How to create a 3D shapefile from a raster?

Question

I´m trying to make a 3D shapefile from a raster feature but not working. Tried already to run the command in OSGeo4W but nothing happen. I'm using the Qgis 1.8.0 Lisboa. The purpose is to get a final 3D dxf file! Can anyone help me?

Answer 1

For the moment, the geometry system of QGIS does not handle z-values but z-coordinates can be read by parsing the binary format of the geometry (see Why is it not possible to extract the z value of a wkb geometry with PyQGIS while it is possible with ogr or Shapely ?) of which I take the example:

With a 3D point shapefile in the Python console of QGIS:

mylayer = qgis.utils.iface.activeLayer()
mylayer.geometryType() == QGis.Point
True
# select the first object
sel = mylayer.selectedFeatures()[0]
qgisgeom = sel.geometry()
qgisgeom.wkbType() == QGis.WKBPoint
False
qgisgeom.wkbType() == QGis.WKBPoint25D
True
# QGIS recognizes that the WKB geometry is 3D but is is impossible to extract the z value 
qgisgeom.asPoint()
(205553,89857.7)
# no z in the functions of geom.asPoint, only x and y in dir()
dir(qgisgeom.asPoint())
['__class__', ..., 'x', 'y']

But using other Python modules like osgeo.ogr or Shapely, you can extract the z value from the binary format of the QGIS geometry:

# with shapeky
from shapely.wkb import loads
wkb = qgisgeom.asWkb()
transfshapely = loads(wkb)
transfshapely.has_z
True
list(transfshapely.coords)
[(205552.628666262, 89857.66932836606, 222.1999969482422)]

# with osgeo.ogr
from osgeo import ogr
wkb = qgisgeom.asWkb()
geom_wkb = ogr.CreateGeometryFromWkb(wkb)
geom_wkb.GetGeometryType() == ogr.wkbPoint25D
True
geom_wkb.GetX(),geom_wkb.GetY(),geom_wkb.GetZ()
(205552.628666262, 89857.66932836606, 222.1999969482422)

So, for create a 3D shapefile in the Python console or in the Python shell you must use modules that are able to handle the z value like ogr, Fiona with Shapely or Pyshp (CreateElevationValues)

Example with Fiona:

import fiona
from shapely.geometry import mapping, Point
# example of a 3D point (Shapely)
point = Point(0, 0, 3)
# schema of the shapefile
schema = {'geometry': '3D Point','properties': {'id': 'int'},}
with fiona.collection('my_shp3D.shp', 'w', 'ESRI Shapefile', schema) as output:
   ## If there are multiple geometries, put the "for" loop here
   feat = {}
   feat['geometry'] = mapping(point)
   feat['properties'] = {'id': 1}
   output.write(feat)

Control:

d = fiona.open('my_shp3D.shp')
d.schema
{'geometry': '3D Point', 'properties': {u'id': 'int'}}
d.next()
{'geometry': {'type': 'Point', 'coordinates': (0.0, 0.0, 3.0)}, 'id': '0', 'properties': {'id': 1}}

As an example of extracting the z value from a DEM and the color value from a geological map, see "Python: Using vector and raster layer in a geological perspective without GIS software", Python: utilisation des couches vectorielles et matricielles dans une perspective géologique, sans logiciel SIG , in French but the Python scripts are universal...