IQmulus projekt: a téradatok felhasználásának új dimenziója magyar részvétellel

„IQmulus” projekt: a téradatok felhasználásának új dimenziója – magyar részvétellel Dr. Kristóf Dániel

09_05_18_Vorlage.ppt

FP7 Információs Nap, Nemzeti Innovációs Hivatal, 2012.09.17.

TÉKA Tájérték Kataszter Rendszer, szakmai műhelyvita Földmérési és Távérzékelési Intézet Budapesti Corvinus Egyetem, 2012. szeptember 15.

© 2012 SINTEF

1


Előzmények • A Földmérési és Távérzékelési Intézet (FÖMI) a kormányzat térinformatikai háttérintézménye, jogelődje 1967-ben jött létre, a Vidékfejlesztési Minisztérium irányításával működik • Aktívan és sikeresen részt veszünk nemzetközi K+F+I pályázatokban a térinformatika, távérzékelés és téradatinfrastruktúrák területén • Közvetlen előzmény: HUMBOLDT projekt (FP6 IP, 20062011) • Az akkori konzorcium-vezető intézmény (Fraunhofer IGD) kereste meg Intézetünket az új pályázat ötletével IQmulus: a téradatok felhasználásának új dimenziója FP7 Információs Nap  NIH, Budapest, 2012. szeptember 17.

2 © 2012 SINTEF

2


Előkészítés • A pályázat nyers ötletét először 2011. áprilisában vetették fel, a FÖMI-t a kezdetektől bevonták az előkészítésbe • Az ötlet alapján szakmai-ismeretségi alapon, a pályázati téma és elvárások figyelembevételével indult meg a konzorcium-építés, az ötlet finomítása • Email- és telefon-kommunikáció, majd 2011. novemberében többnapos, intenzív műhelymunka során épült és finomodott a pályázati elképzelés • Határidő: 2012. január 17, 17:00 – intenzív online munka és telefonos kapcsolat • Pályázat beadva: 2012. január 17, 16:48:59… • Bizottsági meghallgatás: 2012. március 22. • Tárgyalási meghívás: 2012. május 3. • Első tárgyalási nap: 2012. május 24. IQmulus: a téradatok felhasználásának új dimenziója FP7 Információs Nap  NIH, Budapest, 2012. szeptember 17.

3 © 2012 SINTEF

3

Jelenlegi helyzet


• A tárgyalási szakasz véget ért • A Bizottság előzetesen jóváhagyta a munkatervet és a támogatási megállapodást (Grant Agreement) • A konzorciumi megállapodás (Consortium Agreement) partnerek között egyeztetett verziója kész • Várhatóan néhány héten belül aláírják a támogatási szerződést • Kezdődátum: 2012. november 1., végdátum: 2016. október 31.

IQmulus: a téradatok felhasználásának új dimenziója FP7 Információs Nap  NIH, Budapest, 2012. szeptember 17.

4 © 2012 SINTEF

4


Jelenlegi helyzet


5 © 2012 SINTEF

5


: A High-volume Fusion and Analysis Platform for Geospatial Point Clouds, Coverages and Volumetric Data Sets

IQmulus will leverage the information hidden in large heterogeneous geospatial data sets and make them a practical choice to support reliable decision making

© 2012 SINTEF

6

Scenario and motivation  Current situation 



Information in large geospatial datasets exists but is not integrated in the decision process Often only accessed for damage assessment, „what went wrong?” analysis

 Objectives 



Two test cases 


Make information from large geospatial datasets available on time, at the relevant/right level of decision making



Flood Management and Marine Spatial Planning Economic and social importance to EU

Ajka, October 2010 © 2012 SINTEF

7

IQmulus empowers users

SONAR Seamless DEM

LIDAR


Overlaping datasets in time and space

Point cloud data 5 data users

+ Intelligent Processing =500 User Information data users

These numbers based on HR Wallingford operating data for the SeaZone business in UK © 2012 SINTEF

8

The Core Innovation  New methodology for fusion and analysis of geospatial data:  Independent of data modelling paradigm 

Not bound to predefined data partitioning



Allows expressing basic correlation patterns, advanced analysis and knowledge discovery algorithms



Managing uncertainty

 Requires a new technical approach to the definition, configuration and deployment of functional spatial processing services  Independent of data size and execution architecture




Distributed processing of heterogeneous geospatial data

 Allows new organisational processes taking into account archive and current in-situ data to cope with extraordinary scenarios

© 2012 SINTEF

9

Architecture and Scalability Challenges

WP2, WP3

Vertical Scalability:

Partitioning Data Reduction


Horizontal Scalability: Distribution

© 2012 SINTEF

10

Data processing: challenges & approach WP4: core reserach  High degree of heterogeneity: multi-source data, raw data, spatial and/or temporal resolution, spectral dimension, dimensionality, ... 

today: heterogeneity is handled on few of these aspects only



IQmulus: handling several aspects at the same time (uni-variate and multivariate data sets); matching of semantic information at different scales to drive iterative data fusion/integration

 Uncertainty and its management: noise, incompleteness, instrument precision, processing/approximation errors 

today: global accuracy information, “simple” confidence maps



IQmulus: provenance and thourough quality documentation; fuzzy/probabilistic methods for associating confidence values to integrated data; robust classifiers


 Large data volumes 

algorithmic alternatives for the same task to exploit different parallel architectures

© 2012 SINTEF

11

Data processing: evaluation measures

WP4 Core research

 Uncertainty management, quantitative/qualitative measures  



metadata for the documentation of lineage and reliability of data and integrated models classification performances  cross-validation, auto-qualification of generated data, in-situ validation  Comparing formal errors (from error propagation) User satisfaction in the visualization of uncertainty in decision making

 Processing tools, quantitative performance measures: benchmark


definition for the evaluation of 

gain in processing time, both in theory (complexity) and in practice (execution time)



robustness: how do the results change if (artificial) outliers are added



sensitivity: how do the results change if the settings/control parameters are modified

© 2012 SINTEF

12

Processing and Human Intervention 

Consolidating and harmonizing input data 






WP4

exploitation of semantic information to reduce the need of expert parameter setting to harmonize data sets

Semantic enrichment 

automatic extraction of high-level information by segmentation or classification processes



automatic detection and characterization of changes and dynamic events, for instance, sand dunes and landslides

Methods 

geometry-oriented methods: exploit shape analysis techniques for point clouds (eg, fitting primitives, variational analysis)



machine learning and statistical analysis



combined reasoning to extract features according to rules and context

© 2012 SINTEF

13

Interactive Visual Decision Support

WP5 Visualization

Leveraging modern GPU features on graphics machines and web-clients  single common view for 2D / 3D / nD vector and 2D raster data 

enabling visual communication on large heterogeneous geospatial and temporal data sets

 deferred mapping and deferred fetching 

better visual quality, faster display

 direct manipulation based on data from "deferred processing“ 

joy of use, improved interactivity: speed-up factor ~ 20 on avg.


 3D-Web-based visualization (decision makers, laypersons) 

dedicated apps (ease of use) / device independence / on-site presentation



hybrid rendering, compression and streaming: up to a factor of 30

 visualization-driven data formats 

research, analysis and design of optimally tailored data formats for GPU-based interactive visualization -> future standardisation activities © 2012 SINTEF

14

User engagement 

Expert users, decision makers



Evidence to address user needs 

Users as partners



Core Users (per test bed: start with ca. 6

WP 1, 5, 7, 8

and go up to 20) – regular user workshops 

User Group (approx. 500 persons) –

IQmulus workshops at international events 

When users get access 




involved right from the beginning in user requirement task (WP1) Evaluation of prototype components M13-M18 (WP7)



first user evaluation phase M25-M30 (WP7)



second user evaluation phase M43-M48 (WP7) © 2012 SINTEF

15

Markets & Impact

WP 9

 Markets


 Competitive Environment 

Mainly US-based industries

 Strengthening the role of European industry in the converging market of remote sensing and GIS © 2012 SINTEF

16

European Dimension & Impact  IQmulus an Integrating European Project 

Data providers / users represent diversity in Europe



Hand-picked experts / expertise for implementing the IQmulus infrastructure and services ...



...allowing European organisations to fully exploit data for commercial and political benefit

 Impact on Europe 

Increase Competitiveness of SME sized European geospatial industry; a single SME cannot provide a solution in sufficient time-to-market, but many can build on the IQmulus infrastructure.

“Because EU Policy makers at all levels (local to European) An operational IQmulus will supports EU Knowledge Economy, especially in sectors such marine and offshore engineering which the EUthat is a major world player. are convinced better IQmulus will support European policy implementation (e.g. SEIS, GMES, Floods decisions need better and Directive) by providing a homogeneous and transparent view on European topography more timely information” and its dynamics 








Leveraging on EU track record in image and point cloud processing over 20 years

Basis for improved and advanced data licensing and re-use models under the „data as service‟ paradigm.

© 2012 SINTEF

17

The Consortium  Basic research in geo-spatial information processing 

University College London (UCL): research centre for photogrammetry, 3D imaging and metrology



TU Delft: Department of Earth and Climate Sciences, Man-Machine Interaction Group (usability)



Université de Brest (UBO): European Institute for Marine Studies)

 Applied research





SINTEF - Institute of Information and Communications Technology: 3D information processing, heterogeneous computing and visualization:



CNR-IMATI - Institute of Applied Mathematics and Information Technologies of the National Research Council of Italy: semantic information enrichment



Fraunhofer - Institute for Computer Graphics Research: geo-information management, geo-data harmonization, interactive visualization and visual decision support

GIS industry (SME) 

M.O.S.S. Computer Grafik Systeme (MOSS) : spatial-based information and technology solutions


 National organizations 

French National Mapping Agency (Institut national de l‟information géographique et forestière (IGN)



l‟Institut Francais de Recherche pour l´Exploitation de la Mer (Ifremer)



Hungarian National Mapping and Cadastral Agency (FOMI): Institute of Geodesy, Cartography and Remote Sensing



HR Wallingford (HRW):UK and worldwide customers)

 Regional organization 

Region of Liguria (Italy) © 2012 SINTEF

18

Institute of Geodesy, Cartography and Remote Sensing

 Hungarian National Mapping and Cadastral Agency 

Established in 1967

 Provider and/or custodian of country-wide reference geodata 

e.g.: orthoimagery, large-scale topography, cadastre, administrative units, land cover, land parcel identification

 Geo-information hub of the Hungarian public sector 

Strong professional links with governmental institutions

 Involved in numerous national and international R&D projects and cooperations Leading several significant national geospatial R&D projects



Contributor in FP6-FP7 and eContent+ projects



Active participation in GMES and INSPIRE-related activities




© 2012 SINTEF

19

Institute of Geodesy, Cartography and Remote Sensing

Main role and tasks in IQmulus:  „Data provider”, „data integrator”, „user” and „multiplier” at the same time  Coordinating the „Land” application scenario 

In cooperation with other partners involved in the scenario



By involving other public sector institutions as core users and in the user group

 Lead of WP1 (Requirements) 

Bringing in the user perspective from the beginning



Cooperating with scientific partners in the State of the Art analysis


 Additional significant contributions 

Processing (WP4)



Testing and validation (WP7)



Dissemination (WP8)

© 2012 SINTEF

20


The Work Package Structure

© 2012 SINTEF

21

© 2012 SINTEF 22


Köszönöm a figyelmet! További információ:


Dr. Kristóf Dániel osztályvezető Földmérési és Távérzékelési Intézet Térinformatikai Osztály 1149 Budapest, Bosnyák tér 5. telefon: (1) 460 4090 mobil: (20) 341 7079 e-mail: [email protected] honlap: www.fomi.hu


28 © 2012 SINTEF

23

IQmulus projekt: a téradatok felhasználásának új dimenziója magyar részvétellel

Recommend Documents