ul 2004 
racters. 
arts of 
iclusive 
charya, 
ds: GIS 
13:495- 
jon for 
“Nepal, 
ciences, 
Y 
A study 
iced to 
Nepal, 
nology, 
ary). 
DATA INTEGRATION RELATED TO SENSORS, DATA AND MODELS 
Farhad Samadzadegan 
Department of Surveying and Geomatics, Faculty of Engineering, University of Tehran, Tehran, IRAN 
Email: samadz@ut.ac.ir 
Commission VI, WG VL/4 
KEY WORDS: Fusion, Spectral Information, Spatial Resolution, Region Growing, Change Detection, Potential Evaluation 
ABSTRACT 
Data integration or fusion refers to the acquisition, processing and synergistic combination of information provided by various source of 
data. The scope of this article is to describe three typical applications of data integration in photogrammetry and remote sensing. The first 
study case refers to the evaluation of the potential of different image fusion techniques on integration of two satellite images with 
different spatial and spectral resolution. The second one considers the problem of the object extraction in outdoor situations and the 
solution which proposed base on a feature level fusion. The third one presents the characteristics of a decision level fusion strategy for 
construction of an automatic system for detection of changes between available sensor information and corresponding digital vector map. 
Each study case presents also the results achieved by the proposed techniques applied to real data. 
I. INTRODUCTION 1990; Chavez et. al, 1991; Kathleen and Philip, 1994: 
Rockinger, 1996; Sharma, 1999). The choice of a suitable 
In recent years, there has been a growing interest in the use of fusion level depends on the available information type: when 
multiple source of information to increase the capabilities of the sensors are alike, one can opt for fusion at the pixel-level to 
intelligent machines and systems (Varshney, 1997; Hall and take all data into account. When sensors or information are very 
Llinas, 1997). Due to this, information fusion becomes an area different, decision-level fusion is more suitable and is also 
of intense research activity in the past few years (Varshney, computationally more efficient. Feature-level fusion is the 
1997; Llinas, and Walts, 1990; Abidi and Gonzalez, 1992; proper level when the features as found by the processing of the 
Clark and Yuille, 1990; Luo and Kay, 1995; Hall and Llinas, different sensors can be appropriately associated. 
1997). Information fusion is a process of combining evidence 
from different information sources in order to make a better The application of the fusion approach shows successes with 
judgment. It plays an important role in many application techniques ranging from expert systems to probabilistic 
domains. No single source of information can provide the techniques. As there is no simple rule for selecting the proper 
absolute solution when detection and recognition problems fusion technique, a wide range of techniques has potential 
become more complex and computationally expensive. applicability. The process of selecting optimum algorithm for 
However, complementary information can be derived from fusion is complicated by the fact that data analysis seeks to 
multiple sources (Samadzadegan, 2002). combine incomplete and missing data in a complex 
environment in real time. In this paper three typical applications 
One of the important issues concerning information fusion is to of these techniques are described. 
determine how to integrate (fuse) the information or data. 
Depending on the stage at which fusion takes place, it is often 
divided into three categories, namely, pixel level, feature level 2. DATA FUSION IN PIXEL / IMAGE LEVEL 
and decision level (Abidi and Gonzalez, 1992; Hall and Llinas, 
1997). In pixel level fusion, the combination mechanism works Image fusion (i.e. fusion in pixel/image level) refers to the 
directly on the pixels obtained at the sensors’ outputs. Feature synergistic combination of different sources of sensory 
level fusion, on the other hand, works on image features information into one representational format. We use the term 
extracted from the source images or the features which are image fusion to denote a process generating a single image 
available form different source of information. Decision level which contains a more accurate description of the scene than 
fusion works at an even higher level, and merges the any of the individual source images. This fused image should 
interpretations of different objects obtained from different be more useful for human visual or machine perception. The 
source of information. different images to be fused can come from different sensors of 
the same basic type or they may come from different types of 
Several data fusion algorithms have been developed and sensors (Figure 1). The sensors used for image fusion need to 
applied, individually and in combination, providing users with be accurately co-aligned so that their images will be in spatial 
various levels of informational detail in photogrammetry and registration. 
remote sensing (Anderson, 1987; Burt, 1992; Carper et. al, 
569