Scope of the Edge Matching Service¶
The Edge Matching Service (EMS) is a WPS implementation of a service that aligns edges and points of vector geometries so that they will be gapless. The EMS complains OGC WPS 1.0.0.
Two possible scenarios can be defined for the use of this service:
- Dataset Cleaning: A data set should be without gaps, but there are small gaps in between the individual features that need to be filled.
- Coverage Alignment: Two data sets should have identical geometry over a shared feature (such as a common administrative border), but the geometry varies.
The Coverage Aligner provides two modes of operation:
- Align-to-Reference: In this case, all candidate data sets will be transformed so that points are moved up to a maximum distance also provided as input. As an example, consider a data set with high-quality country borders that a set of regional borders has to use as reference. Another example would be river networks of different scale, where the higher-scale data can be used as reference.
- Distribute-Errors: In this case, there is no reference data set that can be used as "ground truth", therefore all geometries will be transformed. Again, no point or edge will be moved further than a client-specified amount. This approach is useful for data sets of even quality and legal bindingness.
Furthermore, the EMS documents which modifications it did to the features in the resulting data set if its metadata document is provided as input. This lineage documentation is added extending ISO 19115.
Application in the Protected Areas Scenario¶
The HUMBOLDT project has several scenarios which are used to test-drive the developed software. In the case of HALE, the Protected Areas scenario was used.
In the Protected Areas Scenario, different data sets describing administrative boundaries and protected areas zoning are available that often share borders. An example for this is the border of the Ligure area which should be matched to the geometry of a river and also to the geometry of a protected site border. However, since all these data sets have been collected in different ways and have been generalized to different scales, they don't match up, as can be seen in the pictures below.
Administrative Boundaries Italy 1:25.000, Ligure 1:5.000 and Rivers 1:10.000
Administrative Boundaries Ligure 1:5.000 and Special Protected Areas Zoning
Administrative Boundaries Italy 1:25.000 from two sources, with a visible translation because of different reference system transformations
Literature/Links/Guides¶
- JCS Documentation. JCS is the Java Conflation Suite: an API and set of interactive tools which perform conflation on spatial datasets. JCS contains functions for performing various kinds of geo-spatial conflation processes. The major conflation problems it addresses are Coverage Cleaning, Coverage Alignment and Road Network Matching. JCS supports QA operations for detecting and visualizing errors and both automatic and manual cleaning functions to adjust geometry. It is written in 100% pure JavaTM and is open source. See http://www.vividsolutions.com/JCS/
- 52north documentation. The 52N Web Processing Service enables the deployment of geo-processes on the web in a standardized way. It features a pluggable architecture for processes and data encodings. It offers a basic client implementation for accessing the WPS (including the complete XML encoding) and plug-in for uDig and JUMP (Java Unified Mapping Platform). See http://52north.org/maven/project-sites/wps/52n-wps-webapp/index.html