Download presentation
Presentation is loading. Please wait.
Published byJustina Patrick Modified over 9 years ago
1
U.S. Department of the Interior U.S. Geological Survey Dalia Varanka Research Geographer Center of Excellence for Geospatial Information Science Spatial Ontology Community of Practice Workshop October 17, 2008 Ontology Research for The National Map Dalia Varanka Research Geographer Center of Excellence for Geospatial Information Science Spatial Ontology Community of Practice Workshop October 17, 2008
2
Project Description Ontology for The National Map is a framework to more explicitly articulate detailed and shared information about U.S. topographic features A robust exchange of feature semantics enables greater information access for a diverse public Richer data models based on ontology will increase potential data applications National Research Council recommends to establish a research priority to explore use of geographic feature ontologies for geographic features to enable information discovery Integrate data from a variety of formats and platforms Enable natural language queries on features
3
Parts of this Presentation Background Topographic mapping and narrative Narrative as an approach to restore ontology semantics Semantic facets of queries
4
Topographic mapping Base map – an empty container Surface landscape features at local scale Navigational device Natural resource and urban development A record of the changing American landscape National unity and identity These meanings serve as the semantic contexts for topographic feature ontology
5
Topographic maps and narratives Topography as a type of temporal, narrative experience of the land (Lukermann 1961; Curry 2002) American topographic landscape as the stage of our domestic experience Routine activities, narratives, and symbols associated with places Topographic map as the information storage device for that experience Interact with the map via narratives and symbols
6
Application of the ontology to The National Map Scope: Science and Society Science is structured by its practitioners and broader social contexts The USGS is mission-driven to link science to society Common-term concepts For example: “The valley floor of the Grand Canyon,” “The spread of invasive species,” “Sea- level rise”
8
Unify existing stages of an ontology Feature lists Themes/domains Classification Gazetteer Crosswalks What we need DB interfaces Relations / axioms
9
20 th -Century Topographic Mapping A balance of mathematical, regional, and local mapping Combined photogrammetric (air photo) and planetable (field) methods of surveying Were focused on the 15 or 7.5 minute quadrangle, not government units Topographic Instructions were general for regional implementation with state partner needs Memos for verbal mediation and modifying ideas
10
National Databases Transition to a single national database strengthened centralized standardization of feature lists Field offices vied for national center status Feature lists served various functions, such as mapping, digital computation, and data interoperability, not just regional character Specifics communicated via metadata, not two-way discourse Based on governmental unit
11
Objects: Feature and Code Lists DLGBest Practices Cable/Pipeline SiteUtil_Line Dam SiteUtil_Point SubstationUtil_Structure Pipeline Regulation StationUtil_Area
12
Feature List Standards Comparisons Spatial Data Transfer Standard (SDTS) Digital Line Graph-Enhanced (DLG-E) Digital Line Graph-Feature (DLG-F) Federal Geographic Data Committee (FGDC) Framework Best Practices Vector Data Model (The National Map)
13
DLG-E (1988)DLG-F (1993) FGDC Framework Data (1998) The National Map (2004) Cover Cover - Barren Land Non-vegetative surface cover Cover - Built up - StructureBuilt upStructures Cover - Built up - Complex Cover - Built up - Utility Cover - Built up - NetworkTransportation Transportation-Base Transportation-Rail Transportation-Air Transportation-Transit Transportation-Roads Transportation- Waterways Cover - Cultivated Cropland Cover - VegetationVegetative Surface Cover Cover - WaterHydrography Domains – Land Cover
14
DLG-EDLG-F FGDC Framework DataThe National Map Division Boundaries Government Unit BoundariesGovernmental Units Division-Administrative Division-Boundary Division-Census Division-Hydrologic Unit Division-Land ParcelCadastral Division-Locale Division-Maritime Division-Political Division-Survey System Public Land Survey System Public Land Survey System Domains - Surveys
15
Domains – Earth Processes DLG-EDLG-F FGDC Framework Data Best Practices Data Model Ecosystem MorphologyNamed LandformsElevation Hypsography GeopositionGeodetic Control Digital Orthoimagery
16
Domains - Security Historical relation with defense mapping New emphasis on event-based, hazard management Ops_IncidentLine Ops_AccessLine Ops_SearchRescueLine Ops_SearchRescuePoint Ops_IncidentPoint Ops_AccessPoint Ops_EmergencyFacility Ops_IncidentArea Ops_AccessArea Ops_SearchRescueArea Ops_AlertArea Ops_ResourceAssignment Ops_DamageAssessment
17
Assumptions so far The working hypothesis is that we share a common-term vocabulary within the topographic map, but statements and queries are personalized and vary; The tendency to move feature types to data model implementation may be narrowing our discourse about topography; and Syntactic variance of query statements reflects the narrative of topographical experience; feature lists are better than code lists.
18
Approach Base the analysis of feature content and syntactic structures on various narrative forms Feature types, definitions, and domain expertise Production: surveys, instructions, memos, and compilation Gazetteer, particularly the Geographic Names Information System Volunteer geographic information Map reading events Content analysis of text concordances
19
Glossary Definitions A stream is a body of water, with a current, confined within a bed and stream banks. Concept conditions …is a body of water …has a current, a source, a mouth [is] confined within a bed and stream banks
20
Categorization Basic level Appear between super- and sub-classes Have an associated action that comes to mind Have the most attributes Have the most whole-part structures For example, road (driving) and river (flows)
21
Other categories Family resemblance Associated with super-classes For example, vegetation Degrees of membership No precise boundaries For example, wetlands Generator Feature generates new features Glaciers generate cirques, till, moraine, kames
22
Syntax, with addition of names Geographic Name Report Description Bluff Creek is a stream about 10 miles long heading in [sec./twp./range] flowing generally SE to the Leaf River in [sec./twp./range] about 5 miles southeast of Bluffington. [name] [subject] [modifier_length] [start_location] [event] [direction] [end_junction] [end_location] [proximity]
23
Queries What is Bluff Creek?What is [name] [Name] [relation] [object] Bluff Creek is a stream. Where is Bluff Creek?Where is [name] [start-location] [event-direction] [end-location] [proximity] Headed in [sec./twp./range] Bluff Creek flows generally SE to the Leaf River in [sec./twp./range] about 5 miles southeast of Bluffington.
24
Ontology to Database Interface
25
Semantic similarity Types and causes Regionalization (Hollows, Valleys, Bottoms) Borrowings (Playa, Mesa, Prairie) Geographic scale (River, Stream, Creek) Infrastructure change (Condo, Gated Community) Implications for query analysis Synonyms Variants Spatial cognition
26
Semantic context to enhance query functionality Features with indeterminate boundaries carry semantic ambiguity in definition, name, and recognition. By relating landform features with indeterminate boundaries to elevation, a geographical or environmental context, the identification, extent, and naming of these features is defined despite linguistic or spatial perspective ambiguity.
27
Elevation (context) and Landform (feature) Landform term Glossary term definition Coded criteria applied to elevation context Landform feature extraction
28
Features and spatial relations Absolute location: gazetteer coordinates Topological location: Query the topology rather than the feature as the primary subject ‘What is “near” the “lake?”
29
Summary: work in progress Work in progress includes the conceptual design and ontology software development of categorization, relation, attributes. Current research includes integrating the ontology software to USGS databases.
30
Contact Dalia Varanka Center of Excellence for Geospatial Information Science Tel: 573-308-3897 Email: dvaranka@usgs.gov
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.