I've preferred Ian's java app to shp-to-osm.
I just want to put a plug in for a workflow to potential NHD importers that I
have arrived at after many previous mistakes.
1. Convert nhd shp for a subbasin files using shp-to-osm with max-nodes set to
a very high number. This makes one big file per nhd file (waterbody, area,
flowline, point, line).
2. Merge all the files in josm. Make sure you have allocated a lot of memory
to josm as this can take a lot of memory. It will also take a lot of time.
3. Use validator to identify duplicate nodes. There are two main sources of
duplicate nodes:
a. Flowline intersections--NHD has each reach as a separate vector.
These should all be merged.
b. Flowline intersections with riverbanks and waterbodies. These, too
should be merged.
c. Shared boundaries between lakes, wetlands, riverbanks, dams and
other features. I think there is some dispute what to do here. If there are
two multipolygon relations involved, one shared way should be created that is
a member of both relations. If there aren't two relations, I think one could
either create two or simple merge nodes. I have been merging nodes.
4. Quality control. Keep in mind several things about the NHD. They are not
perfect and they overlap between basins. Try your best to keep the import
from duplicating others' and your own work. Using the slippy map plugin,
double check that features in the import do not already exist. They could
either be NHD data, which can be deleted from the new import, or another form
of data--either an import such as PGS or a real human. Use judgment to
determine which is the best. Based on up-to-date orthophotography I can tell
that NHD is not always the best data. However, many water features input by
hand are not always better either-- they are frequently hastily sketched off of
Yahoo pictures and not super accurate.
If you do multiple subbasins in the same basin, you will quickly notice that
there is a ton of overlap in both the waterbodies and the riverbanks between
different subbasins.
5. Upload the entire subbasin using josm's ability to split large changesets.
If your connection is at all flaky, you may want josm to split into smaller
changesets than the maximum (50000 changes).
Sorry for the book, but I wanted to share some of my experiences.
James
On Tuesday, June 22, 2010 03:56:16 pm Ian Dees wrote:
> On Tue, Jun 22, 2010 at 2:36 PM, Dylan Semler <dylan.sem...@gmail.com>wrote:
> > I've patched the nhd2osm scripts to print details about the NHD
> > tag conversions[1]. I'm attaching the output of the scripts with my
> > patch applied if anyone is interested in what it does.
>
> The shp-to-osm Java app I wrote has a more complete set of NHD conversion
> rules here:
>
> http://wiki.openstreetmap.org/wiki/NHD_Rules
>
> Information on the JAR is here:
>
> http://wiki.openstreetmap.org/wiki/Shp-to-osm.jar
_______________________________________________
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us
