GeoBasis-DE LVermGeo LoD1

GeoBasis-DE / LVermGeo LoD1 Import

This page is a work in progress! The content is likely incomplete, inaccurate or empty.

Dataset description

This page documents the LoD1 import for Germany, Land Sachsen-Anhalt. LoD1 datasets are published by federal cadastre offices, modeling buildings with simple 3D structures, generated from footprints ("Hausumrisse"), annotated with elevation, height, address and housenumber (if applicable), and building use (if applicable). For the purpose of this import, we project 3D models back to their 2D footprints, and convert ATKIS functions to OSM building and amenity tags (for exact mapping please refer to the notebook below). Building height and elevation of footprint is also stored OSM tags.

License & permission

License/permission grant (and requirements for attribution in source-tag of changeset) is recorded here: https://wiki.openstreetmap.org/wiki/DE:Permissions/Geobasisdaten_Sachsen-Anhalt.

Import type

To ensure a careful and diligent procedure, a staged import is planned:

1. Import buildings with street and house number address that are not in OSM yet, and do not conflict with existing geometries in OSM.
2. Import buildings without house numbers that do not conflict with existing geometries in OSM.
3. Update existing houses that match geometries on a metric such as 90% overlap as measured by intersection-over-union (Jaccard index).
4. Reconcile conflicting buildings manually and/or publish a list of buildings that are not suitable for automatic reconciliation.

In particular, all imports shall maintain the same source tag (source=GeoBasis-DE / LVermGeo LSA, dl-de/by-2-0, LoD1), and buildings will identify source dataset and dataset date by tags following the schema source:geometry=GeoBasis-DE / LVermGeo LSA, dl-de/by-2-0, LoD1_608_5758_2_ST:DESTLIKA0004H3EV (2019-12-09), so that imported buildings can be easily identified and reviewed.

The initial import will be driven via OSM changeset upload API (for implementation details please refer to the linked notebook below). The notebook however also does emit plain OSM files which can be loaded in JOSM and similar tools for manual conflict resolution and reconciliation. One OSM user expressed interest to adopt the notebook code in this fashion.

Schedule

Considering the complexity of the dataset and the staged import approach outlined above, the following approximate timeline is proposed:

1. Initial import of non-overlapping subset with full addresses, skipping all nontrivial corner cases: T0+weeks (T0=import of first changeset).
2. Import of of remaining non-overlapping subset without addresses, but building elevation, height, function tags: T1+weeks (T1=completion of step 1).
3. Update missing addresses, house numbers and function tags that match existing buildings very closely: T2+months (T2=completion of step 2).
4. Resolution of conflicts, via maproulette or similar semiautomated support for manual mapping: T3+manymonths (T3=completion of step 3).

Example

The following exemplary screenshot from the notebook compares OSM buildings (red, left) with new buildings (blue, right).

File:Http://lists.openstreetmap.org/pipermail/imports/attachments/20220925/1b5bd6c1/attachment-0001.png (please copy image URL in browser as image links see disallowed for edits with this account.)

Community discussion

In advance of the import, community feedback has been requested on the imports@openstreetmap.org mailing list (email 2022-09-25) as well as on the talk-de@openstreetmap.org mailing list (email 2022-10-02). The most active mappers in Land Sachsen-Anhalt in recent months have been approached via OSM profile message (these are only 6 contributors, one of them replied, very positive feedback. Considering that LSA is one of the underserved areas in DE, and this in fact has been motivating driver to plan this import, the low response rate might not come as a surprise).

Code & Tools

A jupyter notebook to compute addresses, house numbers, and footprints from LoD1, and to prepare changeset files for upload is maintained here: https://codeberg.org/j0j/OSM/. All buildings are annotated with tags for elevation, height, building type and function derived from ALKIS code, and city, streetname and housenumber if the building has a house number.

An interactive notebook with slipping map to browse examples with tags as tooltips here: https://j0j.codeberg.page/ (this is an HTML rendering of the notebook above).

Coverage

LoD1 datasets are published as zip-compressed bundle of GML files, for Land Sachsen-Anhalt these are 4485 GML files, each describing one quadrant. The following table lists summary statistics for the top-25 quadrants with buildings in dataset, not in OSM, and buildings with address and housenumber, that are missing in OSM:

	buildings	not_in_osm	not_in_osm_with_addr
Total	1746707	592116	42032
LoD1_680_5694_2_ST	3465	2580	846
LoD1_690_5752_2_ST	2687	2220	828
LoD1_690_5754_2_ST	2783	2259	789
LoD1_676_5748_2_ST	3000	1959	672
LoD1_726_5730_2_ST	3351	2360	659
LoD1_684_5730_2_ST	2489	1503	513
LoD1_694_5770_2_ST	1789	1477	497
LoD1_730_5728_2_ST	1392	1204	490
LoD1_678_5746_2_ST	3199	1943	452
LoD1_680_5752_2_ST	2668	2161	445
LoD1_678_5754_2_ST	1773	1340	399
LoD1_726_5728_2_ST	2488	1647	392
LoD1_710_5682_2_ST	1652	1013	359
LoD1_726_5726_2_ST	1596	929	349
LoD1_678_5748_2_ST	1459	1018	338
LoD1_726_5722_2_ST	2280	1273	329
LoD1_640_5700_2_ST	1127	818	328
LoD1_728_5730_2_ST	2274	1696	328
LoD1_728_5726_2_ST	1866	828	322
LoD1_638_5752_2_ST	2883	1873	295
LoD1_682_5758_2_ST	1092	955	294
LoD1_702_5770_2_ST	902	810	290
LoD1_726_5718_2_ST	2277	1280	288
LoD1_710_5666_2_ST	1697	1202	284

Contact

Please reach out to OSM user J0J (j0j on osm, edits, contrib, heatmap, chngset com.) with feedback, questions, and help.

____