I have a problem (illustrated below) with matching two datasets.

Datasets to Match On the left I have a GIS containing drain network data (in blue), including diameter, length and topology information (from- and to manholes).

I have received some video drain survey data which has neither coordinates nor a reliable key to the original GIS data (illustrated on the right in green). It does however contain useful attribute information like diameter and length and topology information. Not all drains have been surveyed.

I need to match the two datasets. I can do this by matching similar diameters and length attributes (providing a score) and pairing up drains using bipartite best matching. I do not however use the topology which could be extremely valuable information and ensure the matching occurs in the correct order. My approach could therefore deliver poor results (imagine a network where lots of drains have similar lengths and diameters).

I would like to know if anyone knows of an algorithm (Python implementation?) applicable to this type of problem.

I am considering generating the score by also comparing up- and downstream pipes but this just seems like an extension of a bad idea...

  • Just a thought, looking at your example (which I appreciate may not be realistic) you have 2 nodes in the network that have a valency of 3. May be valency could be a property that helps you match the pipes? – Hornbydd Mar 8 '18 at 23:05
  • Having from and to manholes is more than enough. – FelixIP Mar 9 '18 at 3:08
  • Hi, thanks. Valency a good idea. To and from Manholes Important the manholes are not named - only the connectivity of the links (which also have no names) – Mr_Robinini Mar 9 '18 at 8:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.