Re: [OSM-talk-be] CORINE Land Cover 2000 (now with permission)
+1 2010/6/7 Luc Van den Troost luc.a...@gmail.com As it starts to be an early voting here, +1 for me too, at least for the areas with low areal resolution. For the areas with high res areal images we probably have to look first what areas have a good landuse definitions right now. Based on images and personal knowledge they might be much more precise and detailed than corine. Just one remark regarding rendering. Landuse rendering seems too dominant on certain zoom-levels now. I noticed it near the french border here: http://www.openstreetmap.org/?lat=49.8146lon=4.9209zoom=12layers=B000FTF At this zoom level, even tertiary roads, that should be rather clear, fade away in the dark green forest area. Luc / Speedy On Mon, Jun 7, 2010 at 7:37 PM, Renaud MICHEL r.h.michel+...@gmail.comr.h.michel%2b...@gmail.com wrote: Le lundi 07 juin 2010 à 14:42, Ben Laenen a écrit : Lennard wrote: Now that the legal issues have been resolved, and the technical setup has been mostly performed, the question remains about whether to proceed with an actual import? Don't worry about it happening overnight, we have plenty of time to discuss and review the steps involved. +1 for importing from me as well. +1 for me too. -- Renaud Michel ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be -- Mathias Versichele Bio-ir milieutechnologie / Msc. geografie Oudburgstraat 16 9240 Zele 0485/16.07.08 ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be
Re: [OSM-talk-be] CORINE Land Cover 2000 (now with permission)
On 18-5-2010 2:04, Ben Laenen wrote: OSM just works like that: start with big things and then get more and more detailed. For 70% of the country Corine is better than existing data, and I welcome anything which is an improvement. There is now a blanket permission from the EEA to use any data available through the EEA website. This includes Corine Land Cover 2000 http://lists.openstreetmap.org/pipermail/imports/2010-June/000594.html http://wiki.openstreetmap.org/wiki/European_Environment_Agency I have the necessary databases set up and the relevant scripts from the French CLC import on my computer, and can work on generating a test import file. I could also generate an overlay showing the result of the import file, after filtering out CLC polygons that overlap with existing OSM polygons. Ben had the notion that perhaps we should, before the import, delete inaccurate polygons, especially in the Ardennes. The proposed overlays could assist in determining whether that is necessary, and if so, at which scale. Now that the legal issues have been resolved, and the technical setup has been mostly performed, the question remains about whether to proceed with an actual import? Don't worry about it happening overnight, we have plenty of time to discuss and review the steps involved. -- Lennard ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be
Re: [OSM-talk-be] CORINE Land Cover 2000
2010/5/19 Martijn van Exel m...@rtijn.org: Hi Nicolas and others, I spoke to Guido d'Hoker (g...@ngi.be) in IGN/NGI yesterday and all I can say is that he was not quite sure about the status of CORINE 2006 - as far as I can see it is included in http://www.eea.europa.eu/data-and-maps/data/corine-land-cover-2006-clc2006-100-m-version-12-2009 but I might not be quite sure what is being meant here. I'll let you investigate further. Anyway, with regard to licensing issues feel free to contact him directly, he can put you in touch with the right people. have you got a direct telephone number that you could send me privately ? (otherwise I'll go to the IGN website) I would like to discuss the availability of some (all ?) of the datas of IGN under free licences (aka compatible with osm). Thanks, Nicolas Martijn martijn van exel +++ m...@rtijn.org laziness - impatience - hubris http://schaaltreinen.nl/ twitter / skype: mvexel flickr: rhodes On Sun, May 16, 2010 at 10:58 PM, Nicolas Pettiaux nico...@pettiaux.be wrote: 2010/5/16 Martijn van Exel m...@rtijn.org: BTW, nice work on the overlay. I think it looks super sweet especially on lower zoom levels. I will be at IGN this week and try and talk to some people there. let us know what comes out : could you ask who we should contact, what are the blocking points for ING to release its data as the French and British did ? THanks, Nicolas -- Nicolas Pettiaux, dr. sc. - Gsm 0496 24 55 01 April - « promouvoir et défendre le logiciel libre » - www.april.org Rejoignez maintenant plus de 5000 personnes, associations, entreprises et collectivités qui soutiennent notre action ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be -- Nicolas Pettiaux, dr. sc. - Gsm 0496 24 55 01 April - « promouvoir et défendre le logiciel libre » - april.org Abppc - « promouvoir les échanges entre professeurs de physique et de chimie » - abppc.be ArbaRivabella - « une école à dimension humaine qui favorise le développement de la personne » - arbarivabella.be ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be
Re: [OSM-talk-be] CORINE Land Cover 2000
On 19-5-2010 7:01, Martijn van Exel wrote: can say is that he was not quite sure about the status of CORINE 2006 - as far as I can see it is included in http://www.eea.europa.eu/data-and-maps/data/corine-land-cover-2006-clc2006-100-m-version-12-2009 but I might not be quite sure what is being meant here. What you see there is a bitmapped version of CLC 2006, with 100 m resolution. There is also a 250 m version there, and bitmapped images of the changes going from 2000 to 2006. The seamless vector data, however, is not available, and that is the version we would need. -- Lennard ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be
Re: [OSM-talk-be] CORINE Land Cover 2000
On Wed, May 19, 2010 at 9:33 AM, Lennard l...@xs4all.nl wrote: On 19-5-2010 7:01, Martijn van Exel wrote: can say is that he was not quite sure about the status of CORINE 2006 - as far as I can see it is included in http://www.eea.europa.eu/data-and-maps/data/corine-land-cover-2006-clc2006-100-m-version-12-2009 but I might not be quite sure what is being meant here. What you see there is a bitmapped version of CLC 2006, with 100 m resolution. There is also a 250 m version there, and bitmapped images of the changes going from 2000 to 2006. The seamless vector data, however, is not available, and that is the version we would need. Oh of course. I did not notice that this was only the raster dataset provided here. Martijn martijn van exel +++ m...@rtijn.org laziness - impatience - hubris http://schaaltreinen.nl/ twitter / skype: mvexel flickr: rhodes ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be
Re: [OSM-talk-be] CORINE Land Cover 2000
Hi, I also looked at this import but I abandoned it for several reasons: - 25ha may be good for France or even Holland but for a region like Flanders it is simply too big: we have no land use planning (ruimtelijke ordening) here whatsoever. So land use tends to be small scattered patches. - The French conducted this import with the rule: if there is more than 2% overlap with an existing region, the CORINE data will not be imported. This is very crude and you will miss some good data, but I understand the reasoning: overwriting the knowledge of local mappers should be avoided. - The last two reasons show that this import will result in data which will be far from perfect, but imported data will work discouraging for corrections:if something is present people tend to have respect for it and not change it, they tend not even to question it. To sum it up: although it might seem great to have all this colour on our map within a few hours, we should very carefully consider it. There are numerous examples of hasty imports which introduced some strange quirks, quirks which are very hard to get rid of afterwards. Kenny 2010/5/16 Lennard l...@xs4all.nl As some of you may know, last year the French have imported the CORINE Land Cover 2006 (CLC 2006) dataset. This dataset aims to classify every bit of surface of the participating countries. It is a project of the European Environment Agency (EEA) in conjunction with national mapping agencies. For Belgium, I believe this would be the NGI/IGN. While they did get approval to import their CLC 2006 dataset, getting approval from the NGI/IGN in Belgium would probably be problematic, and actually getting the data from them even worse. Although, if we don't ask, we'll never know for sure. (Anyone up for contacting them?) However, the CLC 2000 dataset is fully available on the EEA website, and has usage terms that seem to be very compatible with OSM. Unless otherwise indicated, re-use of content on the EEA website for commercial or non-commercial purposes is permitted free of charge, provided that the source is acknowledged. On IRC, someone wanted to contact the EEA explicitly to obtain permission (for Belgium, but why not ask this for the entire dataset?). I don't see a problem with that, although for me the terms on the EEA site are clear. I have made an overlay which shows the CLC 2000 data on the OSM map: http://mijndev.openstreetmap.nl/~ldp/clc2000/http://mijndev.openstreetmap.nl/%7Eldp/clc2000/ This is so you can already enjoy what we may be able to import, and maybe use it as a backdrop in Potlatch/JOSM or other editors to use. I don't recommend actually tracing the polygons, as a direct import is much easier. The amount of changes between CLC 2000 and CLC 2006 is on the order of 0.5% over the entire covered region. The overwhelming amount of polygons would be unchanged. I think there's not much harm in working with CLC 2000, even if it is 10+ years old. I have received the scripts used for the French CLC 2006 import, and can adapt and use those for an import of CLC 2000 for any area. I'm going to test them and prepare an import file for Belgium, which can be imported later on, if no valid objections are lodged. -- Lennard ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be
Re: [OSM-talk-be] CORINE Land Cover 2000
On 5/17/10, Kenny Knecht kenny.kne...@gmail.com wrote: Hi, I also looked at this import but I abandoned it for several reasons: - 25ha may be good for France or even Holland but for a region like Flanders it is simply too big: we have no land use planning (ruimtelijke ordening) here whatsoever. So land use tends to be small scattered patches. - The French conducted this import with the rule: if there is more than 2% overlap with an existing region, the CORINE data will not be imported. This is very crude and you will miss some good data, but I understand the reasoning: overwriting the knowledge of local mappers should be avoided. The idea is that you manually check the areas where Corine didn't overwrite existing data. So this data can be added later. Albeit with a little work. - The last two reasons show that this import will result in data which will be far from perfect, but imported data will work discouraging for corrections:if something is present people tend to have respect for it and not change it, they tend not even to question it. Then work on changing the mindset of mappers? I don't see this really as an existing problem anyway. I’ve seen big things changing on the map before which wouldn’t happen if the mapper was scared of changing something. Sometimes I even think that some mappers should be a little more carefull. And there are already some big polygons with landuse like the forests in the Ardennes. So why would this just be a problem of Corine and not the existing data? To sum it up: although it might seem great to have all this colour on our map within a few hours, we should very carefully consider it. There are numerous examples of hasty imports which introduced some strange quirks, quirks which are very hard to get rid of afterwards. I agree the data isn't perfect. But I don't see better options to get data like this in areas where yahoo satellite images aren't available for now. And for most of those areas it will actually enhance the existing landuse which is now usually very crudely drawn. If somehow at a later date better data becomes available, it’s always possible to remove Corine again. OSM just works like that: start with big things and then get more and more detailed. For 70% of the country Corine is better than existing data, and I welcome anything which is an improvement. Greetings Ben ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be
Re: [OSM-talk-be] CORINE Land Cover 2000
I fail to see why there would be a problem, legally, to import CORINE if the license is compatible. IGN/NGI would not be a party in this, as the EEA licenses CORINE. If it would be actually useful is another matter altogether. I think it would look nice on lower zoom levels, and could be useful as a starting point for individual mappers to improve upon, but more detailed data may already be available which would make a blanket import action undesirable. Martijn Martijn van Exel +++ m...@rtijn.org Laziness – Impatience – Hubris http://schaaltreinen.nl twitter: mvexel skype: mvexel flickr: rhodes On May 16, 2010, at 9:47 PM, Lennard wrote: As some of you may know, last year the French have imported the CORINE Land Cover 2006 (CLC 2006) dataset. This dataset aims to classify every bit of surface of the participating countries. It is a project of the European Environment Agency (EEA) in conjunction with national mapping agencies. For Belgium, I believe this would be the NGI/IGN. While they did get approval to import their CLC 2006 dataset, getting approval from the NGI/IGN in Belgium would probably be problematic, and actually getting the data from them even worse. Although, if we don't ask, we'll never know for sure. (Anyone up for contacting them?) However, the CLC 2000 dataset is fully available on the EEA website, and has usage terms that seem to be very compatible with OSM. Unless otherwise indicated, re-use of content on the EEA website for commercial or non-commercial purposes is permitted free of charge, provided that the source is acknowledged. On IRC, someone wanted to contact the EEA explicitly to obtain permission (for Belgium, but why not ask this for the entire dataset?). I don't see a problem with that, although for me the terms on the EEA site are clear. I have made an overlay which shows the CLC 2000 data on the OSM map: http://mijndev.openstreetmap.nl/~ldp/clc2000/ This is so you can already enjoy what we may be able to import, and maybe use it as a backdrop in Potlatch/JOSM or other editors to use. I don't recommend actually tracing the polygons, as a direct import is much easier. The amount of changes between CLC 2000 and CLC 2006 is on the order of 0.5% over the entire covered region. The overwhelming amount of polygons would be unchanged. I think there's not much harm in working with CLC 2000, even if it is 10+ years old. I have received the scripts used for the French CLC 2006 import, and can adapt and use those for an import of CLC 2000 for any area. I'm going to test them and prepare an import file for Belgium, which can be imported later on, if no valid objections are lodged. -- Lennard ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be
Re: [OSM-talk-be] CORINE Land Cover 2000
I know a former director at the Belgian IGN/ING who retired very recently. I'll call him to know how to proceed to have the best and most effective results. Could you please help me by writing all elements (in the OSM wiki) that could help to support the fact that IGN should release its data under a compatible with OSM licence eg. by doing like the UK of France ? (a one page summary of facts would be most welcome). Thanks, Nicolas -- Nicolas Pettiaux, dr. sc. April - « promouvoir et défendre le logiciel libre » - www.april.org ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be
Re: [OSM-talk-be] CORINE Land Cover 2000
On 16-5-2010 22:34, Renaud MICHEL wrote: Well, there's no harm in asking, but as you say it seems compatible with the actual CC-By-SA (but I have no idea for ODBL). I just learned the question went out, but we have not received an answer. That's not so strange, with Ascension Day and a weekend in between. It seems like a great idea, how does the import work? We have already some data imported in Belgium from the french import, will those regions need manual editing to integrate the two imports? Should we keep the landuse where they already exists in OSM belgium, or is the CORINE data much more accurate and should be preferred? The CORINE data is not that bad, but is slightly generalised in the sense that features smaller than 25 ha are not present, and are rolled up as part of a larger area. The way the import in France was done is to compare the CLC polygons to OSM polygons. Only CLC polygons with no overlap or a very small overlap to OSM polygons were automatically imported. The non-imported part of the dataset was then added to a web service, where each can select and export a non-imported polygon and add it to OSM themselves: http://clc.openstreetmap.fr/ -- Lennard ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be
Re: [OSM-talk-be] CORINE Land Cover 2000
On 16-5-2010 22:35, Martijn van Exel wrote: I fail to see why there would be a problem, legally, to import CORINE if the license is compatible. IGN/NGI would not be a party in this, as the EEA licenses CORINE. The NGI/IGN reference is for the CLC 2006 dataset specifically. The 2006 dataset is not publicly available from the EEA and has to be obtained through the participating national agencies. We already received permission to use the CLC 2006 dataset in The Netherlands, but we then have to buy the actual dataset. The point is moot, since we already have the much more detailed 3dShapes data. We can still use CLC (2000) to enrich 3dShapes classifications. For Belgium, it's different, and any broad addition of land cover to OSM would mostly be welcomed, I assume. If it would be actually useful is another matter altogether. I think it would look nice on lower zoom levels, and could be useful as a starting point for individual mappers to improve upon, but more detailed data may already be available which would make a blanket import action undesirable. That's why we're not aiming for a blanket import, but a directed (but still national) import. See my previous reply about the import method and the comparison to existing OSM polygons. -- Lennard ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be
Re: [OSM-talk-be] CORINE Land Cover 2000
On 16-5-2010 21:47, Lennard wrote: The amount of changes between CLC 2000 and CLC 2006 is on the order of 0.5% over the entire covered region. The overwhelming amount of polygons would be unchanged. I think there's not much harm in working with CLC 2000, even if it is 10+ years old. To follow up this part, a list of project details, including a contact person and the changes between CLC 1990 and CLC 2000. If we extrapolate the amount of changes for the 2000-2006 update, we can see that overall the 2000 data would still be pretty valid. http://etc-lusi.eionet.europa.eu/CLC2000/countries/be/full -- Lennard ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be
Re: [OSM-talk-be] CORINE Land Cover 2000
On 5/16/10, Martijn van Exel mve...@gmail.com wrote: If it would be actually useful is another matter altogether. I think it would look nice on lower zoom levels, and could be useful as a starting point for individual mappers to improve upon, but more detailed data may already be available which would make a blanket import action undesirable. That depends on the location of course. In areas where we have Yahoo imagery the existing landuse is usually much more detailed than Corine. Other areas, like the forests drawn in the Ardennes, are less accurate. And in many of those places we’d never be able to reach the accuracy of Corine without satellite images or other datasets. I also wonder whether it would be a good idea to delete some of the less detailed landuse polygons beforehand if the import goes ahead. Ben ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be