[HOT] Padang Mapping Task
kate at maploser.com
Sun Apr 22 14:19:58 BST 2012
It is the same exact satellite shot between the Bing and the Digital
Globe Imagery that HOT bought. If you go look at cars on
intersections they are in the same place.
I would suggest adjusting the Bing imagery offset if you are going to
use it, since most of the rest of the tracing happened long before
high resolution imagery was available of Padang through Bing. We
won't have any idea of knowing which is right until there are GPS
tracks of the area, I don't see that happening anytime very soon.
If adjusting the offset is too complicated then please use the Digital
Globe Imagery so it matches up with everything else.
On Sun, Apr 22, 2012 at 6:36 AM, Martin Holmgren <d95marty at gmail.com> wrote:
> Hi all!
> On May 9th I'll be giving a short introduction to OSM, HOT and mapping using Potlatch 2 to a group of mostly GIS and IT professionals here in West Sweden. My idea is to have the participants take tasks in Padang using the tasking manager and do some tracing of buildings. To me it seems like a perfect introduction both to OSM and HOT as tracing and tagging buildings requires a minimum of knowledge about tagging schemes (basically "building=yes, source=...") and the more complex aspects of the data model.
> In my own mapping I have used both the Digital Globe imagery and Bing Aerials, where I have tried to offset Bing to match the previous traces with Digital Globe as source. Now, reading the (about two months old) discussion below, I guess I rather should have either
> a) according to Jaakko's view: kept Bing in its original position and used the source tag to allow for adjustment at a later stage, when we have background material that is more consistent or
> b) avoided using Bing altogether, based on Kate's suggestion.
> Even if b) would give a more consistent map, I see a couple of problems, both in general and pertaining to my planned mapping activity:
> - the Bing (as opposed to the Digital Globe) imagery is available also on zoom level 19, which makes it considerably easier to do tracing.
> - manually adding the Digital Globe images is an extra barrier to getting going. (ok, in the case of our mapping activity it could be managed, since they will all be computer savvy)
> - with Bing imagery available directly in Potlatch there is a risk that people who either don't use the task manager, or who miss reading the instructions, do tracing using Bing unknowing of the Digital Globe material.
> My question is therefore: is it ok for us to trace using Bing (tagged "source=Bing Aerial") with no offset, or should we still take the extra step of configuring the Digital Globe imagery in Potlatch?
> I'd love to get this as right as possible before having a group of people do a lot of tracing, so any suggestions on what approach to take are very welcome!
> Kind regards, Martin
> 13 feb 2012 kl. 21:26 skrev Jaakko Helleranta.com:
>> This is (one of the) reason(s) I think ppl should use source tag (possibly source:geometry) on all mapped objects. .. Makes adjusting easier as one can filter with the tag when traces or other more solid information on imagery accuracy becomes available.
>> Ps. Talking of traces, imagery, and accuracy I think it would be good to try to figure out how to partner with various development / humanitarian players (and why not private sector players / "even" governments to create a "reference trace set" for any area in which we are active in.
>> I'm trying to figure that out here in Haiti as implied in my msg a few hrs ago.
>> Has some tried to do this somewhere else? .. I remember reading somewhere that something along these lines was (one of) the base data set(s) for OSM in London in the beginning of the whole OSM project.
>> ------Original Message------
>> From: Kate Chapman
>> To: HOT at OSM (Humanitarian OpenStreetMap Team)
>> Subject: [HOT] Padang Mapping Task
>> Sent: Feb 13, 2012 14:27
>> Hey All,
>> Just wanted to make you aware there is an offset between the imagery
>> available in Bing and the imagery we've been using.
>> I think the best thing to do right now is to continue using the
>> imagery we were using. Eventually when there is more GPS traces
>> available we might be able to correct. At the moment though there
>> isn't a good way of knowing which is best.
>> Also if you run across some buildings that are offset I think it is
>> best to line them up with the imagery so it is consistent. If someone
>> has a better suggestion let me know.
>> HOT mailing list
>> HOT at openstreetmap.org
>> Sent from my BlackBerry® device from Digicel
>> Mobile: +509-37-26 91 54, Skype/GoogleTalk: jhelleranta
>> HOT mailing list
>> HOT at openstreetmap.org
> HOT mailing list
> HOT at openstreetmap.org
More information about the HOT