APPLICATION OF IMAGE PROCESSING AND IMAGE ANALYSIS METHODS FOR 
LARGE SCALE MAP REVISION 
Mohammad Sohrabinia 3 , Saeid Sadeghian b , Dadfar Manavi c 
a Shahid Beheshti University, Faculty of Earth Sciences, Department of Remote Sensing and GIS, Tehran, Iran 
b Geomatics College of National Cartographic Center (NCC), Tehran, IRAN, P.O.BOX: 13185-1684 
National Cartographic Center (NCC), Tehran, IRAN, P.O.BOX: 13185-1684 - 
ms_gis2003@yahoo.com, (sadeghian, manavi)@ncc.neda.net.ir 
KEY WORDS: Ikonos, Geometric Correction, Image Fusion, Object Extraction, Change Detection, Map Revision 
ABSTRACT: 
The aim of this study is to use different image analysis and processing methods in order to extract information content needed to 
update large scale maps. Recently available high resolution satellite imagery attracted mapping communities to shift their focus from 
aerial photographs to satellite imagery. Obviously, this new source of information requires different methods and algorithms to 
extract needed information for map revision. The purpose of this paper, therefore, is to examine reliability of image analysis methods 
to extract fine groups of classes equired to update large scale maps. Data used include Ikonos and QuickBird images of the study 
area together with aerial photographs and digital maps of the same region. 
1. INTRODUCTION 
Maps are date stamped evidences of a locality preserving the 
state of earth’s surface in their production time. Mapping the 
earth’s surface, resources and ecosystems is an invaluable 
source of knowledge about the past and present state of our 
planet. Mapping is termed as “Land literacy” and a source of 
information for “Traditional use studies” and “Hazards 
management” (McCall, 2003). Maps are prepared based on the 
current state of the areas of interest; however, as far as we know, 
the surface of the earth does not remain intact and goes under 
changes over time. The rate of change in urban areas, which is 
our focus in this study, is faster than natural landscapes. In order 
to preserve the validity of the maps, they need to be updated in a 
specified time intervals and new changes should be included. 
The information we need for the revision process is a vital scrap 
in the whole map updating procedure. Depending on the 
magnitude of the revision project and the extent and nature of 
the area covered by maps to be updated, revision process can 
vary extensively. In general, funding and time constraints are 
major concerns which make us to find more efficient and less 
time consuming methods to acquire essential data and 
accomplish revision in the least time. As a result, finding the 
most efficient source of information with the least cost and time 
involved has been always desirable for mapping communities. 
Traditionally, the information required for the revision was 
collected through land surveying. Aerial and space imaging 
systems revolutionized land surface data acquisition. Aerial 
photography offered great help for mapping community, 
although became available only after the advent of airplanes. 
Yet it had its own difficulties; stitching many number of 
photographs each of which had been collected in different 
interior and exterior orientation of the whole imaging system 
were a tedious task. With parallel developments in space and 
imaging technologies, satellite remote sensing was evolved. 
Satellite systems, on the other hand, are more systematic and 
cover larger tracks compared to aerial photographs. Even in 
terms of high resolution satellite imaging systems (such as 
Ikonos and QuickBird), dimensions of the images are 
considerably larger than any high resolution aerial imaging 
system. Moreover, remote sensing images, by and large, cover 
vast areas with a standard pixel size, making further image 
processing tasks even more straightforward. 
In view of the qualities mentioned above, this paper examines 
potentials of remote sensing images for map revision, putting 
greater emphasis on spectral methods. It is attempted to estimate 
the efficiency of image processing and analysis methods in the 
revision of large scale maps. 
2. BACKGROUND 
Map revision is one of critical discussions for the mapping, 
photogrammetry and remote sensing researchers and enjoys a 
rich literature. We have used a number of these researches in 
this study as a background and as a source of technical help and 
methodology. Below, a number of these works are introduced. 
Patynen (1998) has described revision process of 1:5000, 
1:10000 and 1:20000 scale Finish Topographic DataBase (TDB) 
with digital methods. He has noted that time intervals to update 
1:20000 maps had been specified in 10-20 years but in 
populated areas where changes are more rapid, this has reduced 
to 5-10 years. He has outlined that some major facilities like 
roads, power lines and administrative structure r eed continuous 
updating (every year). He has explained the whole process of 
scanning, geometric registration and ortho-photo creation and 
object extraction methods for TDB updating process. 
Di et al. (2003), Croitoru et al. (2004) and Hu et al. (2004) have 
extensively analyzed geometric correction and sensor modeling 
using rational functions. They have provided a good knowledge 
about rational functions and rigorous and non-rigorous sensor 
modeling. Origins of rational functions and their calculation 
formulas together with rational polynomial coefficients (known 
as RPCs), which has been used by satellite vendors like Ikonos 
to offer sensor modeling parameters, have been fully detailed. 
1351