Thanks to all who replied to this vexing problem. My postdoc Isaac Ullah found what seems to be a good solution to part of this.
v.net.path appears to extract a single object, single cat vector line from a stream network if we can give it a starting and ending point. Michael ____________________ C. Michael Barton Director, Center for Social Dynamics & Complexity Professor of Anthropology, School of Human Evolution & Social Change Head, Graduate Faculty in Complex Adaptive Systems Science Arizona State University voice: 480-965-6262 (SHESC), 480-965-8130/727-9746 (CSDC) fax: 480-965-7671 (SHESC), 480-727-0709 (CSDC) www: http://www.public.asu.edu/~cmbarton, http://csdc.asu.edu On Dec 1, 2014, at 11:21 AM, Michael Barton <michael.bar...@asu.edu> wrote: > v.build.polylines is indeed part of what is needed. But it is not enough to > be able to use v.to.points in the data I have—and this is not a weirdly > meandering stream, just a Mediterranean barranco. > > A major part of the problem is in the raster to vector conversion. I’ve been > using r.to.vect. But it breaks the resulting line into multiple segments and > assigns different cat numbers to the segments. If the segments all connect > cleanly, v.build.polylines may combine the segments. But I’ve tried using > cat=first and it continues to assign multiple cats to the segments. To get a > single line, what I have to do currently is: > > v.clean > v.build.polylines with cats=no > v.category with option=del ## may not be necessary but I’ve had enough > inconsistent results to want to make sure on this > v.category with option=add cat=1 step=0 ## this gets me a vector with cat=1 > for all parts. > > repeat the above a 2nd time > > This gives a sufficiently clean single-object line with a single cat number > that I can run v.to.points on. > > Maybe r.stream.order does a cleaner job in raster to vector conversion. It’s > worth a try. Unfortunately, I can’t use the stream order info to extract a > stream since I want the stream from its headwaters (stream order = 1) to its > outlet (stream order > 1). > > Ideally, there would be a module where I could enter the coordinates of the > upstream end of a stream, it would extract the entire watercourse as a single > object to its downstream end, edge of map, or a defined outlet stop point. > Then I could run v.to.points and v.what.rast to get profile data. OR in a a > stream profile module, it would run these for me. I could do it in Python if > I could extract a clean vector (one object, one cat, no loops or > self-crossing) of a single stream from a stream network created with > r.watershed or r.stream.extract. AFAICT, there is nothing available in GRASS > that can do this. Kind of surprising. > > Michael > > > ____________________ > C. Michael Barton > Director, Center for Social Dynamics & Complexity > Professor of Anthropology, School of Human Evolution & Social Change > Head, Graduate Faculty in Complex Adaptive Systems Science > Arizona State University > > voice: 480-965-6262 (SHESC), 480-965-8130/727-9746 (CSDC) > fax: 480-965-7671 (SHESC), 480-727-0709 (CSDC) > www: http://www.public.asu.edu/~cmbarton, http://csdc.asu.edu > > > > On Dec 1, 2014, at 9:31 AM, Moritz Lennert <mlenn...@club.worldonline.be> > wrote: > >> On 30/11/14 06:28, Michael Barton wrote: >>> >>> ____________________ >>> C. Michael Barton >>> Director, Center for Social Dynamics & Complexity >>> Professor of Anthropology, School of Human Evolution & Social Change >>> Head, Graduate Faculty in Complex Adaptive Systems Science >>> Arizona State University >>> >>> voice: 480-965-6262 (SHESC), 480-965-8130/727-9746 (CSDC) >>> fax: 480-965-7671 (SHESC), 480-727-0709 (CSDC) >>> www: http://www.public.asu.edu/~cmbarton, http://csdc.asu.edu >>> >>> >>> >>> >>> Thanks César, >>> >>> This is pretty much what I’m doing. However, it is not that simple, it >>> turns out. >>> >>> First, for a stream profile, you want to isolate a single watercourse. >>> r.stream.extract creates a stream network. Getting stream order does not >>> help with profiling, however. >>> >>> I looked at r.stream.distance. But this calculates distance from each >>> stream junction. I want it for the entire course of the selected >>> stream—from headwaters to outlet. That is the only way to graph a stream >>> profile for the entire stream. >>> >>> Once I have a set of points with the distance from the beginning or the >>> end of the line stored as an attribute for each point, the rest is easy. >>> It is getting to this step that is hard. >>> >>> If a line is composed of multiple segments—as is the case for any line >>> created with r.stream.extract or r.watershed, and also with r.drain >>> surprisingly—there is nothing that will ignore all of these segments and >>> just give me the distance along the line from one end to the other. This >>> is the case with v.to.points >> >> As MarkusN said: you have to use v.build.polylines with cats=first, but then >> it is easy to get total length with v.to.db or v.to.points if that is what >> you need (or many other options (get end node coordinates and use v.distance >> 'to_along' or MarkusM's suggestion: v.net.distance from start to end node >> along the stream). >> >> Here's an example in the NC dataset, using r.stream.order to approximate the >> main stream: >> >> g.region n=227440 s=221640 w=630220 e=637220 res=10 >> r.watershed --overwrite elevation=elevation@PERMANENT threshold=10000 >> accumulation=accum drainage=direction stream=streams >> r.stream.order --overwrite stream_rast=streams@user1 >> direction=direction@user1 elevation=elevation@PERMANENT >> accumulation=accum@user1 stream_vect=streams horton=horton >> v.extract --overwrite input=streams@user1 where=horton=3 output=main_stream >> v.build.polylines --overwrite input=main_stream@user1 >> output=main_stream_poly cats=first >> >> then: v.to.db (result=11312.5) or >> >> v.to.points input=main_stream@user1 output=main_stream_points dmax=1 >> >> db.select sql="select round(max(along)) from main_stream_points_2" >> round(max(along)) >> 11312.5 >> >> >> Moritz > _______________________________________________ grass-user mailing list grass-user@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-user
