r.extract.buildings.html

<h2>DESCRIPTION</h2>

<em>r.extract.buildings</em> extracts buildings as vectors and
calculates height statistics (minimum, maximum, average, standard
deviation, median, percentile) and presumable number of stories using
an nDSM-raster, NDVI-raster, and FNK-vector (Flaechennutzungskartierung).
As certain area codes from the FNK are used, the class codes have to be
consistent. The extraction can be based on an image segmentation using
<a href="i.segment.html">i.segment</a>, which requires the <b>-s</b>
flag to be activated. Note that this significantly extends the
processing time.

<p>
A generic NDVI threshold is used to separate buildings from trees.
The NDVI threshold can be defined as a fixed NDVI value (on a scale from
0-255). Alternatively, the threshold can be defined by calculating the n-th
percentile (indicated by the <b>ndvi_perc</b> option) of NDVI values from all
vegetated areas. Vegetated areas are defined from the FNK-vector - therefore
the class codes have to be consistent. For this alternative the parameter
<b>used_thresh</b> must be set to <b>ndvi_perc</b>.
<p>
Only buildings with a defined minimum size and minimum height are
extracted. The average story height is assumed to be 3 meters.
<p>
The extraction works via tiles. Tile size has to be set by the
<b>tile_size</b> option. For serial processing use <b>nprocs=1</b>,
for parallel processing set number of cores to be used (e.g. <b>nprocs=8</b>).
The default uses available cores minus one.
For processing the tiles the addon uses the worker
<a href="r.extract.buildings.worker.html">r.extract.buildings.worker</a>.

<h2>EXAMPLES</h2>

<h3>Extraction using default values and the 5th percentile as NDVI threshold
  and a tile size of 2000m</h3>

<div class="code"><pre>
r.extract.buildings ndsm=nDOM_Bottrop_2017_05m ndvi_raster=ndvi_Bottrop_2017 fnk_vector=FNK_Bottrop_2017 fnk_column=code_akt ndvi_perc=5 output=buildings_Bottrop_2017 memory=10000 tile_size=2000
</pre></div>

<h3>Extraction using a fixed NDVI threshold and a differing minimum size of
  buildings with a tile size of 2000m</h3>

<div class="code"><pre>
r.extract.buildings ndsm=nDOM_Bottrop_2017_05m ndvi_raster=ndvi_Bottrop_2017 fnk_vector=FNK_Bottrop_2017 fnk_column=code_akt min_size=30 ndvi_thresh=145 output=buildings_Bottrop_2017 memory=10000 tile_size=2000
</pre></div>

<h2>SEE ALSO</h2>

<em>
<a href="https://grass.osgeo.org/grass-stable/manuals/r.mapcalc.html">r.mapcalc</a>,
<a href="https://grass.osgeo.org/grass-stable/manuals/i.segment.html">i.segment</a>,
<a href="https://grass.osgeo.org/grass-stable/manuals/r.quantile.html">r.quantile</a>
</em>

<h2>AUTHORS</h2>

Guido Riembauer, <a href="https://www.mundialis.de/">mundialis GmbH &amp; Co. KG</a>
<p>
Julia Haas, <a href="https://www.mundialis.de/">mundialis GmbH &amp; Co. KG</a>