Examples of Geoinformatics research at AGIS/UniBwm

30/11/2010
UdC, Los Angeles, Cl, Nov 2010
Examples of Geoinformatics research at AGIS/UniBwm
Prof. Dr.-Ing. Wolfgang Reinhardt
AGIS GI research group
Institute of Applied Computer Science
AGIS
More infos: www.unibw.de/inf4/professuren/geoinformatik
AGIS: GI-Lab at UniBw M, Chair of Geoinformatics
- civil engineering / computer science faculty
- 10 scientists, mostly financed from 3. Party funds
tasks:
• Education, professional training
• Research
• Research transfer
• Teaching
• Geoinformatics
• Basics in Comp. graphics
• Project Management
Diploma courses (Master equivalent)
PhD education
Specific courses (3 days – 2 weeks)
1
30/11/2010
Overview - Research projects 2006-2010 (3. party funds)
Interoperability / standards /
Services
•
•
Web service testbed
GI Standardisation, Quality procedures /
Concepts for meta data
Portrayal service
•
(funded by Geo-office Bw)
Engineering / sewerage pipes
• 3D capture of small sewery pipes
(industry funds)
Concept for residential sewery pipe
documentation in GIS (bbr industry
funds, coop with Prof. Heister)
•
GI Services / Dec. Support
•
•
Indoor Navigation
• 3D modeling and services
(BMBF funded)
Early warning / land slides
• Simulation / Information System
(DFG/BMBF funded, coop. Prof.
Boley)
GI education
• eduGI – exchange of GI modules (EU
funded)
Quality / Quality Management
• Q-assurance (industry funds)
• Process optimisation / data base
update process (fqs funds)
• …..
3
ASYS : Automated System for 3D acquisition and documentation
of underground sewer networks
•
Development of an automated, sophisticated system capable of 3D
acquisition and documentation of underground residential sewer network
drains.
System concept / description:
•
Integrated multi-sensors (IMU,
Distance measuring units, …) for
3D geometry acquisition
•
Different hardware (controlled
robot/carriage) configurations
•
Software (UI) for controlling the
hardware, data processing and
management, visualisation, and
capable of exporting acquired data
to existing software (GIS, CAD)
Funding:
JT-Elektronik Gmbh, Lindau
Kasseler Entwässerungsbetrieb
Also can acquire drains of small
diameters (100mm, 125mm, etc.)
Contact:
Dr.-Ing. Admire Kandawasvika
admire.kandawasvika@unibw.de
2
30/11/2010
Survey of sewerage system with multi-sensor system
Jörg Renter
Diss. in progress
Investigation goal
•
Survey sewerage system with a combination of 3D camera system and INS
by several methods
Problem definition
•
Interaction / Communication with 3D camera
•
Real time modelling of sewer pipes
•
Innovative miniaturised 3D camera
coupling with INS
Methods/Results
•
Conceptual design of an interface
•
3D modelling
•
Kalman filter
Survey sewerage system with a combination
a of 3D camera system and other methods
5
Context-aware visualization on the Web
Overview
Project for BWGIO, Overview
•
•
•
•
•
•
Iris Wiebrock,
Diss. Submitted to faculty
To be finished 10/2010
Introduction
Motivation for consideration of context
Context model
Adaptation Process
Prototype
Conclusion
3
30/11/2010
Introduction
Introduction
Components
Adaptation Process
Prototype
Web service-based portrayal architecture
Conclusion
Client
Geo
data
Request:
Portrayal Service
Area ”XXX”
Portrayal
Specification „YYY“
Rules
Symbology
IF Feature=AL015
AND ATTRIBUTE bfc=0
or 1
THEN USE
SYMBOL ID=123
Portrayal defined in ISO 19117: „presentation of information to humans”
Introduction
• Current Situation
• Map Services are “data-oriented”
• User of geodata have a different background
Introduction
Components
Adaptation Process
Prototype
Conclusion
(domain, task, …)
• The user demands related to geodata,
representation style and user interface are different
• Depending on “Domain” and task, the necessary
geodata and some parameters of the visualization
normally differ
• Portrayal service is not considering that sufficiently!
• Examples
4
30/11/2010
Introduction
Example: Bike tour (Task)
Components
• Task 1: "Planning and creation of a route"
Adaptation Process
Prototype
 Planning of a route with analysis tools based on different
data like DEM and bike route / track data (with attributes like
“condition”)
 Route creation with classification of bike type (mountain
bike, racing bike), length and duration of the route, and POIs
(sightseeing, parking place, ...), …
Conclusion
• Task 2: “Display of a route during the tour"
 The route should be selected depending on the user criteria
 Display of the route during the tour with background map -> nav.
Task 1
Task 1
Task 2
Example: Bike tour (other conditions)
Introduction
Components
Adaptation Process
• Change of weather condition (e.g., strong rain)
 Influences the condition of the road and an
alternative route will be selected
Prototype
Conclusion
• Change of physical condition (e.g., get tired)
 A shorter route is selected
• Change of light condition from day to night
 Display: Bright geodata view and dark background
[http://www.gpsmagazine.com/assets/review-loox/loox_n100_day_night_map.jpg]
5
30/11/2010
Introduction
Goal of Investigation
Components
• Problem: The user demands for kinds of geodata,
Adaptation Process
Prototype
representation style and user interface are different
Conclusion
• Goal / our approach:
• Portray user group specific geodata views, which are generated by a
•
•
•
web-based portrayal service within a specific domain and with respect to
a given task, needs:
User group categories
Development of a suitable model to consider context
The solution should be standard / and web service-based and extent the
portrayal service (-> Adaptation necessary)
 Concept for context-aware geodata visualization is required
Introduction
Concept / Components
Components
• The adaptation process takes in two input models and
generates one output view
Adaptation Process
Prototype
Conclusion
• Portrayal model (input)
• Context model (input but required)
• Context aware user geodata view (output)
Context aware view
6
30/11/2010
Introduction
Portrayal Model
•
The portrayal model contains default portrayal rules and the
symbols for defining how the data should be displayed
•
Portrayal rule:
•
If the ‘query statement’ is true, the portrayal application like a
portrayal service executes the drawing with the referenced symbol.
Components
Adaptation Process
Prototype
Conclusion
Introduction
Context Model
•
Context is "any information that can be used to characterise the
situation of an entity (person, place, or object)." [Dey 2001]
•
A system is context-aware "if it uses context to provide relevant
information and/or services to the user, where relevancy depends
on the user's task" and should support "presentation of information
and services to a user; automatic execution of a service; and
tagging of context to information for later retrieval". [Dey & Abowd
2000]
•
Many other approaches… see paper
Components
Adaptation Process
Prototype
Conclusion
7
30/11/2010
Introduction
User Groups
Components
Adaptation Process
• Examples of User Groups and their roles from topography
Prototype
and military domain
User Group
Conclusion
Immobile
Field Viewer Topo
Mobile
Analyst
Data / map
provider
*
GIS Analyst Topo
*
Field Viewer C2
(Command and Control
System)
*
*
*
Operational Commander
C2
*
GIS Coordinator C2
*
*
*
Important: Profiles (including relevant data, styles etc.) available
Introduction
Context Model – static and dynamic part
Components
Adaptation Process
• Static:
Prototype
Conclusion
domain
the particular area of activity or interest, like topography or
military
role
specifies the users activities within a certain task
system
the parameters of the device which is used for displaying the
view like resolution, size, band with, …
• Dynamic:
situation
location and time of the user
physical
condition
middle and short term influences caused by weather, day/night
effects etc.
•
•Context modeling in XML Schemas (part of Diss.)
8
30/11/2010
Introduction
Adaptation process
Components
Adaptation Process
Prototype
Adaptation rules:
• Object and application (Data) adaptation
rules
Conclusion
 Influence of the geodata filtering (data -> features)
• Portrayal adaptation rules
 Regulation of the portrayal filtering (portrayal
model -> user portrayal model)
• System adaptation rules
 User geodata view can be influenced by the display
system (e.g., brightness)
Introduction
Adaptation process
Components
Adaptation Process
• Portrayal Pipeline
Prototype
Adaptation rules
Display
Client
Conclusion
User geodata
view
Image
Image
Render
Render
WMS+
Display
Display
Elements
Elements
Context
Model
Style
Adaptation rules
Adaptation rules
Display
Display
Element
Element
Generator
Generator
Features
Features
Filter
Filter
River
Road
Bridge
Data
Data
Source
Source
9
30/11/2010
Introduction
Adaptation process / Implementation
Components
Adaptation Process
• Extension of the OGC WMS operations
GetCapabilities and GetMap
Prototype
Conclusion
 "GetCapabilities"-Request: Extension of
"ContextType" element and the new WMS+ operations
 "GetMap"-Request: Add a "context" parameter
• Context documents on the WMS+ server (context id),
on an external server (URL reference) or as "Body" in
the WMS+ requests (BODY_element)
• Two new WMS+ operations are introduced
 "DescribeContextType"-Request: Get context schema
 "GetContext"-Request: Get context data
 "Transaction"-Request: Insert, update and delete
operation for context elements, the user can submit and
update the context information
Introduction
Prototype
• The figure shows built-up areas in general for an overview.
Components
Adaptation Process
Prototype
Conclusion
• The figure shows a land-use plan in more detail with industrial and
residential built-up areas, grassland, and transportation information
like streets.
10
30/11/2010
Concepts for Metadata (BGIO funded project)
• Definitions in standards of ISO TC211
• data about data
• information about data
• INSPIRE Directive
• information describing spatial data sets and spatial data services and
making it possible to discover, inventory and use them
• Extended definition
• all needed information to describe spatial data sets and services
without a direct access to them.
21
Usage of Data and Metadata
Metadata
Request
Response
Metadata
Spatial Data
22
11
30/11/2010
How is metadata generated?
1. Manual collecting in a Metadata Entry Tool
•
•
•
The common way to collect metadata
Great effort and additional work
The huge amount of metadata enable a wide description of a resource
2. Automatic generation by accessing spatial data and sources
•
The quantity of metadata depends on the data format
3. Automated generation by monitoring the production process
of spatial data
•
A controlled environment is needed
Meta data editor
Developed for
BGIO
for various data
products
23
Which Metadata should be generated?
• The purpose of metadata is to inform a user
 metadata should be user oriented
• Knowledge about the user needs related to metadata only
slightly available
 all metadata should be generated
• To decrease “all” in a reasonable quantity
 all metadata of ISO standards should be generated
But the standardized metadata are only an approach to fulfill
the user requirements
24
12
30/11/2010
Typical user request
Background
• An ongoing process is the development of a NATO metadata
profile
• The first draft version of the profile was created as a subset of
ISO 19115
• Typical user request (freetext) were defined by NATO
• There was an analysis if the typical user request can be
answer with the metadata elements of the first draft of NATO
metadata profile
25
Additional information to answer a request
Typical user request:
I need datasets which describe main supply routes of ISAF RC East.
To answer this request with common metadata some additional
information is needed.
Questions:
• Which features contain main supply routes (e.g. roads, bridges, tunnels)?
• Is an attribute named supply classification needed?
• Which area contains ISAF RC East?
Missing information:
• The semantic relation between ”main supply routes” and the
feature types or keywords must be defined
• The area of ISAF RC East must be identified
26
13
30/11/2010
Transformation to answer a request
Typical user request:
I need all obstacles higher than 30 m in the surrounding of 10 km
from Kandahar air field.
The request has to be transformed in a “metadata” request
e.g.: Give me all datasets with:
• Features = rocks, trees and buildings
• Attribute = height
• Attribute accuracy of height = 5 m
• Geographical extend of Kandahar air field + Buffer of 10 km
27
Typical user request (examples)
In what area is no data on rivers available? (different to the
question in which area are no rivers)
I need a 3D urban model of Kabul with textures of all buildings.
• There is no element in ISO 19115 for textures of buildings
28
14
30/11/2010
Typical user request (examples of analysis)
• If the system tells me that snipers cannot see me, can I trust
that information?
• Is the route really the shortest route?
• I need to transport oversized equipment. Can I rely on the
results of the routing analysis?
 Open question: How can I describe the reliability of an
analysis, because it depends not only on the spatial data?
29
Results
• Metadata should be user oriented but often collected in a
production view
• There is a discrepancy between the production view and the
user requirements
• Users are different in:
• Knowledge (expert – non expert)
• Domain (civilian – military)
• Tasks (customer – producer)
It is an ongoing research to solve the problem to create user
oriented metadata.
30
15
30/11/2010
Dissertation Thorsten Bockmühl (in progress)
„Improvement of metadatamangement“
extract
metadata capturing
producer
requirements
edit/complement
validate
store
metadata management
filtering
rendering
access
metadata application
adopt
apply
contribution
provide
requirements
Metadata life
transfer
end user
31
Early warning for landslides
Overview
•
•
•
•
•
•
Introduction
Key technologies
Sensors and networks
GIS combined with Numerical Simulation
Processing of the simulation results (cluster analysis)
Conclusion
16
30/11/2010
Geotechnologies program
Early Warning Systems Against Natural Hazards,
projects, goals:
LASTMILE –> Tsunami Early Warning (that was the starting point! early
warning in general can make catastrophe management unnecessary
EGIFF -> Early Warning Systems for landslides
EXUPERY -> Volcano Fast Response System
RAPID -> Rapid Automatic Determination of Seismic Source Parameters
EDIM - Earthquake Disaster Information Systems
ILEWS - Integrative Early Warning Systems for Landslides
EWS TRANSPORT - Early Warning Systems for transport lines
G-SEIS - GPS Stress Evaluation within Seconds (Seismic)
alpEWAS - integrative 3D early warning system for alpine instabile slopes
SLEWS - infrastructure for early warning systems for landslides
Introduction
• Focus on landslides
• The problem of Early Warning in landslides today:
 Normally based on direct measurements,
example:
 Radio transmission, if threshold value exceeded
 Reaction time, e.g. 30 minutes!
 Research goal: earlier (and saver) warning!
17
30/11/2010
Factors
• The problem: many factors have influence on the slope
stability:
 Geometry (slope)
 Land use
 Geohydrology
 Geology
 Soil mechanics
 Climate /
 Metrology (rainfall)
…
 Various approaches to improve early warning, examples:
 Sensors
 Combination of sensors / sensor networks / infrastructures
 Combination of sensor measurements and numerical simulations
…
Key Technologies
•
•
•
•
•
•
Sensors
Sensor networks
Infrastructures
Geographic Information systems (geographic data, methods)
Numerical simulations
Decision support
Also important:
• Structured procedures / processes / methods / resposibilities / for early
warning
• Risk Management (which risk is acceptable, forecast of damages ..)
• Regulations by public law
 not in focus in this presentation
18
30/11/2010
Sensors - overview
Src.: A. Kandawasvika, PhD thesis
Sensors for landslide detection (examples)
Examples of High Cost to Low Cost GPS receivers
Examples of total stations / laser scanners
And many others!
outer conductor
isolator
inner conductor
Extensometers (wire, tape or rod)
Time Domain Reflectometers (TDR)
An example for sensor based EW-systems can be found at:
(http://www.alpewas.de/)
19
30/11/2010
Geographic Information Systems
• Well known technology:
• A GIS is a system of hardware, software
•
•
and procedures to facilitate the
management, manipulation, analysis,
modelling, representation and display of
georeferenced data to solve complex
problems regarding planning and
management of resources (NCGIA, 1990)
Can be used for many purposes in
EW and CM
Examples on the next slides
Automated Hazard Maps based on GIS
• Susceptability can be
calculated based on GIS
(various topics)
• Different models of calculation
available
• To be used as an indicator
• High susceptible regions have
to be observed by sensors
and / or investigated by
numerical methods!
Example from EGIFF Project (http://herakles.fzi.de:81/wiki/index.php/Hauptseite)
20
30/11/2010
Geographic Information Systems
• Provides various methods to analyse geographic data, also in
3D
Digital terrain model
Digital underground model (Geology)
Src.: http://www.egiff.uos.de/dokuwiki/egiff:projektseite
Coupling GIS and numerical simulation
New approach, the usage of simulation systems (SIMS)
predominately only by experts and scientists, Currently no
availability for disaster prevention and management
Goals / benefits of Coupling GIS and numerical
simulation (SIMS)
• Sophisticated data management through the GIS
 Data input
 Data storage and administration
 Presentation of results
 Support of the interpretation of the simulation results, decision
support
 User-friendly and intuitive control of the complete system
• Numerical calculation of the slope stability!
More info: http://www.unibw.de/bauv11/geoinformatik-en/forschung/projekte/slide
21
30/11/2010
Study Area „Isarhänge Grünwald“
Study Area
Available data:
• digital elevation model (2m x 2m)
• geodetic surveys
• orthophotos
• digital topographical data
• groundwater level measurements, discharge rate
• inclinometer measurements
• extensometer measurements
• dynamic probing tests
• drilling profiles (borehole data)
• geotechnical material parameters
(laboratory test results)
Multi-step data transfer between GIS and SIMS
Geometrical information:
Non-geometrical data:
• Digital terrain model and 3D-ground
• Material parameters and
•
model
FE-mesh generator
•
•
•
constitutive equations
Boundary conditions
Impacting loads
Calculation steps
GIS
Input File
Wire
Frame
Model
FE-Mesh
Slope
Geometry
nongeometrical
Data
Output File
Results
Project in cooperation
with IGRB, UniBw
(Prof. Boley)
Simulation System
22
30/11/2010
Simulation of a break
Movements for preset loads
Result of a simulation
FE-Network
GIS
“Very many shift vectors” describing the movement of the nodes
-> Goal: find the areas in danger to break
23
30/11/2010
Processing of the simualtion results in the GIS
Calculation of areas of different danger levels
(classes)
 Cluster analyse methods
Determination of the sliding body
Vectors before classification (schematic)
Cluster analysis
• Evaluation of several methods, best results: Bivariate Cluster analysis and
separate neigbourhood consideration (adaption of the method)
d  (a  b )²  (a  b )²  ...  (a  b )² 
1
1
2
2
n
n
( L  L )²  ( R  R )²
1
Bivariate Cluster analysis (direction and length of
vectors)
2
1
2
Eva Nuhn
Diss. in progress
Neighbourhood consideration
24
30/11/2010
Preparation for Decision Support
More details:
http://www.unibw.de/bauv11/geoinformatik-en/forschung/projekte/slide
Decision support
• Decision is normally made by a politician
• Complex simulation results have to be translated to simple
ones (but considering uncertainty of information)
25
30/11/2010
Conclusion / Further work
• Early warning is most important for preventing catastrophes
• Up-to-date early warning systems for landslides (very often)
allow for a too short reaction time
• Our work has shown good results for certain types of soils and
slopes - > research with other types necessary
• Several research activities based on new sensors, sensor
networks and numerical simulations are promising and give
hope that we can come to a really Early Warning
Examples of other dissertation work
52
26
30/11/2010
Quality of Geoinformation
•
Quality:: “fitness for use”
• Quality elements (ISO 19113)
 “completeness”,
 “logical consistency”,
 “accuracy (positional, temporal, thematic)”
• Quality Management (consideration of the production process)
„75% of all errors are caused by mistakes in the product definition phase!, but are
detected when the product seems to be finished.“ (Adams, Rademacher, 1994)
 This is also true for the production process of Geodata
 QM for geodata production process
Issues of GI Quality research (enquiry, conferences, 2008)
Quality is important and
treated at many conferences / papers, e.g. dealing with:
• Modeling uncertainty and vagueness
• Processing of uncertain and vague data
• Spatial relations for uncertain and vague objects
• Aspects of Error Propagation
• Logical consistency, Database Integrity, Integrity Constraints
• Semantic aspects of uncertainty
• Quality of Digital Elevation Models
• Quality aspects in Data Mining
• Quality management
• New metadata types
• Communicating metadata to users
27
30/11/2010
Integrity constraints
• General
General semantic integrity constraints are more complex
explicitly
specified
during
modelling
express
• and
domain
constraints
restrict
the allowed
typesand
of values
of an
application
attribute logics, rules or even laws, Examples:
• Roads/railroads and rivers can only be connected through a
• keybridge
and relationship
constraints refer to the possibility to
or a tunnel
define key values for entity classes, cardinalities for
• A house connecting pipe must be connected to a main pipe
relationships between entity classes and participation
•requirements
A watermill is always connected to a river
• If you can turn from street a to street b they have to be
• Well
known from data bases, not further treated here
topologically connected
• A petrol station must be at least 500 meter away from a school
• Sewery sludge only spread on certain parcels….
Integrity constraints - Modelling
General
• For conceptual modeling in GIS UML is used
Class 1
attributes
methods
relation
Class 2
attributes
methods
• UML offers limited possibilities to express complex integrity constraints
Main research topics:
• Types of constraints? Unique treatment? ...
• How to detect conflicts, redundancies
•• Class
level they
/ instance
levelimportant
(1:1, 1:n, but
n:mare
, all1:all2)
In practice
are pretty
defined separately from the
data model and implemented in a proprietary environment.
•• Modelling
these
complex
constraints?
There areoftools
and
languages
(OCL, SWLR …) which provides better
• Are
all
constraints
strict?
Consideration
severity
levels necessary?
possibilities, but how to integrate in the of
whole
processes?
• How to integrate in relevant processes? What to do if constraints are violated?
• ....
Examples from own research
28
30/11/2010
Integrity constraints - taxonomy
Taxonomy of integrity constraints (src.: Mäs and Reinhardt, 2008)
Checking Consistency of Semantic Integrity Constraints
Methods / Results:
•
•
Stephan Mäs
Finished 12/2009
Definition of cardinality properties of class relations
Investigation of logical characteristics of these properties (symetry, composition)
Some examples:
Overlaps RDLT (Watermill, Stream) -> every watermill at a Stream, but just at one
58
29
30/11/2010
Checking Consistency of Semantic Integrity Constraints
Methods / Results:
•
•
•
Investigation of the combinations of class relations
Reasoning methods
Prototyp implementation (to detect conflicts/redundancies)
Stephan Mäs
Finished 12/2009
examples:
59
Extended conceptual data modeling
Dissertation Wang Fei (finished 08/2008):
Extension of the model, constraints in a CDT
It bases on Event Condition Action (ECA) rule which has semantics:
“When an event occurs, check the condition and if the condition is
satisfied, then execute the actions”
Definition of severity, and “to do” – rules if the constraint is violated
Transfer to other process steps like data acquisition in a standardized way (GML
extension)
Tested in a landslide application
CDT: constraint decision table
30
30/11/2010
Integration of heterogeneous multi-sensors in geoscientific
monitoring applications
Admire Kandawasvika
Finished 09/2009
Investigation goal:
• Concept for disparate multi-sensors integration based on interoperable, open
standards and service oriented architectures (SOAs).
Problem definition:
• Management of heterogeneous sensors in a single application (e.g. different
communication protocols and data interfaces,…)
• Time-consuming, costly and difficult integrating new sensors into existing legacy
systems
• Existing sensor-based
frameworks not easily
implementable (e.g. global
in nature or too generic, …)
Methods/Results:
• Definition of application scenario (e.g. landslide monitoring); sensor requirements
• Analysis of different sensor specs., modeling; and analysis of existing sensorbased frameworks
• Conceptual framework: domain-specific and based on interoperable solutions
61
Conclusion
I hope it was better understandable than this letter …
31
30/11/2010
Thank you!
questions?
Contact
Univ.-Prof. Dr.-Ing. Wolfgang Reinhardt
AGIS / Institut für Angewandte Informatik
Universität der Bundeswehr München
D-85577 Neubiberg
Telefon +49 (0)89 6004-2450
Telefax +49 (0)89 6004-3906
Wolfgang.Reinhardt@unibw.de
http://www.unibw.de/bauv11/geoinformatik
32