Re: [OSM-talk] Introducing "Fix waterway direction"
On 29/08/2014 12:43, Sylvain Maillard wrote: for Europe, you could also use the EU-DEM (http://www.eea.europa.eu/data-and-maps/data/eu-dem) "a hybrid product based on SRTM and ASTER GDEM data fused by a weighted averaging approach" => full compatibility with SRTM, but with better data Nice ! Is that the only processing involved in the production of this data ? More specifically, I was wondering if anything is done specifically to detect and smooth ASTER's well-known artifacts before letting them pollute the end-product. ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Introducing "Fix waterway direction"
Hi, for Europe, you could also use the EU-DEM ( http://www.eea.europa.eu/data-and-maps/data/eu-dem) "a hybrid product based on SRTM and ASTER GDEM data fused by a weighted averaging approach" => full compatibility with SRTM, but with better data ;) Sylvain 2014-08-28 20:49 GMT+02:00 Peter Barth : > Hi all, > > I created a new challenge for MapRoulette: "Fix waterway direction". > > As the waterway's direction in OSM denotes the direction of the water's > flow, I started to compare them to SRTM. This leads to more than 1 > million wrong directions for Europe only. Many times SRTM is right, > however, due to data errors, not every time. That's why I created this > challenge. > > A more detailed description can also be found on my user diary here: > http://www.openstreetmap.org/user/Peda/diary/23632 > > I started with the really easy ones and only Europe for now. I will > upload further tasks soon, worldwide coverage will follow, too. > > Hope you enjoy this challenge, > Peda > > -- > > ___ > talk mailing list > talk@openstreetmap.org > https://lists.openstreetmap.org/listinfo/talk > ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Introducing "Fix waterway direction"
Not to brag, but New York State is already all done. I mean, there may be one or two mistakes, but when I was entering all of the streams listed in Wikipedia for NY, I set the flow direction properly. So if you have a way to exclude New York, you will have more productive users. -russ Peter Barth writes: > Hi all, > > I created a new challenge for MapRoulette: "Fix waterway direction". > > As the waterway's direction in OSM denotes the direction of the water's > flow, I started to compare them to SRTM. This leads to more than 1 > million wrong directions for Europe only. Many times SRTM is right, > however, due to data errors, not every time. That's why I created this > challenge. > > A more detailed description can also be found on my user diary here: > http://www.openstreetmap.org/user/Peda/diary/23632 > > I started with the really easy ones and only Europe for now. I will > upload further tasks soon, worldwide coverage will follow, too. > > Hope you enjoy this challenge, > Peda > > -- > > ___ > talk mailing list > talk@openstreetmap.org > https://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Introducing "Fix waterway direction"
Am 28.08.2014 um 21:48 schrieb Peter Barth: > Hi, > > Christoph Hormann schrieb: >> Great to see the waterways are getting some attention. Does this only >> analyze SRTM elevations at start and end point or are the surrounding >> waterways considered? > > currently start and end point only. I have code to do more, but for the > easy tasks that was not necessary as I have a fairly high number of true > positives. > >> If a waterway meets another waterway at one side >> but is unconnected at the other for example this is usually a strong >> indicator for the direction independent of the elevations data (which >> is only helpful usually in mountain areas). > > Actually I found many streams where start and end point are not > connected to anything at all. However, for those cases (not mountain > areas) the aerials didn't help neither in most cases ;) But as I'd like > to get a mostly complete waternetwork for another project, I'm planning > on extending the challenge at some point. > >> In terms of usability it would of course be good if you could see the >> direction in MapRoulette so you can verify the data without actually >> loading it in the editor. > > Serge suggested this, too. But I did never understand what the direction > would tell the user? Does it denote the current way's direction or the > supposed correct direction? How would the user know? And my biggest > concern is, that in many cases you need aerial anyway to be really sure. > > Anyways, would be happy to add this, once I got, how this should be done > :) You may add arrows in both directions, color-coded for "assumed" and "current" (and explained in the challenge description on the top right) regards Peter ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Introducing "Fix waterway direction"
On Thursday 28 August 2014, Peter Barth wrote: > ]...] But as I'd > like to get a mostly complete waternetwork for another project, I'm > planning on extending the challenge at some point. If you are hoping for a complete network right from the OSM database without post processing you have a long way ahead of you... > > In terms of usability it would of course be good if you could see > > the direction in MapRoulette so you can verify the data without > > actually loading it in the editor. > > Serge suggested this, too. But I did never understand what the > direction would tell the user? Does it denote the current way's > direction or the supposed correct direction? How would the user know? > And my biggest concern is, that in many cases you need aerial anyway > to be really sure. Well - cases like http://maproulette.org/#t=waterways-direction/8300823723664637270 are pretty clear cut - no need for other data. If you'd be able to switch to a background map with relief rendering like the cyclemap even less clear cases can often be properly determined. It is also a matter of motivation i think - if you can clearly see the error you are much more motivated to fix it. Of course showing the direction of the other waterways around would be very useful - but probably difficult to implement. I would always show the currently mapped direction - so the potential answer 'this is not an error' matches that rendering. -- Christoph Hormann http://www.imagico.de/ ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Introducing "Fix waterway direction"
On 28/08/2014 19:49, Peter Barth wrote: A more detailed description can also be found on my user diary here: http://www.openstreetmap.org/user/Peda/diary/23632 Your statement that canals may flow uphill is incorrect. Where canals pass over a summit, the water supply is fed into the topmost pound and then flows in both directions (i.e. downhill) from that feed. Therefore the water flow should always match a topological analysis. In the case of contour canals, no actual water flow may take place, therefore the direction of the way would be arbitrary. Also, for navigational reasons, canal administrations [in Europe (CEVNI)] have to define a left & right bank to determine the colours of buoys & nomenclature of signage. These may conflict with the actual water flow, since the left/right choice is arbitrary. Likewise the direction of the way. ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Introducing "Fix waterway direction"
Hi, Christoph Hormann schrieb: > Great to see the waterways are getting some attention. Does this only > analyze SRTM elevations at start and end point or are the surrounding > waterways considered? currently start and end point only. I have code to do more, but for the easy tasks that was not necessary as I have a fairly high number of true positives. > If a waterway meets another waterway at one side > but is unconnected at the other for example this is usually a strong > indicator for the direction independent of the elevations data (which > is only helpful usually in mountain areas). Actually I found many streams where start and end point are not connected to anything at all. However, for those cases (not mountain areas) the aerials didn't help neither in most cases ;) But as I'd like to get a mostly complete waternetwork for another project, I'm planning on extending the challenge at some point. > In terms of usability it would of course be good if you could see the > direction in MapRoulette so you can verify the data without actually > loading it in the editor. Serge suggested this, too. But I did never understand what the direction would tell the user? Does it denote the current way's direction or the supposed correct direction? How would the user know? And my biggest concern is, that in many cases you need aerial anyway to be really sure. Anyways, would be happy to add this, once I got, how this should be done :) Peda -- ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Introducing "Fix waterway direction"
On Thursday 28 August 2014, Peter Barth wrote: > Hi all, > > I created a new challenge for MapRoulette: "Fix waterway direction". > > As the waterway's direction in OSM denotes the direction of the > water's flow, I started to compare them to SRTM. This leads to more > than 1 million wrong directions for Europe only. Many times SRTM is > right, however, due to data errors, not every time. That's why I > created this challenge. > > [...] Great to see the waterways are getting some attention. Does this only analyze SRTM elevations at start and end point or are the surrounding waterways considered? If a waterway meets another waterway at one side but is unconnected at the other for example this is usually a strong indicator for the direction independent of the elevations data (which is only helpful usually in mountain areas). In terms of usability it would of course be good if you could see the direction in MapRoulette so you can verify the data without actually loading it in the editor. -- Christoph Hormann http://www.imagico.de/ ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk