapping at a scale of 1"=25' or smaller. The error of +1.5 in Z is acceptable for information about road side features 
but not for preliminary or location surveys, which may require + 0.1 ft accuracy. 
Table 3 also suggests that the strict photogrammetric model used in the Calib may not be acceptable for calibrating a 
digital imagery. However, by limiting the object distances from 25 to 50 ft from the camera and the camera locations 
from 0 to 15 ft, it may be possible to get accuracies better than + 0.5 ft in all (X,Y,Z). 
Fig 12 shows a video logging system of Transmap which is equipped with a pair of digital cameras unlike the Mandli's 
system which used only one camera. Four pairs of the Transmap's video imageries , 25 ft apart (See figure 13), and the 
soft plotter, which is a soft photogrammetric workstation by Autometrics Inc, were used to produce the GIS map 
showing the roadside features. First, the digital cameras were calibrated using the Calib software. Then, using the 
interior orientation elements, a triangulation was done using the soft plotter, followed by collecting line and point data, 
viewing one stereo pair at a time on the stereo plotter. See fig 14. 
The Fig 13 shows that the location of features between stereo pairs are satisfactory. However, locations of features 
obtained beyond 50 ft from the camera are not acceptable for mapping at a scale of 1" —25' scale. 
  
Fig 12. TransMap Van Fig 13. TransMap Video Image 
  
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B5. Amsterdam 2000. 411