terrestrial) images that can be applied within an automated or 
semi-automated process of stereoscopic exploitation. 
3D model inserted in a SIT 3D on topographical bases of 
neighboring relations as well as relations between different 
primitives ( initial or final nodes, arcs and sides was created 
for the present case by the exploitation of aerial 
photograms at a scale of about 1: 6 000 in 1994. It concerns a 
restricted area from Bucharest city. Since data collected 
using aerial photograms are incomplete considering only 
aerial photograms exploitation in urban areas, the shadows 
surfaces, the object contrast (taking into account the nature of 
building material specific to urban area and their reflectivity) 
do not allow the accurate identification of all details. These 
data were completed by data obtained through the 
digitisation of topographical plans at a scale 1:2 000 referring 
tothe same area. 
Thetesting area consist ofthe districts from Bucharest city 
central area where the building are very old and whose 
Structure could not resist during an earthquake. This area 
consists of about 80 buildings of different heights, shapes, 
structured and age. 
The topographical details collected from the stereomodel 
are geometrical represented by their planimetric position to 
which we can add elements that specify connections and 
adjacencies for each detail. The attributes added to every 
detail specify, the characteristics and the connections 
between all these. The details are grouped in classes of 
objects as for instance punctiform, linearly details and 
surfaces details. The data base is topographical structured 
the details being connected between them through entities 
as for instance arcs, nodes. 
The information referring to the method of data collecting 
namely if it is digitized on ground or at a z heights or an if 
it is a fictitious arc, are associated to each arc ( defined 
through the point final or initial). The data base consists only 
of vectorial data. 3D post disaster model can be created 
on classical technologies basis through the classic 
aerophotogrammetrical process using optico- mechanical 
sensors or CCD opto-electronic sensors on define interest 
area, but only during the first step. Many times this steps 
must be processed in real time. Video images taken through 
rapid intervention means, helicopters, airplanes for low 
altitude must be used. 
The use of video images for geometric data base in urban 
areas was developed by (Mills P.J. and Newton I., 1996). 
Video images can be used for 3D model of urban areas after 
a preliminary processing for an automatic stereoscopic 
exploitation. Video images of the tested area using S-VHS 
high resolution camera mounted on a helicopter at three 
different heights were taken within the experimental 
researches, that were performed for this project. The defined 
technologic process and the respective processing are due to 
be achieved and good results are foreseen for the tested areas. 
3.3 The detection of the changes by comparing the two 
3D models 
The difference between 3D reference model and the 
models issued after the disaster is performed. This difference 
will generate other two new models, one in which the 
differences are positive and one 3D model in which the 
differences are negative in relation with the reference 
model. A new 3D model of notable changes is obtained 
through the combination of these two models of positive and 
negative differences (it is indicated that the two types of 
changes should be represented through two complementary 
764 
colors). This model is conceived only for the polygon 
defined zones that presents significant changes in a first 
step. The positive changes correspond to new structures, 
to the reconstruction of the urban zone, of its 
modernization or to natural phenomena as for instance 
earthquake. 
3D model negative changes correspond to damages due to 
disaster. These changes must be analyzed in detail using 3D 
model on a large scale. The analyze is performed for zones 
that define positive changes in order to define their 
nature but also for areas that indicate negative changes 
that can define secondary effects that follow natural 
disasters or phenomena indicated by these. They will be 
classified corresponding to different types of changes 
adopted for this analyze. 
The accuracy of building damages determination is 
connected with the type of data used and with their geometric 
resolution. The type of changes detected refers only to 
structures or to parts of collapsed buildings that can be 
directly determined be stereoscopic measurements or can be 
deducted by the image photointerpretation. There are 
elements that cannot be pointed out by the analyze of these 
model. These natural phenomena (earthquakes, fires etc.) 
induce profound changes to the inner structure of the 
building as for instance cracks, sliding, collapses of inner 
walls. These changes can be pointed out using SITh 3D 
conceived for inner organization for different networks 
(heating, electric light, telephone etc.) for each building 
concerning the shape, dimension group location of different 
structure element and specific inner space of every building 
according to its use are conceived in order to accomplish SITh 
3D information system. This SITh 3D will be accomplished 
using all the other data bases (social and economic, 
demographic). The analyze of the changes the type and the 
class of every change at these level must be performed 
through subsidiary terrain works and than comparison to 3D 
when they will be analyzed in view of final evaluation. 
4. DISCUSSIONS 
The principle proposed for the development of a monitoring 
and evaluation system for the natural disaster effects is 
adapted to the development of the intervention steps in case of 
natural disasters. The construction of the 3D reference model 
can be accurately and completely performed for the entire 
urban and periurban zone. The model built after the disaster 
must be more rapidly performed using rapid intervention 
means together with an adequate technologic processes. 
After some catastrophic disasters the identification of 
enough control points cannot be performed. The use of other 
elements like linear features, identified from the reference 
model can be used for the conceiving of this model. The 
implementation of GIS or SIT 3D in our project is at its 
beginnings. The serial analysis of data bases of this type 
must be adapted to this type of problems. As for the future 
our interest goes to the establishing and practical 
implementation of the steps of video images used in the 
construction of after disaster 3D model together, with the 
detailed analysis of changes in order to answer the necessity 
of a precise evaluation quichly and efficiently in disasters. 
S. BIBLIOGRAPHY 
Airault Sylvain, 1993, Extraction du réseau routier sur 
photographies aériennes. Bulletin d'Information de l'IGN nr.61 
pg.19-24.. 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, Budapest, 1998 
Kofl 
1996 
Supp 
Arch 
Part | 
Kone 
Inter 
Sensi 
Korox 
Rese: 
Plann 
Nago' 
Fig! .