A RATIONAL APPROACH TO DIGITAL ANCILLARY DATA ACCESS AND
VISUALIZATION IN LAND COVER, LAND USE PROJECTS
Milena Danailova
Department of Remote Sensing and GIS,
Space Research Institute, Bulgarian Academy of Sciences, Sofia, Bulgaria
milena@space.bas.bg; m_ves_dan@yahoo.com
Commission VII, WG VII/6
KEY WORDS: Land Cover, Land Use, Interpretation, aerial photographs, Rectification, Orthorectification, database, GIS
ABSTRACT:
During the last decade many national and regional Land Cover/Land Use projects are implemented based on computer aided satellite
image interpretation. An essential procedure for the European CORINE Land Cover 2000 and EROSION projects is forward and
backward change detection. The use of three generations of Landsat images (MSS, TM and ETM+) with different spatial and
spectral resolution raises some specific geometric and thematic problems. Topographic maps and aerial photographs are
indispensable sources of support information. Usually old archives are not in digital form. The use of hardcopy topographic maps
and aerial photographs is a cumbersome and time-consuming operation. À rational approach to the conversion of these data and
advanced methods for visualization are vital for the interpretation process optimization. In this paper the alternatives of using either
geo-referenced or ortho-rectified aerial photographs are studied so that project specification requirements concerning planimetric
and thematic accuracy could be achieved. In this case optimisation means minimal hardware, software and human resources used. À
GIS data base management system is extended with a new structure in order to facilitate the fast and accurate choice of appropriate
ancillary data. This approach for data organization is very useful in the verification phase, when data for large territories is
processed. A fast and accurate choice of spatially overlapping data sets is achieved. The developed methodology is applied
throughout the implementation of the projects mentioned above. Examples of making decisions in some cases of uncertainty during
the interpretation are given.
1. INTRODUCTION e orthophotographs and aerial photographs 1:5 000 to
1:25 000 scale;
Throughout the last decade many national and regional Land e thematic maps — city, forest, vegetation, geological
Cover/Land Use projects are implemented based on computer and soil maps 1:5 000 to 1:600 000 scale;
aided satellite image interpretation. The European CORINE e land and forest restitution vector data bases, obtained
Land Cover 2000 (CLC2000) and EROSION projects are from 1:5 000 and 1:10 000 scale maps.
executed in Bulgaria during 2003 and 2004 based on
harmonized technology (Bossard ef al, 2000). The purpose of Most of the historical data are available in an analogue form.
CLC2000 is a revision of CLC90 data (based on hard copy The use of hardcopy maps and aerial photographs is a
technology) mapping of 1990-2000 changes and CLC2000 cumbersome and time-consuming operation. To facilitate the
database creation by an overlay procedure. The same computer aided interpretation a conversion to digital form and
technology is applied in the EROSION project for retrospective georeferencing is an indispensable procedure. The enormous
1975-1990 change detection for Bulgarian Black Sea coastal mount of this heterogeneous data has to be arranged and stored
zone. Three generations of orthorectified Landsat imagery are in a proper database with an user’s friendly interface. The
used — 1975 MSS, 1990 TM and 2000 ETM+ resampled to 50 database creation and management raises two main problems —
and 25 m pixels respectively. The mapping scale is 1:100 000 the appropriate date preprocessing and flexible
with a minimum mapping unit of 25 ha and 5 ha for changes. A accessivisualization.
lot of ancillary data of different types should be used in order to
reach the thematic and geometric accuracy specifications. The
main purpose of this paper is to describe the acquisition and 3. METHODOLOGY
rational access to all ancillary data used during the
implementation of the projects. According to the CORINE methodology satellite images are the
main source of information for land cover/land use mapping.
Most of the interpretation problems could be solved using the
2. ANCILLARY DATA interpreter's knowledge and proper combinations of Landsat
HERR : 3 channels. For the classes of non-homogeneous colour and
To assist the computer aided satellite imagery interpretation (exture like complex cultivated pattern, orchards, vineyards etc.
process the following sources of ancillary data are used: in many cases it's difficult to make the write decision. Similar
problems arise when relief or cloud shading masks the
* topographic maps 1:100 000 and 1:25 000 scale; vegetation cover. Ancillary data is obligatory in such cases.
e satellite images — SPOT, KVR-1000, Landsat;
818
by
ge
st
or
G
or
pr
To
da
Fo
bo
an
ZO!
an
The
gra
to
attr