Full text: XVIIIth Congress (Part B4)

  
APPLICATIONS OF SPACE IMAGERY IN THE DEVELOPING COUNTRIES 
Dan Rosenholm 
Department of Geodesy and Photogrammetry 
Royal Institute of Technology 
100 44 Stockholm, Sweden 
ISPRS commsion - WG IV/2 - International Mapping from Space 
KEY WORDS: Land Use Mapping, Natural Resources Inventory, Topographic Mapping, Satellite Images, Developing Countries, 
ABSTRACT 
The paper describes a number of mapping projects based on satellite images that the Swedish Space Corporation and its subsidiary SSC 
Satellitbild have been involved in since 1987. The first large vegetation mapping project, considered a major breakthrough in the earth 
observation business, was a land use mapping of the Philippines finalised within one year. A land use mapping, followed by a forest 
suitability study on South Sumatra, Indonesia, was carried out in the early 1990’s. During 1992 and 1993 a land use, forestry and biomass 
mapping of Malawi was carried out. A forestry mapping project of Northern Namibia started in late 1992 and is beeing finalised now. 
In Ethiopia SPOT panchromatic stereo images were used to derive a DEM, contour lines with 20 metre intervals and orthophotos for 
mapping at 1:50,000. Twenty-seven 1:50,000 topographic map sheets of the islands of Cebu and Bohol in the Philippines were produced 
based on existing old topographic maps and SPOT images. As a part of the National Population Census of Nigeria, 177 planimetric maps at 
1:50,000 were produced in early 1990. This mapping led to an updating of the regular 1:50,000 series by the Federal Survey Department of 
Nigeria of 350 more map sheets. In the Baltic States ongoing projects are producing a new series of topographic maps based on 
orthocorrected SPOT Satellite Image Maps. 
Earth Observation technology is a fully operational tool for land use and vegetation mapping and revision of topographic maps in developing 
countries. Major issues are related to availability, where cloudiness is one major problem and the programming and distribution system 
another. Strict handling of geometrical problems is essential in all applications. The trade-off between spectral, spatial and temporal 
resolution is treated. A typical land cover project design is described. Conclusions are made. 
1 INTRODUCTION 
The Swedish Space Corporation (SSC) and its subsidiary SSC 
Satellitbild have since 1987 been involved in a number of 
mapping projects in developing countries using satellite data. 
The first large vegetation mapping project, considered a major 
breakthrough in the earth observation business, was a land use 
mapping of the Philippines finalised within one year. The 
project was carried out at a time when remote sensing was seen 
more as a tool for researchers, using scenes or parts of scenes, 
than for practical use in larger areas. Since then thematic 
mapping and topographic mapping projects have been carried 
out by SSC on all continents. 
2 PURPOSE 
The purpose of this paper is to share the experiences from 
SSC’s international mapping projects and point out the 
important issues in mapping from space in developing 
countries. 
3 PROJECT DESCRIPTIONS 
3.1 Land Cover Mapping of the Philippines 
During one year’s execution time, between 1987 and 1988, a 
full Land Cover Mapping at 1:250,000 of the Philippines was 
carried out using SPOT scenes (Rasch et al, 1988, Rasch, 
1994). The primary goal of the project was to establish land 
cover statistics for project planning at the World Bank. The 
choice of SPOT was primarily based on the programming and 
tape recorder capabilities in combination with the flexibility and 
repeatability due to the side-viewing capabilities. In the 
beginning of the project almost no useful images existed and at 
that time it was practically more or less impossible to get a 
coverage with Landsat. A sufficiently cloud-free coverage was 
achieved during the project period, partly due to the fact that 
there were more cloud-free periods than usual during the year. 
The mapping was performed by interpretation of precision 
corrected scenes at a scale of 1:100,000. 24 land cover classes 
were interpreted on the 187 scenes used, 43 map sheets at 
1:250,000 were produced and land use statistics by region were 
computed. To obtain ground truth and give the interpreters an 
understanding of the area 19 air reconnaissance and 7 ground 
surveys were carried out. 
Twelve months was, especially in 1987-1988. a short time for a 
project of this size. It was possible due to the characteristics of 
the SPOT system and the enthusiasm of the team. The accuracy 
of the classification was good - the only frequent problem was 
two forest classes that were mixed up - especially taking into 
account that the number of actual classes was twice as much as 
stipulated by the contract. The project has since then got many 
followers internationally. 
3.2 Forest Suitability Mapping on South Sumatra 
As part of a study on management plans for industrial forest 
plantations of parts of the Musi River watershed, South 
Sumatra, two sequential studies of the same area were carried 
out jointly by SSC Satellitbild and Swedforest International. 
This is a part of Indonesia with intensive land use and rapid 
changes. A land use mapping in 20 classes was followed by a 
forest suitability mapping based on the land use information. 
Forest Suitability Maps are powerful instruments for land use 
planning and decision makers (Fig. 1). Satellite Image Maps of 
multispectral SPOT Images, complemented with Landsat TM 
scenes were used. The initial study was of pilot character and 
was carried out on twelve map sheets at a scale of 1:50,000. The 
project was carried out jointly by SSC and Swedforest 
International and was finalised in 1991. Sumatra is, like most of 
Indonesia, extremely cloud covered. The areas were selected at 
that time based on existing satellite data. 
The results were considered very satisfactory and have led to an 
ongoing full scale project covering 109 1:50,000 map sheets on 
East Kalimantan (Borneo) and Sumatra. 
704 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996 
  
71 
AN 
THT Tn o> 
  
w+ 
my 
  
  
  
REAM EERE
	        
Waiting...

Note to user

Dear user,

In response to current developments in the web technology used by the Goobi viewer, the software no longer supports your browser.

Please use one of the following browsers to display this page correctly.

Thank you.