ISO/TS 19129:2009 Geographic information - Imagery, gridded and coverage data framework

From ICA Wiki
Jump to navigation Jump to search


Full name ISO/TS 19129:2009, Geographic information – Imagery, gridded and coverage data framework
Version Edition 1
Amendments None
Corrigenda None
Published by ISO/TC 211
Languages English
Online overview
Type of standard ISO Technical Specification
Meta-meta level
Related standard(s) ISO 19107:2003, Geographic information – Spatial schema
ISO 19109:2015, Geographic information – Rules for application schema
ISO 19115:2003, Geographic information – Metadata
ISO 19115-2:2009, Geographic information – Metadata – Part 2: Extensions for imagery and gridded data
ISO 19118:2011, Geographic information – Encoding
ISO/TR 19121:2000, Geographic information – Imagery and gridded data
ISO 19123:2005, Geographic information – Schema for coverage geometry and functions
Application This standard provides a framework for applications dealing with imagery, gridded, coverage and/or raster data, such as remote sensing, photogrammetry, image processing, digital elevation and terrain models, and modelling using discrete surfaces (polygons with homogenous values) or continuous surfaces.
Conformance classes Continuous Quadrilateral Grid Coverage
Riemann Hyperspatial Multi-Dimensional Grid Coverage
Triangular Irregular Network Coverage
Discrete Point Coverage
Discrete Grid Coverage
Fundamental geospatial dataset Category: Base geography
Data Theme: Rectified imagery


ISO 19129:2009 defines the framework for imagery, gridded and coverage data. This framework defines a content model for the content type imagery and for other specific content types that can be represented as coverage data. These content models are represented as a set of generic UML patterns for application schemas.

Implementation benefits

There are many different ways to encode imagery, gridded and coverage data because of structural complexities (e.g. square grids, hexagonal grids, quad trees, TINs or Thiessen polygons), traversal methods for sequencing grid cells (e.g. linear, Morton, spiral or Hilbert), distributing attributes over a space, interpolation, break lines, sensor information, geo-referencing, look-up tables, metadata, compaction and compression. There is also different and even contradicting terminology between specifications, such as the terms ‘imagery’, ‘raster’ or ‘matrix’. The result is the legacy problem of needing to deal with many old, but still very relevant, datasets in different formats.

ISO/TS 19129:2009 does not attempt to provide a new, universal format to replace the existing ones. Rather, it provides a framework for understanding the critical aspects of these formats and how imagery, gridded and coverage data fit into the general feature model used in the ISO 191xx series of standards. Further, ISO/TS 19129:2009 provides template application schemas for different types of imagery and gridded data.

Implementation guidelines

A feature is the fundamental unit of geospatial data and is defined as an “abstraction of real world phenomena”, that is, something in a database representing some identifiable thing in the real world (or the imaginary world: as planned, as envisioned, as speculated, as simulated, as fancied, etc.). Traditionally, the concept of a feature might have been applied only to vector data, where a feature had spatial attributes (geometry, possibly with multiple representations), non-spatial attributes (including multimedia), a classification or feature type, topological relationships to other features (intersection, containment, adjacency, etc.), logical relationships to other features (compound features and associations), symbology (suitable graphic representations of the feature) and metadata.

However, these can all apply to a coverage, a “feature that acts as a function to return values from its range for any direct position within its spatial, temporal or spatiotemporal domain”. The key difference is that in a vector-type feature, the spatial and non-spatial attributes are independent of one another, while in a coverage, some of its non-spatial attributes are associated with its spatial attributes (i.e. with positions in the coverage), depending on the coverage’s function or rules. For example, while a non-spatial attribute (say, its official name) might be homogenous for the whole of a vector-type feature, a non-spatial attribute of a coverage (say, the strength of reflectance in a particular radiometric frequency band) might vary throughout the feature, depending on where it is in the feature.

A coverage could be a raster image, polygon overlay, digital elevation model, Delaunay triangulation, triangulated irregular network, or Lidar point set or cloud, for example. The real benefits of considering both vector-type data and coverages as features is that it facilitates mixing them together in the same datasets, models and databases, and allows for a single application schema.

For the user wanting to understand better the nature of imagery, gridded and coverage data and the formats used for such data, Clauses 5 to 8 inclusive of ISO/TS 19129:2009, and ISO 19123 (see above), will be most useful. They provide details of:

  • The value of separating carrier and content, as many imagery and gridded data transfer formats ‘hardwire’ the encoding format into the data model, such as the encoding format determining the bit lengths of numbers, which means that the transfer format cannot be used for imagery with a higher spectral resolution.
  • The content model is the ‘information view’ of an application schema, that is, it describes the semantics of the data, independently of the transfer format or how the data are portrayed. It includes the definitions, types and valid domains of the attribute values (or the coverage function), spatial referencing information, metadata and quality information.
  • An explanation of how coverages are features, which is outlined above.
  • Spatial referencing of imagery, gridded and coverage data are handled differently, depending on the nature of the data. For example, a grid can be referenced as a whole, given the location of its origin, the spacing of cells in the grid, and the orientation of the grid. On the other hand, in a point dataset (such as a Lidar point cloud), each point has its own direct position.

In addition, Clause 7 of ISO/TS 19129:2009 provides the framework for imagery, gridded and coverage data, that is, how they fit in the general feature model. It caters for five patterns or types of coverage, for which template application schemas are provided in Section 10:

  • Continuous quadrilateral grid coverage, such as a typical satellite image;
  • Quadtree grid coverage (Riemann hyperspatial multidimensional grid coverage), such as a large set of classified satellite imagery that benefits from the compression of quadtrees, because of adjacent cells with the same value;
  • TIN coverage, a unique set of non-overlapping triangles, often used for elevation data;
  • Discrete point coverage, such as a Lidar point cloud;
  • Discrete surface grid coverage, with mutually exclusive polygons that provide continuous coverage, often with irregular boundaries, such as a classified satellite image.

Section 7 includes the diagram in the figure below to show the packages from the ISO 191xx suite of standards that are applicable to imagery, gridded and coverage data.

Imagery, gridded and coverage data can be described at an abstract level (see ISO 19123), a content model level (application schema, type of coverage, spatial referencing, portrayal and metadata) and an encoding level (such as XML, BIIF or JPEG2000). Data compaction is part of the content model level while data compression is part of the encoding level.

Clause 9 of ISO/TS 19129:2009 then provides the formal components of the imagery, gridded and coverage data structure, which in Clause 10, are assembled into template application schemas for different types of imagery, gridded and coverage data. The components are:

  • IF_DataSet, the logical entity, defined by a data product specification (see the section on ISO 19131) and which can be all or part of a collection and can have one or more tiles;
  • IF_Transmittal, the entity used for the transfer of the IF_DataSet, including encoding and physical medium;
Packages applicable to imagery gridded and coverage data (Source ISO/TS 19129:2009)
Packages applicable to imagery gridded and coverage data (Source ISO/TS 19129:2009)
  • IF_DiscoveryMetadata, the core metadata defined in ISO 19115;
  • IF_Collection, a set of IF_CoverageData (IF_GridCoverage, IF_TINCoverage, IF_PointSetCoverage and/or IF_DiscreteSurfaceCoverage) and IF_CollectionMetadata;
  • IF_CollectionMetadata, the metadata for the IF_Collection as a whole, including the IF_DiscoveryMetadata, IF_StructuralMetadata, IF_AcquisitionMetadata and IF_QualityMetadata;
  • IF_StructuralMetadata, the structure of the collection, which is optional, including grid sizes and tiling;
  • IF_AcquisitionMetadata, the source(s) of the data, essentially the lineage of the collection, which is optional;
  • IF_QualityMetadata, the quality of the data in terms of ISO 19157 or its predecessors, which is also optional;
  • IF_GridCoverage (either an IF_QuadGriddedData or an IF_RiemannGridded Data type of grid), IF_TINCoverage, IF_PointSetCoverage and/or IF_DiscreteSurfaceCoverage, which contain the actual data;
  • IF_Tiling, which describes the tiling scheme used, outlined briefly in Clause 11 of ISO 19129:2009.

The five templates in Section 10 are all structured similarly and are straightforward to interpret, if one understands the different types of imagery, gridded and coverage data. They cater for aspects such as interpolation method, sequence type, scan direction, dimensions, axes, break lines (for TINs), domains and ranges.

Annexes B and C are very brief and describe use cases and portrayal.