OGC adopts new revision to GML in JPEG 2000 (GMLJP2)

16 October 2018: The membership of the Open Geospatial Consortium (OGC) has approved version 2.1 of the GML in JPEG 2000 (GMLJP2) Encoding Standard.

The GMLJP2 Encoding Standard defines how Geography Markup Language (GML) is used within JPEG 2000 images and other gridded coverage data. GML, in this case, can be used to add geospatial content to imagery, such as the geographic coordinates of the image, annotations, or references to features.

With newly added support for referenceable grid coverages, GMLJP2 transitions from an imagery format restricted to rectified ‘map’ images to a more general format with support for a much wider set of imagery types, including sensor model imagery.

Sensor model images are useful because they contain geospatial information that is lost upon processing into rectified ‘map’ images. Besides being processed into map images, sensor model images may be used, for example, for precision mensuration or to construct 3D models of ground structures.

The 2.1 revision adds support for referenceable grid coverages (and thus sensor model images) via the Coverage Implementation Schema (CIS) extension, ReferenceableGridCoverage.

The ReferenceableGridCoverage Extension has a referenceable grid element that allows coverages to contain sensor models as described by the OGC SensorML 2.0 standard. By providing a raw unprocessed image with a SensorML 2.0 sensor model placed in a referenceable grid description, one defines a sensor model image.

An example is a visible or infrared light sensor consisting of a rectangular array of light sensing pixels used as part of a camera mounted in an aircraft. The light entering a given pixel element comes from a ground location that is determined with a sensor model that is part of a referenceable grid description.

Simple XML examples using the standard may be found on the OGC GMLJP2 v2.1 schemas page. Full examples, including one closely based on the Community Sensor Model for frame cameras, are also available from Aeroptic.

Lucio Colaiacomo, Chief Technology Officer of SatCen, the European Union Satellite Centre, commented: “Our daily business is the exploitation of high resolution satellite imagery in the field of Common Foreign and Security Policy (CFSP). In order to fulfill our objective, it is necessary to support appropriate replacement and physical sensor models. Although most high resolution satellites have pushbroom imaging systems, instead of providing standardized physical sensor models, vendors normally provide highly accurate bias-corrected replacement sensor models using a set of rational polynomial coefficients (RPC). However, as RPC coefficients are in some instances not provided by vendors, and because at times the provided RPC coefficients for a given image have problems, being able to support standardized physical sensor models of the imaging process, such as the metadata profiles of the Community Sensor Model standard, is also necessary. By allowing replacement and physical sensor models, encoded in SensorML 2.0, to be associated with a given image, by either an embedding in the coverage encoding or by a reference to an external file, this version 2.1 release of the GMLJP2 imagery standard has the potential to support our mission.”

Frank Suykens, CTO of Luciad, a Hexagon company, commented: “Hexagon’s software components power many industry solutions to analyze, visualize, and act on geospatial and real-time data. Metadata is key. Standardized GML content within JPEG2000 is essential to understand and use the imagery correctly and efficiently. The addition of sensor model imagery is an important evolution to exploit new UAV and satellite data. Interoperability is key as well. Hexagon Geospatial is an early adopter of the new GMLJP2 version to support activities with SatCen and Hydrographic Institute of Portugal.”

Eric Hirschorn, a Software Design Engineer at KeyW Corporation, Remote Sensing Division, commented: “A few years ago we developed a SensorML 2.0 representation of the Community Sensor Model metadata profile for frame cameras, but we also wanted to embed this imaging geometry data into the JPEG2000 image frames produced by our Aeroptic line of airborne imagery collection systems. Now, with GMLJP2 version 2.1, we are excited that there is an OGC standard supportive of our work that will enable open exchange of imagery data, for example from oblique aerial sensors such as those we have developed.”

As with any OGC standard, the open GML in JPEG 2000 (GMLJP2) Encoding Standard is free to download and implement. Interested parties can view and download the standard, which also includes useful hints and best practice for implementation, on OGC’s GML in JPEG 2000 (GMLJP2) Encoding Standard Page.