A Survey of Industry Data Models and Reference Data Libraries
4 Graphical Information Systems
4.1 Open Geospatial information
4.1.1 Defining organization
GIS models are developed and standardized by Open Geospatial Consortium (OGC) and ISO/TC 211 "Geographic information/Geomatics".
4.1.2 Objectives and scope
The OGC core GIS model is the basis for many domain specific models which are concerned with the representation of things distributed over the surface of the Earth. Some of these are standardised within OGC, and other outside. Domain specific models standardised within OGC include:
- the CityGML (clause 5.2) and Landinfra (clause 5.3) models for the built environment;
- the GeoSCiML (clause 4.3) model for geology;
- domain specific models, such as PipelineML (clause 4.2) model.
The OGC core GIS model is also at the heart of the EU INSPIRE spatial data infrastructure (clause 5.4).
The OGC also defined models for observations and measurements and sensor networks. In June 2020, the OGC adopted the HDF5 standard for the efficient representation of structured data.
4.1.3 Structure of the model
At the heart of the GML schemas is the element feature. This is a spatio-temporal object comparable to physical object in ISO 15926-2. A feature can have states or time varying relationships. An outline of the approach to feature represented in UML is shown in Figure 4.
Figure 4 - Feature in GML
Comments on Figure 4:
- A dynamic feature exists at an instant in time or during a period of time. If a dynamic feature has states, then it is regarded as a dynamic feature collection.
- The relationship between a dynamic feature and time is directly analogous to the relationship between a feature and space.
- An abstract time slice is specialised by relationships which are different at different times.
- The GML time model also contains the time topology relationships of Allen’s interval algebra.
The relationship valid time is a specialisation of time primitive property. Other specializations can be defined in an application schema.
The top of the model for geometry is shown in Figure 5.
Figure 5 - Geometry in GML
Comments on Figure 5:
- The segments of a curve are explicitly defined finite bounded curves, but which are not part of the geometric hierarchy. Similarly the patches of a surface are explicitly defined finite bounded surfaces, but which are not part of the geometric hierarchy.
- The overall structure of this model is very similar to that of ISO 10303-42. A line string corresponds to a polyline in ISO 10303-42, and a linear ring corresponds to a poly_loop and a polygon corresponds to a polygonal_area.
- GML seeks to separate geometry from topology in a way that is similar to ISO 10303-42. Nonetheless, the relationships shown explicitly in this figure are all topological.
The relationships between a feature and geometry are not defined in GML, but are left to an application schema. GML defines curve property as a relationship with an abstract curve. An application schema can specialised this to be a centre line.
The GML model recognises that as a feature changes through time, its geometry can change whilst its topology remains the same. Therefore there is a separate topology model, and a feature can be independently associated with topology. The GML approach to topology is discussed in document “OWS (OGC Web Services) 3 GML Topology Investigation” (http://portal.opengeospatial.org/files/?artifact_id=14337). The document contains - Geometry and topology, which shows an approach similar to that of ISO 10303-42.
Figure 6 - Geometry and topology in GML
A major different between geometry in GML and geometry in ISO 10303-42, is that the geometric objects in GML, such as points, curves and surfaces, exist in physical space rather than in the space of real pairs or triples, according to dimension, with the Euclidean metric.
The models are represented initial UML, from which the XML schemas that make up the Geography Markup Language (GML) are derived. Ontologies in OWL are also being derived. The architecture of the standards is shown in Figure 7.
Figure 7 - Architecture of OGC standards
NOTE The architecture provides extensibility at both the UML and XML schema levels. As a result, there is some complexity. Some relationships are defined partially in the GML schemas and then left to the application schemas to make precise. This is analogous to the “management resources” in ISO 10303.
4.1.5 Maintenance and usage
Both the OGC and ISO TC 211 are active organizations, supporting a family of standards for geospatial information with regular revisions.
The OGC standards are ubiquitous for geospatial information.
4.2.1 Defining organization
This standard was developed by the OGC PipelineML Standards Working Group.
4.2.2 Objectives and scope
The OGC overview says:
The OGC PipelineML Conceptual and Encoding Model Standard defines concepts supporting the interoperable interchange of data pertaining to oil and gas pipeline systems. PipelineML supports the common exchange of oil and gas pipeline information. This initial release of the PipelineML Core addresses two critical business use cases that are specific to the pipeline industry: new construction surveys and pipeline rehabilitation. This standard defines the individual pipeline components with support for lightweight aggregation.
Additional aggregation requirements such as right-of-way and land management will utilize the OGC LandInfra standards with utility extensions in the future. Future extensions to PipelineML Core will include (non-limitative): cathodic protection, facility and safety. PipelineML was advanced by an international team of contributors from the US, Canada, Belgium, Norway, Netherlands, UK, Germany, Australia, Brazil, and Korea.
4.2.3 Structure of the model
PipelineML is an application schema for GML. The objects in the PipelineML model are shown in Figure 8.
Figure 8 - Objects in PipelineML
Comments on Figure 8:
- This is a straightforward implementation of the OGC architecture shown in Figure 7.
- The subclasses of component have appropriate properties, and each has a reference to GML geometry. However the components have no topology, so the topology of the pipeline has to be deduced from the geometry.
The PipelineML is documented according to the guidelines established for OGC standards, as shown in Figure 7.
4.2.5 Maintenance and usage
PipelineML was published in August 2019. The standard seems to be a natural companion to process industry standards, such as MIMOSA CCOM and ISO 15926, but there is no evidence of such usage.
4.3 Geoscience information
4.3.1 Defining organization
GeoSciML (Geological Sciences Markup Language) and EarthResourceML (Earth Resources Markup Language) are standards produced by the Commission for the application and management of Geoscience Information (CGI) under the International Union of Geological Sciences (IUGS).
4.3.2 Objectives and scope
The GeoSciML home page says:
GeoSciML is a data model and data transfer standard for geological data - from basic map data to complex relational geological databases.
The EarthResourceML home page says:
The model describes the geological features of mineral occurrences, their commodities, mineral resources and reserves. It is also able to describe mines and mining activities, and the production of concentrates, refined products, and waste materials.
4.3.3 Structure of the model
GeoSciML is a well-documented GML application. There is even a postersized UML diagram - https://portal.opengeospatial.org/files/?artifact_id=72895 . The dependencies on OGC standards are shown in Figure 9 are taken from the GeoSciML web site.
Figure 9 - GeoSciML schemas
The SWE Common Data Model referred to in Figure 6, is the Sensor Web Enablement (SWE) Common Data Model Encoding Standard, which is defined by the OGC for low level sensor data - see https://www.ogc.org/standards/swecommon .
EarthResourceML is an extension of GeoSciML, in the same style. It is documented with simple and easily understood diagrams, such as Figure 10.
Figure 10 - EarthResourceML schema for a mine
The GeoSciML and EarthResourceML are documented according to the guidelines established for OGC standards, as shown in Figure 7.
4.3.5 Maintenance and usage
The CGI is an active organisation. The latest [in August 2020] release of GeoSciML was in May 2015 and of EarthResourceML in October 2018.
The CGI standards are ubiquitous for geoscience information.