TTT 
red to some 
irs. In this 
an optimal 
ues, we cy 
verage. 
lite image; 
ults of such 
ea (Contact 
[Information 
raction and 
, 63(8), pp. 
I Imagen. 
^ pplic., 21, 
Landforms, 
M from the 
Society of 
& Remote 
lite Stereo 
te Sensing, 
Yotography: 
| Journal oi 
Ammatzia Peled 
TOWARD AUTOMATIC UPDATING OF THE ISRAELI NATIONAL GIS — PAHSE III 
A. PELED*, B. HAJ-YEHIA** 
University of Haifa, Department of Geography, Haifa, 31905 Israel 
*a.peled@uvm.haifa.ac.il 
**basheer(@geo.haiga.ac.il 
Working Group IV/3 
KEY WORDS: Updating, GIS-Driven, Change Detection & Identification, Geo-referencing, Edge Detection. 
ABSTRACT 
Spatial Information revision and updating is the main concern and production effort of maintaining the ever-growing 
GIS systems and spatial Databases. Developing easily effected automatic updating methods of spatial information 
becomes the key to the successful maintenance of the large GIS data bases established by many mapping agencies all 
over the world. The paper presents the third phase in a three-year research effort for the development of automatic 
capabilities for updating the Israeli National Spatial Database. This ongoing research in funded by the Survey of Israel 
within the framework of maintaining the Israeli National GIS. Discussed are the experiments and the evaluation steps to 
define an efficient method for Change Identification and Feature Extraction. Presented are the GIS-Driven change 
identification and the texture classification methods. 
1 INTRODUCTION 
There are two approaches for updating spatial databases. The first one is to, gradually, establish a new database which 
will replace the old one. This approach is guiding many private vendors who are mapping road networks. A second 
approach is to detect, identify and update only the changes. This approach is suitable, for instance, were customers are 
attaching value added information of their own. In this research, the later approach was chosen. A modular Change 
Identification System was developed, which comprises three major stages: (a) Change Detection; (b) Change 
Recognition; and (c) Change Identification. 
The steps taken by the Survey of Israel for the transition from traditional updating methods to advanced digital-domain 
methods, were reported on the ISPRS’ IC meeting in Haifa, September 1997 [Peled and Raizman, 1997]. The 
experiments of change detection algorithms, were discussed and presented in the ISPRS commission IV meeting in 
Sttutgart, Germany, september 1998 [Peled and Haj-Yehia, 1998]. This paper deals with the experiments of the change 
recognition and identification methods. A GIS-Driven method to identify changes is presented in this paper. The 
objectives of GIS-Driven Identification algorithms, suggested by Peled [1992], are to exploit the existing spatial 
information, stored in the old database. This enables the building of training sets for the classification and identification 
algorithms. In addition, edge detection algorithms were evaluated, to extract geographic objects. Also, geometric 
parameters (size, shape, etc.) classifier was tested to recognize the regions of changes and the geographic objects. 
Furthermore, statistical and texture parameters filters were developed to identify the radiometric signatures of the 
objects. The above algorithms, for the recognition and identification steps, were integrated together according to a 
strategy of developing a modular system. This enables applying identification rules related to recognition (geometric) 
with rules related to identification (statistical and texture). 
These experiments were applied for two layers (streets and buildings) from the Israeli National Database. This was due 
to the limitation of using black&white aerial photographs which served the Survey of Israel, originally, to remap the 
country and, nowadays, to update the National GIS. 
2 GIS-DRIVEN CHANGE IDENTIFICATION 
The GIS-Driven change identification method was suggested, by Peled [1992; 1993; 1994], as an advanced algorithm 
for identifying changes. By this method, the existing geographic database is used to build the spectral training sets, 
Which are used for classification. 
  
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B3. Amsterdam 2000. 713