analysis ! 
-based system 
‘EE Trans. on 
317-329 
asic organiza- 
n: Yovits MC, 
oceedings sec- 
ystems, Spar- 
Magazine, 7: 
emantic mod- 
s from aerial 
O, Agouris P 
-made objects 
7-316. Basel: 
for the polyg- 
ss. CGIP, 1: 
ttklassifikation 
nferenzmecha- 
M-Bericht Nr. 
alysis of aerial 
rammetry and 
ja) Structural 
a blackboard- 
an A, Kuebler 
ction of man- 
e images, 53- 
Analyse von 
iutomatischen 
z städtischer 
it dI Ettlin- 
versität 
A NEW METHOD OF BUILDING GIS APPLICATION MODEL BASED ON GRAPHICAL VARIABLES 
Zhu Zesheng, Sun Ling 
Nanjing Navy Institute of Electronic Engineering, Jiang Pu, Nanjing, JiangSu, 210017, P. R. China 
JiangSu Academy Of Agricultural Sciences, Nanjing, JiangSu, 210017, P. R. China 
Commission III, Working Group IV 
KEY WORDS: GIS, Technology, Modeling, Model, Method 
ABSTRACT 
Geographic information system (GIS) has become a very important tool used to implement the integration of complex 
application system and to complete various complex graphic processing. In general, GIS application is mainly relevant 
to a lot of complex graphic processing, so that how to develop an effective method to fast build GIS application model 
with the help of the graphic processing ability for guiding GIS application system implementation has been being a 
challenge problem in GIS application and research area. In the paper, we present a new method for the design and 
implementation of a graphic-oriented GIS application model. This method is based on a new design principle of GIS 
application model, which uses a graphic variable to construct various complex GIS application model. The model 
makes use of GIS graphic processing functions, graphic variable properties and relationships among those variables to 
implement the design of complex application system. Otherwise, we discuss also a typical example for showing how to 
construct a practical GIS application system based on graphic variable, which was used to implement scientific 
management of stored grain. Finally, this paper discusses the general method for designing graphic variable model 
and some key techniques. 
1. INTRODUCTION 
Geographic information systems(GIS) are computer- 
based systems designed to store, analyze and display 
spatially referenced data(ESRI,1994) and have emerged 
as an useful tool for natural and other resource 
management. Thus, the most important feature of GIS is 
its excellent ability to process, store analyze the spatial 
information. Otherwise, GIS is usually used as a general 
tool or framework for integrating many other 
commercial software packages to build more complex 
and efficient application systems. However, how to use 
these GIS functions to design and build a successful and 
satisfactory GIS application system model becomes one 
key of successful GIS application system construction in 
most cases. Unfortunately, most of existent methods for 
building GIS application models are mainly based on a 
numeric analysis method where data analysis model in 
one dimension space is used as 2D or 3D space model, so 
that the difference between model design space and its 
operation or implementation space makes it much 
difficult to efficiently use advantages of GIS information 
processing in 2D or 3D space. Thus, the quality or 
performance of GIS application system implemented by 
the model based on traditional numeric method is usually 
unsatisfactory in most of more complex applications, 
which affects directly GIS application extent and 
increases total expenses of the system design, 
construction and management. 4n fact, the most 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B3. Vienna 1996 
important reason for using 1D rather than 2D or 3D 
method for building, GIS application model is because of 
the lack of an efficient method directly based on graphic 
variable analysis in 2D or 3D space for designing and 
building the application model or system in the above 
situations. However, our recent research shown that, 
really, graphical variables in 2D or 3D space can be 
reasonably defined to construct one new method that not 
only is directly used for describing GIS application task 
and designing its GIS application model, but also 
overcomes many disadvantages of the traditional 
numeric method and obtains a very satisfactory success 
in some typical GIS applications. Those successes show 
that the shift from numeric model toward graphical 
variable model is a profound. Further, some major steps 
of the method application are as follows: scooping of an 
application development project, in which the project 
area, requirement and relevant graphical variables are 
defined by the application plan objective, based on the 
GIS application system user requirement and graphical 
variable operation rules; collecting data relevant to 
graphical variables in the application field, in which 
these data must be represented as graphical variable form 
or transformed into the form by the method for 
generating graphical variables; developing, evaluating 
and selecting all important logical relationships between 
the graphical variables, in which the relationships are 
used to build basic framework of GIS application model; 
integrating the relationships and the data relevant to 
graphical variables, in which final GIS application