Attachment of quality information

SANY D2.3.4 Specification of the Sensor Service Architecture V3 Doc.V3.1 Copyright © 2007-2009 SANY Consortium Page 201 of 233 with a matching targetrule for the given request. In the example above an XACML PDP would decide that the requester is permitted to access the resource SoapBindingsSOSv3WS01 because a policy exists which addresses the resource in a target. Furthermore, the request has one subject which possesses the group attribute with the required value SOS User and the action attribute getObservation . Therefore the rule evaluates to the effect Permit and because of the fact that there are no further rules which could evaluate to Deny the resulting decision of the PDP is Permit.

10.6. Processing of Quality Information

10.6.1 Attachment of quality information

The OGC Sensor Observation Service see section 8.2.2 is used to access observation results. Uncertainty and quality information is relevant at two different places within this service specification: - At the level of the observation process, uncertainty and quality information may be included in the SensorML document returned by the describeSensor operation. All required information about the observation process, the uncertainty of the observations as a collection and the quality assurance processes applied may be included in this document. Time dependent uncertainties, e.g. due to instrument deterioration, could be expressed in SensorML. Client applications may, however, find it more convenient to have the resulting uncertainty of observations expressed directly for individual observations as in the next item. - At the level of the individual observations, uncertainty and quality information may be included in the Observation Measurement documents as returned by the getObservation operation. Again, the observation-specific information regarding the observation process, the uncertainty of this observation value and the quality assurance process that this observation value has undergone may be included here. Figure 10-23 shows an example of a getObservations result with an UncertML block quantifying the uncertainty of the observations. A href in the result block definition provides the link between the property observations and the associated uncertainty data. There is a tacit but natural assumption that the order of the uncertainty information is the same as the observation values. The XML file can be parsed by clients not able to evaluate the uncertainty data. A model based calculation usually makes several assumptions about the nature of the physical process under study e.g. ground water flow in a saturated, homogeneous aquifer with uniform hydrological parameters. These assumptions shall be described in the associated SensorML of the model procedure. It is good engineering practice to always include quality and uncertainty information. SANY D2.3.4 Specification of the Sensor Service Architecture V3 Doc.V3.1 Copyright © 2007-2009 SANY Consortium Page 202 of 233 Figure 10-23: UncertML block in a getObservations result 10.6.2 Multi-level measurement chains When working with quality- assurance processes, data have different “levels” of quality control information, as illustrated in Figure 10-24. For example, in the air quality domain the “raw data” sampled from the sensor undergo some automatic quality assurance process “QC level 1”. In a second step, a manual quality control process is applied to the data “QC level 2”. Sometimes a user may be interested in some specific level of quality controlled data e.g. raw data. Other application scenarios require querying the “best available” data, which means that for each measurement taken, the data point with the highest level of quality control should be returned. Within an OGC Sensor Observation Service this can be handled using different offerings with different procedures. The procedure describes the level of quality control that this specific data set has undergone. For each level a procedure has to be defined, and, if required, one or more additional procedures defining the “best available” data can be defined. SANY D2.3.4 Specification of the Sensor Service Architecture V3 Doc.V3.1 Copyright © 2007-2009 SANY Consortium Page 203 of 233 Data store R a w d a ta Q C l e v e l 1 Q C l e v e l 2 „B es t a va ila bl e“ Se n s o r O b s e rv a ti o n Se rv ic e Quality control operator End user „Best available“ QC level 2 QC level 1 Raw data Time O ffe ri n g s Figure 10-24: Example for a multi-level measurement chain in an SOS 10.6.3 Visualisation of Uncertainty Information It is desirable to represent uncertainty with graphic variables on maps created with the Map and Diagram Service see section 8.4.3. Several promising techniques have been already identified for the visualisation of uncertain information in static maps. MacEachren 1992 promotes the use of transparency for uncertainty depiction based on a metaphor of fog obscuring the view proportional with the amount of uncertainty. He also includes additional modalities for uncertainty visualization such as colour saturation, crispness contour crispness and fill clarity and degradation of the resolution of raster images. A similar technique comes from Drecki 2002. He proposed an opacity display, where opaque objects are the certain ones. The last identified technique comes from Hengl 2003. His work suggests that uncertain data should appear increasingly white or “pale,” depending on the magnitude of uncertainty. Whereas the first techniques apply mainly to coverages, the last two techniques opacity and colour bleaching may prove to be especially important for the visualisation of discrete geographical objects such as moving sensors. Based on the techniques presented above, it should be possible to obtain map representations that display the amount of uncertainty by varying transparency, varying colour, transparency blending, colour bleaching, use of fill patterns and adjusting resolution of geographic detail. Moreover, considering the specificity of sensor data as processed in sensor service networks which primarily consists of interpolated measurements and coverages, the following three additional techniques have been investigated in the SANY project on a conceptual level: - perpendicular colour and transparency variation along contour lines - varying contour widths, and SANY D2.3.4 Specification of the Sensor Service Architecture V3 Doc.V3.1 Copyright © 2007-2009 SANY Consortium Page 204 of 233 - use of graphic filters e.g. blurring as a generic mechanism for localised colour manipulation. Note: Currently, the SensorSA suggests that the use of colour, transparency and texture are the best candidates for representing uncertain information for static maps in an efficient manner. Therefore, these techniques will be further researched in the context of the Map and Diagram Service. The other techniques will remain only as concepts due to the complexity of implementing such graphically demanding techniques in the Map and Diagram Service.

10.6.4 Unit conversion