NEW TECHNIQUE FOR COMBINING PANCHROMATIC AND MULTISPECTRAL 
SPOT IMAGES FOR MULTIPURPOSE IMAGE-MAPS 
Mohammed ESSADIKI 
Département de Cartographie Photogrammétrie, IAV Hassan II , B.P. 6202 — Rabat- Morocco 
m.essadiki(@iav.ac.ma 
  
Commission IV, WG IV/7 
KEY WORDS: Cartography, panchromatic, multispectral, segmentation, recognition, processing, image-map 
ABSTRACT: 
Satellite images offer an efficient and quick data source to obtain the most recent and the urgently needed topographic or thematic 
maps. Present technology provides the possibility of accelerating the production of these maps. A new technique for combining 
SPOT satellite images with two different ground resolutions is described. Using various image processing methods, including 
enhancement techniques, a good quality image suitable for multipurpose image-maps can be produced within a short time. 
1. INTRODUCTION 
From early times, maps have served to give an image of the 
earth's features, as a guide for the traveller, as a record of 
locations, and as an identifier of geographical features. 
Systematic national development is not possible without 
adequate accurate and synoptic information on the nature, 
amount and distribution of human and natural resources. The 
conventional methods of production of topographic and 
thematic maps do not provide a satisfactory answer to the 
demand for maps because they are slow and expensive. 
Developing countries are still at the stage of completing their 
map coverage and in many cases most of the existing maps 
are of low quality and/or outdated. 
During the past decades, the use of satellite remote sensing 
data for generating topographic and thematic maps has been 
the subject of much research and study. 
The urgent need for maps, on the one hand, and the 
development of technology, on the other hand, will 
unquestionably provide countries with the opportunity of 
making active use of the new data acquisition technology 
from satellites and the possibility of accelerating base map 
production. 
In this paper, a new technique is described for producing an 
improved image by combining SPOT panchromatic (10 m) 
and multispectral (20 m) bands. This product is the result of 
data pre-processing, feature enhancement, colour enhance- 
ment and image combination (Essadiki, 1987). 
2. DATA PRE-PROCESSING 
The basic processing (level IB) of the raw data was 
performed at the receiving station in Toulouse, France. This 
level includes radiometric and line-by-line geometric system 
corrections with regard to the Earth's rotation and curvature, 
viewing angle, registration, etc. Some other processing 
needed for this study is described in the following 
paragraphs. 
2.1 Haze correction of the multispectral bands 
Haze correction (atmospheric correction) using only a first 
748 
order additive haze model was applied to the multispectral 
bands. The cut-off points were derived from the histogram 
showing the minimum value of each band. The estimated 
haze contribution to the signal is subtracted from the photon 
count value in each band for all elements in the numerical 
image. 
2.2 Image to image registration using ground control 
points 
The selected test area (Kasserine, Tunisia) gave a sufficient 
variety of terrain types, je, urban and rural, forest, arid, 
agricultural, etc. Figure 1. 
  
Figure 1. The standard image (original) 
2.2.1 Ground control point file 
Image resampling involves the generation of a new digital 
data set which normally has scene element size, geometry 
and radiometric characteristics different from the original 
data set (Colvocoresses, 1986). The panchromatic and 
multispectral images in this study were acquired in the same 
orbit on the same day. For this reason, an affine 
transformation was used. The specific error model was 
derived from: 
e the position of actual grid of scene elements (x, y) and 
Internal 
the indi 
et 
corresps 
For res: 
element 
was nee 
(six wel 
the 20 n 
“master 
were lir 
roads. 
The re 
interacti 
the cont 
2.2.2 
Three n 
neighbo 
convolu 
The N^ 
readily : 
but it hz 
as does 
may be 
spectral 
tortion v 
The dif! 
and the 
regressi 
The res 
spacing 
registrat 
represen 
were me 
3. I 
The use 
been de 
about la 
interpret 
the data 
Image e 
which u 
dimensic 
or huma: 
Features 
features. 
homoger 
detectior 
indicator 
is suffic 
Detected 
the evide 
set of fee 
a- 
b- 
C- 
These