errors due to the 
“arth curvature, 
the ground ue errors, the data 
processing errors using multiple spatial 
correction, data conversion errors (lower 
in the ease ot veetor-to-raster or raster- 
to-vector conversion and bigger in the case 
of raster-to-raster conversion). 
(platforms), the 
atmosferic refraction and F 
ici Bw 
4. LOCATIONAL TOLERANCES 
OF DATA 
On the base ot KMSE and cofactor 
matrix of redundant conversion 
(transformation) there are established 
tolerances on the basis of statistical 
tests. The tolerances have a measuring 
unit of locationa! data (in the case of 
iaster data, the ineasuiing unit is the pixel 
edge equal with 1). 
The tolerance of map resolution 
may be used for vectorial data at the 
definition of topological ares (e. g. the 
definition of minimum length of the are, 
when the ares with the length lower than 
the tolerance are eliminated). The dangle 
lengih tolerance haa ihe same value as the 
map resolution tolerance and defines the 
taet that all the ares with the distance 
between their ends lower than this value 
are considered incident in the same node. 
The mnverted coordinate tolemnce 
iz estimated differently regarding to the 
processing — type, using the RMSE 
muliiplication with the  Siudeni 
distribution factors, regarding to the 
confidence level. 
For altitudinal tolerance is estimated 
a value. different to the transformation 
model and to the adopted function for 
the terrain approximation, which, from 
ihe theory and practice [5]. là 
v(d) = k d* 
where d iz the distance between points 
with altitude, R 15 ihe rugosity of ihe 
terrain, experimentaly determined for the 
different terrain forms; the variance v is 
equal with k when d = 
5. ARNOWLEDGEMENTS 
The final locational accuracy in 
the integrated supersystem data base or in 
the digital map given by a lot of indicators, 
but the most important is the RMSE of 
the  senemlized transformation. The 
locational error and tolerances may be 
defined in the 2D or 3D space, like the 
locational data, the passing from one 
space to another may be rigourously 
controlled. 
The later on studies must refer to 
thematic error estimation or the 
classification — error, generally, in the 
spaces upper than the 3D space. Is also 
necessary the standardization of all types 
of errors, accuracy indicators and 
tolerances and their introduction in the 
nonnatives, so that the GIS calitative 
factor underlining becoming prevalent. 
References 
i. ESRI, 1989. Users Guide, Vol. IL: 
1U.15-10-26 
2. Jeipalan, K.. 1972. Calibration of a 
comparator. In: Phot. Eng. 38(5y 
472-478 
3. Kubik, K., 1988. Digital Elevation 
Models: review and outlook. In: Int. Arch. 
274 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B2. Vienna 1996 
  
  
Ey 
m. 
etx 
« "p 
Led "e 0