International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B2. Istanbul 2004 
  
operations 
A set of rotation, scaling & translation factors were calculated 
based on least square method technique, which will transform 
original feature to the transformed feature. These parameters 
may be modified to take higher order transformation. But using 
higher order will allow bending of objects (which spans across 
several triangles), thus improving relative accuracy but 
geometric errors gets introduced. The original object is 
transformed with these rotation, scaling and translation 
parameters, to re-establish original geometry. : 
2.4 Updation 
The updation of geo-data requires the detection of changes 
between up-to-date geo information and the existing data set. In 
case of manual updation of the data, the operator decides if the 
stored object has changed. The automatic updation process 
requires a similar decision, which means that the extracted 
object have to be compared with the stored object. As the PAI 
considers the latest orthophoto or stereo model is concerned, 
manual changes are required to be captured for new objects. The 
updation is carried out using unit area over which PAI operation 
is carried out as tile. For obtaining absolute accuracy for the 
changes on the ground the following checks are required. 
1. Nodes are not breaking in the tile edge. 
2. There should not be any duplicate tile edge beneath 
the original tile. 
3. UPDATION AND POSITIONAL ACCURACY 
IMPROVEMENT FOR 3 DIMENSION 
3.1 Preparation 
Another innovation developed at KIL made use of existing 
scanned paper maps for production of updated new maps in 3 
Dimensions using stereo images. 
The existing stereo aerial images are used for obtaining digital 
terrain information by automatic tetrain extraction. An in-house. 
program was developed to obtain the height information (from 
digital terrain model) of corresponding plan metric points in the 
vectorized maps. The height information is obtained from 
automatically generated digital terrain model as TIN data of the 
corresponding stereo images. Thus the Automatic program 
merges Z value from DTM data with X, Y values from 2D 
digitised file. 3D vector data so obtained is now ready for 
Positional Accuracy Improvement and updation. 
    
  
Old Vector 
Data s 
  
  
  
qiue rag 
Figure.5, Links are provided for shifting the vector, black 
features are old vector data and green circles show the shift 
required for Positional Accuracy Improvement 
3.2 Applying PAI 
After the links are placed and triangles are generated the PAI 
tool does the Positional Accuracy Improvement over 3- 
dimension. The red marks show the displacement required with 
respect to old vector data. 
A typical out put of the Positional accuracy improvement of 
vector data is shown below. 
  
  
  
  
Figure.6, Transformation before (features in black) and after 
PAI (features in magenta), desired shift or links (in red) 
Inter 
New 
prod 
with 
spec 
retail 
old r 
The 
prod 
chan 
posit 
been 
subs: 
The 
Kam 
Dim 
using 
Refe 
Refe 
Wise 
for S 
Ball, 
Recr 
Coxe 
Geor 
pp. 9 
Hilbe 
Imag 
Stott. 
from 
der K 
Refe 
Smit] 
Repo 
Arlin