3) 
suspension of the camera. In space applications this error is irrele- 
vant because of the slow and linear changes of the orientation para- 
meters. 
Interpolation error Ip 
  
Orientation t 
parameter Computed parameter of orientation 
XYZ pi^ True course of orientation parameter 
Q, Q, ae ee 
9 . — Qo — — Pj+1 — € 
-—77 | TW — 2 . 
—Q—— — f. — 
Pj 
| Update points A ; | 
| | | 
[* 777 DOR — 951 . 
Nj-! Nj Ni+1 Time t 
  
Image error o — lp : F(X, Y, Z, w, @, X) 
Figure 6. Interpolation errors 
The weight coefficient Ark is the appropriate diagonal element of the 
cofactor matrix, the inverse of the normal equations. It expresses the 
influence of the camera and flight parameters and the shape of the 
terrain on the accuracy of the DEM-coordinates k. From this weight co- 
efficient Q,, the flight altitude h in units of the focal length c, 
which deterfifnes the image scale factor m_, can be extracted. This for- 
mula (3) reveals in general form the accul'acy of the DPS-strip tri- 
angulation. The coordinate errors Ov. Ov; 0, of the DEM-points are 
proportional to 
the image error O5: 
the image scale factor mn, * h/c and 
the square root of the reduced weight factor Qe 
The weight factor Q' vk is influenced by 
the other camera parameters, e.g. by the sensor geometry, the 
angle of convergence y, the image angle of view perpendicular 
to the flight direction; 
the length and width of the strip; 
the intervals between the update points £L 
the DEM-intervals along and across the flight direction; 
the shape of the terrain. 
For all camera- and flight configurations with equal geometrical pro- 
portions and angles the Qu, -factorsare equal. 
- 351 -