INSPIRE Community Forum

SF_samplingXXX Types

236 Views

Hiya,

quick question - am I correct in my realization that there aren't any schemas for the various SF_samplingXXX types? Definition is in ISO 19156, but no specializations of SF_SpatialSamplingFeature exist in either http://schemas.opengis.net/sampling/2.0/samplingFeature.xsd nor http://schemas.opengis.net/samplingSpatial/2.0/spatialSamplingFeature.xsd

However, we require various derivatives of SF_SpatialSamplingFeature within the specialized observations. Should we provide specializations for those data providers not capable of specializing their own?

:?

Kathi

Replies

    • Public

    By Ilkka RINNE

    Hi Kathi,

    As far as I understand, the shape-specific spatial sampling features UML classes of O&M have all been implemented by the same SF_SamplingFeatureType in OMXML (OGC 10-025r1):

    There are different Schematron tests for each class (the shape must be a point etc.) in http://schemas.opengis.net/samplingSpatial/2.0/

    • Public

    By Katharina SCHLEIDT

    Hi Ilkka,

    thanks - knew I was missing something, but not sure what!!!

    But, still confusing, especially as within INSPIRE we've done the effort of providing the specialized observation classes that Simon has nicely folded back into the basic om:OM_Observation. Makes me wonder if we shouldn't then also dump the observation specializations in the serialization as this is indicated by om:type.

    I'm also chewing on the aptness of the specialized observations period! The more I look, the more grey areas pop up, i.e. the profile/trajectory mixes we've been discussing.

    Thankful for all further thoughts!

    :)

    Kathi

    • Public

    By Nicolas de Ville de Goyet

    Hello Katharina,

    I'm interested in continuing this topic as I try to use the SF_SamplingFeature element in the PointObservation schema.

    My dataset correspond to underwater noise measured at some points in the North Sea (only used for one study, not a permanent Monitoring Facility). I want to encode the geometry in the FeatureofInterest/AbstractFeature/SF_SamplingFeature/location/choice/AbstractGeometry/Point/choice/pos

          <om:featureOfInterest>
            <sam:SF_SamplingFeature gml:id="_1e603c8d-d4f4-4a80-8344-95325c1e71c0">
              <gml:location>
                <gml:Point gml:id="_d1ee7a82-0053-4aff-b75b-c11fbb434172" srsName="http://www.opengis.net/def/crs/EPSG/0/3035" srsDimension="2">
                  <gml:pos>3181160.06 3823733.175</gml:pos>
                </gml:Point>
              </gml:location>
              <sam:sampledFeature xsi:nil="true"/>
            </sam:SF_SamplingFeature>
          </om:featureOfInterest>

    However the XML created is not valid as it seems that the element SF_SamplingFeature is not allowed under the element FeatureOfInterest but then I don't understand how I should proceed and why it is included by default in the xsd schema??

    A way to make it valid is to use AbstractFeature/PointObservation instead of AbstractFeature/SF_SamplingFeature. However in the AbstractFeature/PointObservation the schema ask again for a "result", "phenomenonTime", "validTime",... which are already defined.

    Any suggestion on the best way to encode the point geometry?
    Thanks, 

    Nicolas

    P.S. Sorry to bother you but hopefully these discussions will help other people ;)

    • Public

    By Katharina SCHLEIDT

    Hi Nicolas,

    no bother at all, and nice to have new names to write to instead of just the known suspects! ;)

    Short answer: you've got some errors in the FoI XML, check my examples at https://themes.jrc.ec.europa.eu/pages/view/30357/best-practices-of-implementation-of-integrated-ef-and-om-data-specifications

    For FoI: https://themes.jrc.ec.europa.eu/file/view/168608/sdn-foi-effoi-d278-temppr01

    It does work ;) we even have complete services online (check out the Austrian INSPIRE EF-based AQD services listed on the best practice page)

    And, if you still have questions, please continue to bother! Only way we can make this work! Also, will you be in Antwerp? We'll be holding a Workshop on O&M&SOS on Thursday, September 20, 2018 - 14:00

    :)

    Kathi

    • Public

    By Nicolas de Ville de Goyet

    Hello Katharina,

    Sorry for the silence I was abroad for the last two weeks.
    I'll be in Antwerp so see you during your Workshop :)

    Cheers,

    Nicolas

    • Public

    By Nicolas de Ville de Goyet

    Proper use of FoI and publishing multiple measurements at same location

    Hello everyone,

    I’m still a bit confused on the proper use of the Feature of Interest :(

     Here is my case study info:

    • My feature of interest is a windmill in the sea.
    • I measure the sound created by this windmill in the water at two different locations (let’s say at 500m and 1000m from the windmill)
    • I measure the peak sound and the average sound (two different ObservedProperties)

    Am I right if I:

    Use one PointObservation featuretype from the SpecialisedObservation

    Describe the type of measure performed in the ObservedProperties

    Describe the distance from the windmill in the Procedure

    Describe the Windmill and its location in the FeatureOfInterest

    ??

    My remaining questions are:

    • Which FoI to choose? In the SpecialisedObservation.xsd visualized in Hale there are many different AbstractFeature available. It seems to me that, except for the SF_SpatialSamplingFeature, the complex types available describe more the ObservedProperty than the FeatureOfInterest itself?! And as the SF_SpatialSamplingFeature is for observations that varies within the scope of the feature it doesn’t seems fit for my example.
    • What is the most efficient way to link the two measurements? Possible candidates for me are:
      • ObservationSet in FoI (even if I don’t really get how and observationSet can be a FoI)
      • relatedObservation: then it means I have to create two .gml files, one for each measurement, and then link the two files with the identifier?! (this seems really heavy for two simple measurements)
      • I create a PointObservationCollection and add the ID in the members list

    Isn’t there a way to simply publish data that were taken at the same location, same time but that are not MultiPointObservation, or GridObservation, or timeseries,…

    Thanks for your help!

    Cheers,

    Nicolas

    • Public

    By Katharina SCHLEIDT

    Hi Nicolas,

    Identifying the correct FoI tends to be the main challenge in sorting O&M stuff. While the windmill seems central in your setup, you must first take a step back and question what are you actually measuring in what. While the windmill is the source of the noise, the measurement pertains to oscillations in the air around your 2 measurement locations (which neither provide the noise level at the location of the windmill!). Thus, you'll need a samplingFeature co-located with your 2 measurement points to represent the air at those locations, sampledFeature of these 2 bubbles is then either general atmosphere or a larger feature representing the air around the windmill. The windmill itself would then be linked to this as a related feature (I did a similar setup for Air Quality, samplingFeature is the air bubble around the intake, and provides info in the intake height (essential!) in addition to the basic geo lat/long)

    It could actually be an interesting use case for the MultipointObservation stuff! (do you only have 2 measurement locations around noise source, or more?)

    I very much like this use case, as it finally allows for concrete context in the more ethereal air/noise measurement world. In the water domain we've been doing interesting work on this via the OGC Environmental Linked Features Interoperability Experiment (ELFIE, should be reaching finalization, I hope to learn more next week at the OGC Meeting)

    Hope to know more by Antwerp!

    :)

    Kathi

    • Public

    By Nicolas de Ville de Goyet

    Hello Kathi,

    Just to be clear, the measurements are made in the water, not in the air ;)

    The dataset could maybe be a  MultipointObservation as it contains:

    - four windmills

    - for each windmill, measurements made at three or four different locations on the same date

    - for each location, four parameters are measured (mean sound, peak sound, ...)

    The measurements date are different for each windmill.

    So basically in a simple excel file it takes less than 20 lines, in an INSPIRE gml file I have no idea :) (still need to find out how to encode it in Hale)

    Nicolas

     

     

     

     

    • Public

    By Nicolas de Ville de Goyet

    And to continue on the FoI;

    If you go into a subcategory of, let say, MultiPointObservation/FeatureOfInterest/MultiPointObservation, you'll find the same attribute as in the parent MultiPointObservation (in fact it's a never ending list of sub-menu).

    What should be the distinction between the "result" attribute (for instance) present in both location??

    as we try to make a difference between what feature is being observed and what property is being measured I wouldn't put the same information (in addition to the fact that repeat twice the same info is not really interesting).

    This problem is not present in the three SF_XXX complex data types. Are they the only one that should be used in the FoI submenu??

    Thanks,

    Nicolas

Environmental Policy

Environmental Policy

Environmental Monitoring Facilities, Observations and Measurements