CONTROL PROCEDURE FOR TOPOGRAPHIC MAPDATABASE SETUP AND UPDATE 
J. Kamper 
National Survey and Cadastre - Denmark, Rentemestervej 8, DK 2400 Copenhagen NV, Denmark, jk@kms.dk 
KEY WORDS: Mapping, Databases, Quality, Updating, Geometry, Accuracy, Orthorectification, Semi-automation 
ABSTRACT: 
In this paper, strategies for development and evaluation of data control processes to database setup and update are discussed. 
Examples of production-workflow and control procedures used for the national topographic map-database of Denmark are presented. 
Map-objects to the database are collected and extracted photogrammetrically by subcontractors and controlled by in-house 
procedures. In the database setup phase data control was focused on data-specification, data-completeness and data-topology. 
Geometric accuracy was then controlled by large-scale photos in random test-areas. In the database update phase, efforts are now to 
ensure that all single map-objects are geometrically correct. The control software used has been extended to support highly accurate 
verification of map-object geometry, by Vector Transformation. The Vector Transformation methodology is build on fundamental 
equations of photogrammetry and is designed to work on PCs without stereoscopic view. It projects new vector data 
photogrammetrically on top of the unrectified image, accounting for inner orientation and rotation, by use of camera properties and 
already verified map-objects in X, Y and Z. The paper presents theory, workflow, features and experiences about Vector 
Transformation. To fulfill a need of verifying large areas of updates with little manpower, effort in making user interaction as 
automated and time saving as possible has been made. The paper points out elements in the workflow that are especially facilitated 
by automation. Experiences suggest that verification of map-data in 3D can be done with good results in a 2D environment, by a 
software solution offering functions as the presented. 
1. INTRODUCTION The case presented in this paper, is part of the production 
workflow for the setup, update and control of the nationwide 
This paper deals with data quality and the need of elaborating topographic map-database of Denmark. An overview of its 
the correct level of data control, at all times, through set up data control processes, as part of the production workflow, 
and updating of a database. To secure data quality and to will be given, with focus on the object-geometry control. 
define a control procedure in the case of database updates in 
general, one must be aware of both quality of data already in In the map-database setup phase the control process for 
a database, and of data to update it. This points out the geometric accuracy was in two dimensions, which was 
essential need for a continuous evaluation and knowledge of: extended to include the third dimension in the database 
- Quality of data currently in the database update phase. Object geometries are extracted from aerial 
- Quality of data-source and of current data-collection photos in 3D by photogrammetric methods. A methodology 
- Quality of data-extraction methodologies and of current for control of objects geometries with 3D accuracy in a 2D 
data-extraction environment is presented. and results and gains from 
- Quality and availability of both technology and implementing this methodology is described. 
manpower for data control 
The overall data quality of a database, can not reach a better 
quality than the least satisfactory of these evaluation points 2. DATABASE SETUP, UPDATE AND CONTROL 
described. Nevertheless, control procedures often give focus 
to the need of high quality data-collection and/or high quality The Danish national topographic map-database in 1:10000 is 
data-extraction, giving little attention to data-quality of the named TOPIODK (Frederiksen et. al., 2004). It was designed 
database content and to quality of control procedures. in the 1990's and in 1995 data from the first part of Denmark 
was put in the database and since 2000 TOPIODK covers all 
One goal of this paper is to point out the need for accepted Denmark. The map-database is owned, maintained and 
long-term control procedures for database updates, and the distributed by National Survey and Cadastre — Denmark 
need for a continuous evaluation of data control processes. (KMS). 
The paper argues that, in respect of data quality, control 
should be neither “ad hoc assignments” nor “static An overview of improvements in data-control flow through 
processes”. Control procedure should be planned, with the production period of the TOPIODK will be described 
scheduled points of elaboration and optimization of the according to the four evaluation points described in the 
control processes. In addition, it is shown that focus on data- introduction. This is aerial photographing as the “data- 
control does not necessarily result in more time and cost collection” process, stereoscopic object extraction as “data- 
demanding workflows. It is pointed out though, that having extraction” and in-house developed and performed processes 
continuous change in control procedures makes it difficult to for "data-control". Different assumptions according to the 
ensure and describe the overall data-quality of a database. setup or update production phases are also pointed out. 
Interne 
2.1 D: 
The de 
setup | 
1:250€ 
20% s 
proces 
when 
are no 
Denm: 
detaile 
When 
aerial 
specifi 
2000, 
specifi 
to be 
exposu 
end-qu 
Contac 
produc 
will be 
2.2 D: 
Extract 
photog 
databa: 
other | 
calibra 
parame 
returne 
For the 
new d 
parame 
area - t 
for the 
Map-ol 
to rules 
new U] 
schedu 
object 
into : 
specifi 
Physic: 
compat 
contrac 
databas 
2.3 Da 
A nun 
product 
databas 
differei 
which i 
by both 
are exe