International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B4, 2012
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia
MULTITEMPORAL QUALITY ASSESSMENT OF GRASSLAND AND CROPLAND
OBJECTS OF A TOPOGRAPHIC DATASET
P. Helmholz *^!, T. Büschenfeld", U. Breitkopf*, S. Müller?, F. Rottensteiner*
? [PI — Institute für Photogrammetrie und GeoInformation, Leibniz Universität Hannover,
Nienburger Str. 1, 30167 Hannover, Germany; (helmholz, breitkopf, mueller, rottensteiner)@ipi.uni-hannover.de,
? Department of Spatial Sciences, Curtin University of Technology,
GPO Box U1987, Perth WA 6845, Australia; p.helmholz@curtin.edu.au
* TNT - Institut für Informationsverarbeitung, Leibniz Universität Hannover,
Appelstr. 9a, 30167 Hannover, Germany; bfeld @tnt.uni-hannover.de
Commission VI, WG VI/2
KEY WORDS: Automation, Quality, Inspection, Updating, GIS, Crop, Multitemporal, Classification
ABSTRACT:
As a consequence of the wide-spread application of digital geo-data in geographic information systems (GIS), quality control has
become increasingly important to enhance the usefulness of the data. For economic reasons a high degree of automation is required
for the quality control process. This goal can be achieved by automatic image analysis techniques. An example of how this can be
achieved in the context of quality assessment of cropland and grassland GIS objects is given in this paper. The quality assessment of
these objects of a topographic dataset is carried out based on multi-temporal information. The multi-temporal approach combines the
channels of all available images as a multilayer image and applies a pixel-based SVM-classification. In this way multispectral as well
as multi-temporal information is processed in parallel. The features used for the classification consist of spectral, textural (Haralick
features) and structural (features derived from a semi-variogram) features. After the SVM-classification, the pixel-based result is
mapped to the GIS-objects. Finally, a simple ruled-based approach is used in order to verify the objects of a GIS database. The
approach was tested using a multi-temporal data set consisting of one 5-channel RapidEye image (GSD 5m) and two 3-channel
Disaster Monitoring Constellation (DMC) images (GSD 32m). All images were taken within one year. The results show that by
using our approach, quality control of GIS- cropland and grassland objects is possible and the human operator saves time using our
approach compared to a completely manual quality assessment.
1. INTRODUCTION After reviewing related work in section 2, we will present our
method for the quality assessment of cropland and grassland of
a GIS data set with respect to the thematic accuracy. The
thematic accuracy is the percentage of correct objects in the
GIS-database. In our approach we verify GIS cropland and
grassland objects automatically comparing them with the real
world in the form of remotely sensed images. Input data into the
system are up-to-date multi-temporal satellite images taken in
one year and a GIS which has to be verified. The system verifies
the GIS-objects using automatic image analysis approaches
introduced in section 3. The result of the automatic comparison
The basic methodology to represent the real world in a GIS is to of the GIS-objects and the images is the decision whether an
define objects using a data model (e.g. a feature type catalogue object in the GIS data set is correct (accepted from the system;
also called GIS-object catalogue) which defines the objects to labelled green) or incorrect (rejected from the system; labelled
be contained, as well as their properties and structure. In the red). The results of the automatic procedures are passed on to a
European Norm (DIN EN ISO 8402, 1995), quality is defined human operator. All the accepted objects do not have to be
Today, many public and private decisions rely on geospatial
information. Geospatial data are stored and managed in
geographic information systems (GIS). In order for a GIS to be
generally accepted, the underlying data need to be consistent
and up-to-date. As a consequence, quality control has become
increasingly important. A high degree of automation is required
in order to make quality control efficient enough for practical
application. This goal can be achieved by automatic image
analysis techniques.
as the “Degree to which a set of inherent characteristics fulfils reviewed, while for all rejected objects an interactive check by
requirements”. In the context of GIS this means that, first, the the human operator is necessary. The human operator saves
data model must represent the real world with sufficient detail time using our system for quality assessment because an
and without any contradictions (quality of the model). Second, interactive check of all GIS-object accepted by the system is not
the data must conform to the model specification (quality of the necessary anymore. We call this efficiency time efficiency.
data). There are five important measures for the quality of geo- Because the final decision of rejected GIS-objects is done by
data: logical consistency, completeness, positional accuracy, the human operator, our approach is a semi-automatic one.
temporal accuracy and thematic accuracy (EN ISO 9000:2005, However, given the fact that quality assessment is essentially
1995). Only the consistency can be checked without any carried out to remove errors in the GIS data base, classification
comparison of the data to the real world. The other quality errors from analysing the satellite images have to be avoided
measures can be derived by comparing the GIS data to the real because these errors can lead to undetected errors remaining in
world as it is represented in satellite images. We call this step the GIS database. The main goal of our approach is to achieve a
verification or quality assessment (Gerke and Heipke, 2008). certain thematic accuracy of the GIS database after the
verification process. To evaluate our approach and to prove that
! Corresponding author
64
