LAND USE/COVER REGISTRATION USING REMOTE SENSING DATA 
B. Ruzgiene, J. Stadalninkaite 
Dept. of Geodesy and Cadastre, Vilnius Gediminas Technical University, Sauletekio al. 11, LT-2040 
Vilnius, Lithuania — Birute.Ruzgiene@ap.vtu.lt 
Commission IV, WG IV/7 
KEYWORDS: Photogrammetry, Remote Sensing, Land Cover, Interpretation, Updating, Database, Imagery, Accuracy 
ABSTRACT 
Remote sensing has provided an efficient method and perfect alternative of data acquisition for land use administration and land 
cover maps updating. Establishing Land information system in Lithuania there is taken into consideration not only Lithuania needs, 
but also requirements of European Union in land use classification. For such purpose it is investigated land cover/land use 
classification methods using satellite imagery and data interpretation defining imperfection of current situation, developing 
perspective recommendations. The strategy and approaches of geo-objects interpretation using LANDSAT 7 imagery is outlined. The 
determination of most reliable land cover data have been carried out for maps at a scale of 1:50 000. Developed methodology 
includes determination of suitability criteria for certain group of geo-objects, applying mathematical statistical analysis and 
determining the confidence interval. Such methodology helps to emphasize the possibility of maps interpretation using remote 
sensing data. The research and accuracy investigation is based on a sample of satellite image, orthophoto map and digital vector 
database of Lithuanian Space Imagery Map. The accuracy analysis is referred to identifiable topographic objects that satisfy the 
defined suitability criteria calculating confidence intervals for each feature types. The results of research present mapping potentials 
for land use/land cover maps update purpose using remote sensing data. 
1. INTRODUCTION 2. A SAMPLE OF REMOTE SENSING, 
PHOTOGRAMMETRY AND GIS DATA 
The creation of the topographic maps at a scale of 1:50 000 in 
Lithuania with the usage of satellite images started in 1993 The combination and separate usage of satellite images allow to 
within the agreement between National Geodesy Service and evaluate the area to be investigated, perform visual and 
SSC Satellitbild (Swedish Space Corporation). Lithuanian Space computer-aided land interpretation. Satellite data serves as an 
Imagery Map at a scale of 1 50 000 (digital vector database input data for further investigation. The digital ortophoto 
LTDBK 50 000-V) was compiled using SPOT panchromatic and images supplies useful information about land use and provides 
geometrically corrected multispectral images. The database is visible ground representation. It is important information source 
periodically updated on the base of orthophoto images at a scale for satellite image georectification and interpretation purpose, 
of 1:10 000. Large interest is directed to application of satellite rather good background information for vector data edition. The 
images for land parcel identification. digital database supplies vector information about current study 
area and refers to a GIS data with possibility to recognize and 
Satellite imagery from LANDSAT is used for land use/cover analyze the spatial relationships among mapped phenomena. 
registration and maps updating at a medium scale, so the 
objective of research was to analyze on a sample of LANDSAT 7 The investigation area (50x50 km) covered by LANDSAT 7 
satellite images suitability for constructing and updating maps at satellite imagery (multispectral and panchromatic with 15 m 
a scale of 1 50 000. Such task was introduced determining ground resolution) is central part of Lithuania (Figure 1) 
identifiable land cover objects by defining the suitability Various ground objects gives a possibility to detect and evaluate 
criterion of objects, extracted from satellite imagery and land cover elements and to perform visual interpretation and 
investigating theirs planimetric accuracy. The vector database comparison within different data sources. The satellite images 
LTDBK was taken as a reference for evaluation of recognizable supplied by Eurimage have a 1G System Corrected level 
and interpretable objects quality from satellite imagery (Ohlhof, (radiometrically and geometrically corrected, including output 
T., 2000) and theirs suitability to update this database. The map projection, image orientation (UTM) and resampling 
different types of data for investigation were used: LANDSAT 7 algorithm (Cubic Convolution), is free from distortions related 
imagery (panchromatic image with 15 m ground resolution and to the sensor, satellite (attitude deviations from nominal) and 
multispectral image with 30 m ground resolution); orthophoto Earth (rotation, curvature). No atmospheric corrections are 
images; digital vector database LTDBK 50 000 (first edition applied for images and do not employ ground control or relief 
according to SPOT imagery) (Sjovall, A., 2001) as well as models to obtain absolute geodetic accuracy. The purpose for 
updated vector database using orthophoto images. The ERDAS usage of satellite image is based on information extraction. On 
Imagine software package have been use for pre-processing of process which includes the determination of spatial object 
satellite images (Ebadi, H., Azizmohammadi, M., 2001). position and land use identification. Considering the goal of 
LANDSAT 7 imagery was converted to vector format that to investigation is more related to object position and geometric 
determine the accuracy of land cover points position by location. 
digitizing in different data sources. 
456 
Internat 
  
Figur 
Other d 
Vector ( 
of 1:50 
compile 
investig 
constru 
resoluti 
LTDBK 
Lithuan 
maps ir 
orthoph 
rectifie( 
sheets. 
image 1 
databas 
GRS80. 
Lo = 24 
investig 
satellite 
3.1 
3.1 Lar 
During 
satellite 
main tc 
does ne 
(Koarai 
After v 
five ty 
(vegeta 
objects 
topogra 
interpre 
respons 
The dis 
using £ 
land cc 
defined 
followt 
- many 
- spect 
cover c 
Mixed 
areas. 
approac