Dear mailing list, I received two more questions about the data, and wanted this just to share with you as well. Q: Most buildings seem to have a 22-25 meters offset, while some other could have a 5-10 meters one. Either the offset is inconsistent which can make the correction a bit harder than expected, or I matched the wrong buildings, but being that large it really requires some correction. We have looked at the offset more in depth and there is no ‘standard’ offset. So there is no other solution than to correct it manually. Q: Some buildings have strange trapezoidal shapes, how are you planning to correct them without accessing the imagery? There are not that many of these. Since we are not able to check back with the original imagery, we’ll have to trust the dataset we want to import for this and will upload the data. Of course by looking at the data manually we would be able to spot areas with real issues with this and then we can still decide to not upload that data.
So yes there is still a lot of manual work to do. Splitting the dataset in chunks of 500 buildings, will allows us to do this proper manual verification. I have just update the wiki and workflow so this is well written down. Best wishes, Jorieke From: Jorieke Vyncke <[email protected]> Sent: 17 July 2020 16:55 To: 'Pierre Béland' <[email protected]>; '[email protected]' <[email protected]> Subject: Re: [Imports] Import building footprints Lopa & Lingo Hi Pierre and all, Thanks for looking at the data and sharing your thoughts! I’ve talked through the proposal with Claire from the OSM DRC community on a few calls and included this feedback in the proposal. Eg to include links to the import to the wiki page of OSM DRC, the conflation of data or not, etc. And of course also our MSF team in Ituri is requesting to have this data in OSM, mostly to use it on OsmAnd, so you could see them as well as local community I discussed with. But yes, there hasn’t been that much discussion on the DRC mailing list or here, hence I was asking you all for some more feedback. So thanks!! The original imagery is with Z_gis at Salzburg University and not available for us to use. The buildings retrieved from the imagery are extracted automatically, but all received a manual treatment from the team at Z_gis. This means that all buildings are very nicely squared or circulized. There are few issues with buildings touching each other or slightly overlapping each other, but the plan is to filter these out with a manual validation through JOSM. We’re still working on the smaller chunks of data. If there would be someone here who could tell me how to do that easily in Qgis, I’d be happy to hear that! Otherwise I have to wait for one of my colleagues to share this with you… Best, Jorieke From: Pierre Béland <[email protected]<mailto:[email protected]>> Sent: 17 July 2020 16:37 To: '[email protected]' <[email protected]<mailto:[email protected]>>; Jorieke Vyncke <[email protected]<mailto:[email protected]>> Subject: Re: [Imports] Import building footprints Lopa & Lingo Hi Jorieke, There was no detailed examination of the data quality on the OSM DRC mailing list. What is the image quality and what is the quality of buildings automatically traced from this imagery ? Are they orthogonalized or not ? We only have access to a DropBox directory to download shapefiles and no access to the imagery used to trace the buildings. I suggest that you provide link to the imagery and access to smaller chunks of data in OSM format has it is planned to be used for the import. Regard Pierre Le vendredi 17 juillet 2020 04 h 34 min 53 s UTC−4, Jorieke Vyncke <[email protected]<mailto:[email protected]>> a écrit : Dear all, If you would have other suggestions or feedback, please let me know. Otherwise we’ll go ahead with the preparation of the import and we’ll start with it after the weekend. Best wishes, Jorieke From: Jorieke Vyncke <[email protected]<mailto:[email protected]>> Sent: 08 July 2020 18:35 To: '[email protected]' <[email protected]<mailto:[email protected]>> Cc: 'Claire Halleux' <[email protected]<mailto:[email protected]>>; Roxana NAZARIE <[email protected]<mailto:[email protected]>>; Nienke Meeuwissen <[email protected]<mailto:[email protected]>> Subject: [Imports] Import building footprints Lopa & Lingo Dear import list, We would love to import a dataset of building footprints in Lopa and Lingo in the Ituri province in the East of DRC. It concerns a dataset of 29,620 building footprints, detected automatically by our partner Salzburg University on imagery of January 2020. We have tried to write an import plan as good as we could, so are now also interested to hear your feedback. Please find the wiki page here: https://wiki.openstreetmap.org/wiki/Lopa_%26_Lingo_building_footprints And the direct link to the raw data here: https://www.dropbox.com/sh/vp6sffkmkcrc51t/AABc4d2uM4XaqGzdTVYgtBpfa?dl=0 I have shared and discussed the proposal already with the OpenStreetMap DRC community. For the discussion on the OSM DRC mailinglist see here: https://lists.openstreetmap.org/pipermail/talk-cd/2020-July/thread.html At the OSM DRC mailing list was suggested to add the source tags on the object itself since there will be happening another import in the area. Could someone have a closer look at them, is the proposed usage of them a right one? Thanks a lot! Jorieke Jorieke VYNCKE Missing Maps Coordinator & OCA GIS Focal Point Manson Unit - MSF UK [MSF_dual_English_CMYK] MSF UK, 10 Furnival Street, London, EC4A 1AB www.msf.org.uk<http://www.msf.org.uk/> || +44 (0)20 7404 6600 || UK Charity Reg. No. 10265888 Don't print if you don't need _______________________________________________ Imports mailing list [email protected]<mailto:[email protected]> https://lists.openstreetmap.org/listinfo/imports
_______________________________________________ Imports mailing list [email protected] https://lists.openstreetmap.org/listinfo/imports
