Skip to content

Latest commit

 

History

History
107 lines (86 loc) · 5.88 KB

ProvidingCoverages.md

File metadata and controls

107 lines (86 loc) · 5.88 KB
Raw

Providing Coverages

Overview

Data Sources

The original elevation source data is often point/vector based, this must first be transformed to a grid in accordance with the requirements to Geographical grid systems. For the purposes of this PoC, EPSG::4258 will be used. GeoTIFF is a useful format for the transfer of this gridded data from the source database to the WCS server.

As gridded data rapidly becomes unmanageable due to size, one can subset the dataset into tiles for transfer and import to the WCS. WCS can import multiple input grids and merge these to one coverage. Such tiling greatly reduces the amount of memory required by the server. The elevation data should be transformed to the target CRS close to source. While the gridded data can also be transformed later, this reduces accuracy!

Gdal library was proven to be very useful for both CRS transformation and tiling of larger elevation datasets.

Additional INSPIRE Attributes

XML Snippets must be created for the additional INSPIRE attributes, that can be provided via the metadata element of the Coverage. One XML Snippet must be created per Coverage ID available via the WCS.

ElevationGridCoverageMetadata

This requires the provision of the following information, with the values from the example below provided:

  • beginLifespanVersion: the time this electronic object was created [2021-04-09T00:00:00+01:00
  • domainExtent: spatial extent of the coverage in WGS84 coordinates
    • westBoundLongitude: western extent of the coverage 3.9632793
    • eastBoundLongitude: eastern extent of the coverage 4.8789582
    • southBoundLatitude: southern extent of the coverage 51.8526434
    • northBoundLatitude: northern extent of the coverage 52.0129335
  • inspireId: INSPIRE identifier of the coverage
    • localId: A local identifier, assigned by the data provider. For consistency, use the CoverageID INSPIRE_WNZ_5_NAP
    • namespace: Namespace uniquely identifying the data source of the spatial object https://www.rijkswaterstaat.nl
  • propertyType: the elevation property represented by the elevation grid coverage. Must be 'height' or 'depth' depth
  • surfaceType: the type of elevation surface that the coverage describes in relation to the Earth's bare surface. Must be 'DTM' or 'DSM' DTM

In the following XML example, the values listed above are encoded:

<?xml version="1.0" encoding="UTF-8"?>
<el-covmd-ks:ElevationGridCoverageMetadata xmlns:el-covmd-ks="http://inspire.ec.europa.eu/schemas/el-covmd/4.0"
    <el-covmd-ks:beginLifespanVersion>2021-04-09T00:00:00+01:00</el-covmd-ks:beginLifespanVersion>
    <el-covmd-ks:domainExtent>
        <gmd:EX_Extent>
            <gmd:geographicElement>
                <gmd:EX_GeographicBoundingBox>
                    <gmd:westBoundLongitude><gco:Decimal>3.9632793</gco:Decimal></gmd:westBoundLongitude>
                    <gmd:eastBoundLongitude><gco:Decimal>4.8789582</gco:Decimal></gmd:eastBoundLongitude>
                    <gmd:southBoundLatitude><gco:Decimal>51.8526434</gco:Decimal></gmd:southBoundLatitude>
                    <gmd:northBoundLatitude><gco:Decimal>52.0129335</gco:Decimal></gmd:northBoundLatitude>
                </gmd:EX_GeographicBoundingBox>
            </gmd:geographicElement>
        </gmd:EX_Extent>
    </el-covmd-ks:domainExtent>
    <el-covmd-ks:inspireId>
        <base:Identifier>
            <base:localId>INSPIRE_WNZ_5_NAP</base:localId>
            <base:namespace>https://www.rijkswaterstaat.nl</base:namespace>
        </base:Identifier>
    </el-covmd-ks:inspireId>
    <el-covmd-ks:propertyType>depth</el-covmd-ks:propertyType>
    <el-covmd-ks:surfaceType>DTM</el-covmd-ks:surfaceType>
</el-covmd-ks:ElevationGridCoverageMetadata>

Importing Data

Putting it all together, in order to import one coverage, you need:

  • 1 Coverage ID
  • Gridded data, e.g. Tif
    • Can be multiple files
  • 1 XML Snippet (as described above in Additional INSPIRE Attributes)
  • Language (usually default English)
  • Metadata URLS for
    • Service Metadata (ISO 19119)
    • Dataset Metadata (ISO 19115)

In addition, you need a WCS server. In this PoC, we utilized rasdaman for this purpose. Details on importing data with rasdaman are available

Provision Issues

In the section below, we describe some issues we faces when importing elevation data to WCS, together with the solutions we found to enable provision.

Wrong CRS

While ideally the gridded data is generated to correct CRS from its source, gridded data can be transformed. Using GDAL, following requests provide a compressed and transformed grid:

gdalwarp -tap -tr 30 30 -t_srs EPSG:4258 -of vrt input_file.tif output_file.vrt
gdal_translate -co compress=LZW output_file.vrt output_file.tif

Alternately, one can only utilize gdalwarp as follows, but that may lead to excessively large files:

gdalwarp -t_srs EPSG:4258 input_file.tif output_file.tif

Too little Memory

Importing the data for a large coverage in one single operation requires a great deal of memory available on the server. While one can compensate for not available RAM by increasing the swap space on the server, an alternative option is to split the file providing the elevation data into individual tiles. In the following sections, we provide GDAL commands for both projected and not projected CRS.

gdal_retile.py -ps 512 512 -targetDir .. AHN3_ZuidLimburg_subset.tif