search

ngageoint / hootenanny

Hootenanny conflates multiple maps into a single seamless map.
GNU General Public License v3.0
357 stars 74 forks source link

Manually match more areas #1102

Open bwitham opened 8 years ago

bwitham commented 8 years ago

For the areas.child regression test.

By areas I mean non-building polys. The two best examples I can give right now are park polygons which might contain other park poi/polys like tennis courts, baseball diamonds, etc. and school polys which might containing several school buildings. There should be several more types, just haven't ran across too many of them yet.

The only dataset in which we currently have manual matches of this kind is the San Francisco poi poly dataset C. Getting more manual matches of this kind, if possible, would help out the quality of this type of conflation. Currently area to area conflation is being added as an optional, experimental generic conflation type (based off of generic polygon conflation) as part of #205 since doing so helps clean the poi to poly output up quite a bit at the end. The current results against dataset C are pretty good, but we need more data to test against before area to area conflation can be considered robust.

Datasets with ~ a couple hundred matches are ideal.

bwitham commented 8 years ago

@curranMapper Not urgent by any means, but this is an issue to track possible manual conflation of area polygons. If you're ever able to snag any data like that, let me know. thanks.

bwitham commented 7 years ago

Closing all area matching related issues until interest in it is shown.

bwitham commented 4 years ago

The data from #3978 would be a good addition.