Hi all
First of all, sorry for posting in English, but I’m French and it’s been too long since my German classes.
I have a question regarding the source of some data in Germany.
In the city of Bremen to be more precise. I wonder where the buildings come from.
In France, we use the cadaster from IGN (with official permission), but I have found no online cadaster for Bremen, and the buildings don’t seem to precisely match the satellite imagery.
Thanks for your help !
Do you have an example? I just checked some buildings, they are matching the BING imagery.
Hi cocobingo,
in Germany the community prefers manual mapping over mechanical imports. This concerns buildings just as well as streets etc.
Details of buildings are as exact as wanted from the mapping people and especially in rural areas a rectangle on private ground is very often seen and not more details wanted. Footways to the entrance or garages are very often not marked and sometimes even removed.
We have also similar data like cadasters from official side we are allowed to use. For myself I cant remember a mechanical import of cadaster data. Most mappers use them only as background info in JOSM and tag the buildings and infos manually using templates and plug-ins.
One advantage is, that you will hardly find “addr:” tags without buildings over great areas or doubled imported POI’s or chaos tagging/mapping like in USA and Japan. You can build up an area which is much more fun than just cleaning up.
Another source is the Aerowest aerial imagery. These images are available for most big cities in Germany and provide high quality images with 5 and 10 cm resolution.
I just checked a random location in Bremen (here: http://tools.geofabrik.de/mc/#18/53.0871/8.8384&num=3&mt0=mapnik&mt1=google-satellite&mt2=nokia-satellite ), and could not see huge differences to imagery. Maybe that is different in other parts of the city, because, as already mentioned, OSM germany does not like imports. So data quality is highly dependent on the person who added the data. The upside of this is that we do not suffer from errors in the official data, and use mor manual labor. More labor=more people and a larger, more active community.
Also, because is germany regional autonomy is very high, there are different access restrictions on cadastre data in different Bundesländer. Nordrhein-Westphalen, for example, has permitted the use of ALKIS (the cadastre) as background imagery for tracing. Other Bundesländer, as far as I know, have not.
Sorry, I had written a lengthy answer but I must have wrongly submitted the form.
So, in shorter terms, here is why I came to wonder about this :
How could the editor guess the separation between the buildings, based only on the aerial imagery ? Check here for instance.
There is almost no hint has to where the buildings should be split. On some buildings you can see pathways, but on others not. You could think that the buildings have been split evenly, but if you look close to the center of the map, 9/11/13/15 Gevekohtstrasse, this is not the case.
How can you split the buildings this way if you have no cadastre information ? There are not hints on the roofs to help you here.
You could think that the editor knew this place, but it is the same almost everywhere I look. I see nicely split buildings everywhere, and that’s what made me wonder where this information came from if not from cadastre ?
Hi,
the splitting of buildings is often done with the terracer plugin. So, if there is one building with housenumbers 9-17, the mapper assumes that there are 5 single buildings (9,11,13,15,17) and uses the plugin to split the building.
But this can not explain your example, where building #9 is smaller than the others. 
Yes, that’s why I chose this example. 
I noticed that many buildings were evenly split, but a lot of them seem to have “real” splitting that I cannot explain without cadastre information.
I would then maybe assume local knowledge. “I know this middle house is smaller”. Maybe the different building sizes are not correct when compared to a cadastre? looked at google street view, No 9 is smaller. But thats just me grabbing at straws ;-). Maybe someone illegally used a cadastre, maybe someone illegally used street-level photography of a certain competitor, maybe there is mapillary imagery available? Maybe they used own photography?