Download presentation
Presentation is loading. Please wait.
Published byAntonia Melinda Booker Modified over 7 years ago
1
Linking Big Data from Space to Apps on Earth
Location Powers: Big Linked Geodata workshop Colocated with the 102nd OGC Technical Committee Delft, The Netherlands Uwe Voges (con terra GmbH, 22 March 2017 Copyright © 2017 Open Geospatial Consortium
2
Copyright © 2017 Open Geospatial Consortium
Problems Start summarizing problems of end-user / app- developer less experienced with Geo services wanting to find and use EO data Finding EO data (ideally as online service) via Google & Co not easy: Result: general info, news, pictures (NASA),… Needs to search specific EO portals EO data often not online -> stored in archives user has to regis- ter -> try to find, order, download the data Copyright © 2017 Open Geospatial Consortium
3
Copyright © 2017 Open Geospatial Consortium
Problems Data delivered as huge datasets, requiring preprocessing,… If online services exist (e.g. OGC based view, download, MD) … client has to deal with different URI´s, offering different service interfaces need to construct specific service calls client cannot just “follow links” (-> HATEOAS) Data / services described by specific metadata models (ISO, O&M..) not providing common encodings, not using common vocabularies objects often re-modeled instead of linking to existing representations prevents (meta)data being part of a wider network cannot be indexed by Google & Co or Open Data Portals Copyright © 2017 Open Geospatial Consortium
4
Copyright © 2017 Open Geospatial Consortium
Targets Reliable, performant and easy web accessible EO data services Easy REST-based data-access and visualization services Download of Online Data in different formats (incl metadata) ideally cloud based Example going in this direction: AGOL providing REST-based ArcGIS Image Services (also providing WMS/WCS) Finding EO data services via Google & Co or Open Data Portals To become indexed: need for commonly understood metadata Based on Resource Description Framework (RDF) – using common vocabularies / ontologies E.g provided as annotated (schema.org) HTML pages Copyright © 2017 Open Geospatial Consortium
5
Metadata Approach: LOD/RDF
Linked Open Data (LOD) is based on RDF Elements identified by URLs, can directly be looked up Links to other URLs by using concepts of common vocabularies which provide their semantic (Geo)DCAT-AP, schema.org, FOAF, EUROVoc, OGC (EOP).. Links allow navigation to related things and minimize replications RDF Encoding e.g. in XML or JSON-LD (optimized for client consumption) Can be stored in RDF-store for which standard query language exists: (Geo)SparQL Open Data Portals often provide both and allow harvesting of RDF JSON-LD can provide RDF data (e.g. datasets, OpenSearch results or semantic annotations of HTML pages) Copyright © 2017 Open Geospatial Consortium
6
REST-API for data / metadata access
web developer friendly REST API would be important download, map view, annotated HTML, metadata (RDF/JSON) etc of EO resource -> just from one single base URL Example: EOP Resource (URI): <root>/Collections/{Coll-ID}/Products/{Product-ID} Div representations available using different mimeTypes / format parameters: Download: application/octet-stream Metadata: application/ld+json -> provides JSON-LD EOP metadata (OGC ) incl links to collection etc HTML: text/html -> describing EOP incl preview, download URL, URL to collection (HATEOS). Annotated with schema.org WMS: application/vnd.ogc.wms_xml, link to Capabilities / getMap Some resources e.g. <root>/Collections/{Coll-ID} provide query support -> OSDD for conformant search ->JSON-LD Copyright © 2017 Open Geospatial Consortium
7
Limited Example: Online Data Provision
Simple Use Case: weather conditions during forest fire in Portugal based on three EUMETSAT SEVIRI products :00 High Rate SEVIRI Level 1.5 Image Data-MSG Multi-Sensor Precipitation Estimate (GRIB) Active Fire Monitoring (GRIB) - MSG (presence of fire within a pixel) Derived buffer surrounding fires Data loaded into ArcGIS mosaic datasets and published with ArcMap as ArcGIS Image REST Service and OGC WMS/WCS) Copyright © 2017 Open Geospatial Consortium
8
Limited Example: Metadata Indexing & Search
Services (not data) described with RDF metadata (using GeoDCAT-AP, EuroVoc, INSPIRE and FOAF) Harvested into test version of new EU Data Portal (EDP), running RDF store During ingestion every RDF description gets assigned a URL Via this URL the metadata can be accessed in RDF, or annotated HTML Annotated HTML indexed by Google targeted searches using vocabularies spatio-temporal filters, etc not possible Copyright © 2017 Open Geospatial Consortium
9
Limited Example: Metadata Indexing & Search
In EDP the services can be searched and discovered via UI and GeoSparQL. much more targeted searches (facetted) possible based on categories, tags, key- words, boundingBox, format, etc. EDP is based on Virtuoso RDF store, allowing SparQL queries Copyright © 2017 Open Geospatial Consortium
10
Limited Example: Online Data Access
From search results it´s possible to open the EDP map client which loads the associated WMS Client is based on conterra map.apps: able to interpret GeoDCAT-AP and to identify the WMS access point Result entries pointing to ArcGIS Image Services can be used by a proper client (e.g. ArcMap) for visualization, applying on the fly processing, or downloading image parts a: SEVIRI b: a+Fire c: a+b+ Precipitation estimation -> incl. download options Copyright © 2017 Open Geospatial Consortium
11
Questions to be answered
RESTful Service Interfaces needed Which vocabularies to use Not suitable for EO data, currently no standard for EO Data -> OGC under development -> (Geo)JSON-LD Flexible enough (!) descriptions of service bindings: In a perfect REST world it is much easier But we have different service interfaces, protocol bindings, fileCompression, service description, parameter definitions, formats,… owc:offerings -> used in EO metadata (17-003) -> limited to some use cases Defined already ideas for a description of service access in DCAT Copyright © 2017 Open Geospatial Consortium
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.