Ubuntu Manpage: 3D raster maps in general

3d raster data in grass gis

Provided by: grass-doc_7.0.3-1build1_all

3D raster data in GRASS GIS

3D raster maps in general
GRASS GIS is one of the few GIS software packages with 3D raster data support. Data are stored as a 3D
raster with 3D cells of a given volume. 3D rasters are designed to support representations of trivariate
continuous fields. The vertical dimension supports spatial and temporal units. Hence space time 3D
raster with different temporal resolutions can be created and processed.

GRASS GIS 3D raster maps use the same coordinate system as 2D raster maps (row count from north to south)
with an additional z dimension (depth) counting from bottom to top. The upper left corner (NW) is the
origin. 3D rasters are stored using a tile cache based approach. This allows arbitrary read and write
operations in the created 3D raster. The size of the tiles can be specified at import time with a given
import module such as r3.in.ascii or the data can be retiled using r3.retile after import or creation.
The 3D raster map coordinate system and tile layout of the RASTER3D library

Terminology and naming
GRASS GIS uses terminology where continuous 3D data represented by regular grid or lattice is called 3D
raster map. 3D raster map works in 3D in the same as (2D) raster map in 2D, so it is called the same
except for the additional 3D. Some literature or other software may use terms such as 3D grid, 3D
lattice, 3D matrix, 3D array, volume, voxel, voxel model, or voxel cube. Note that terms volume and
volumetric often refer to measuring volume (amount) of some substance which may or may not be related to
3D rasters. Note that GRASS GIS uses the term 3D raster map or just 3D raster for short, rather than 3D
raster layer because term map emphasizes the mapping of positions to values which is the purpose of 3D
raster map (in mathematics, map or mapping is close to a term function). While the term layer emphasizes
overlaying or stacking up. The former is not the only only operation done with data and the latter could
be confusing in case of 3D raster data. 3D raster map is divided into cells in the same as the (2D)
raster map. A cell is a cuboid or cube depending on the resolution. The resolution influences volume of
one cell. Some literature or other software may use terms such as volume, volume unit, volumetric pixel,
volume pixel, or voxel. Note that voxel can be sometimes used to refer to a whole 3D raster and that for
example in 3D computer graphics, voxel can denote object with some complicated shape. Type of map and
element name in GRASS GIS is called raster_3d. The module family prefix is r3. Occasionally, for
example in Python, 3D raster related things can be referred differently, for example according to a
programming language standards. In GRASS GIS 3D rasters as stored in tiles which are hidden from user
most of the time. When analyzing or visualizing 3D rasters user can create slices or cross sections.
Slices can be horizontal, vertical, or general plains going through a 3D raster. Slices, especially the
horizontal ones, may be called layers in some literature or some other software. Cross sections are
general functions, e.g. defined by 2D raster, going through a 3D raster. When 3D raster is used in the
way that vertical dimension represents time 3D raster can be referred to as space time cubes (STC) or
space time cube 3D raster. Some literature may also use space time voxel cube, space time voxel model or
some other combination.

3D raster import
The modules r3.in.ascii and r3.in.bin supports generic x,y,z ASCII and binary array import.
Alternatively, 3D rasters can be generated from 3D point vector data (v.to.rast3). Always the full map
is imported. 3D raster can also be created based on 2D elevation map(s) and value raster map(s)
(r.to.rast3elev). Alternatively, a 3D raster can be composed of several 2D raster maps. 2D rasters are
considered as slices in this case and merged into one 3D raster map (r.to.rast3).

3D region settings and 3D MASK
GRASS GIS 3D raster map processing is always performed in the current 3D region settings (see g.region,
-p3 flags), i.e. the current region extent, vertical extent and current 3D resolution are used. If the
3D resolution differs from that of the input raster map(s), on-the-fly resampling is performed (nearest
neighbor resampling). If this is not desired, the input map(s) has/have to be reinterpolated beforehand
with one of the dedicated modules. Masks can be set (r3.mask).

3D raster analyses and operations
Powerful 3D raster map algebra is implemented in r3.mapcalc. A 3D groundwater flow model is implemented
in r3.gwflow.

3D raster conversion to vector or 2D raster maps
Slices from a 3D raster map can be converted to a 2D raster map (r3.to.rast). Cross sectional 2D raster
map can be extracted from 3D raster map based on a 2D elevation map (r3.cross.rast).

3D raster statistics
3D raster statistics can be calculated with r3.stats and r3.univar.

3D raster interpolation
From 3D vector points, GRASS 3D raster maps can be interpolated (v.vol.rst). Results are 3D raster maps,
however 2D raster maps can be also extracted.

3D raster export
The modules r3.out.ascii and r3.out.bin support the export of 3D raster maps as ASCII or binary files.
The output of these modules can be imported with the corresponding import modules noted above.

NetCDF export of 3D raster maps can be performed using the module r3.out.netcdf. It supports 3D raster
maps with spatial dimensions and temporal (vertical) dimension.

GRASS GIS 3D raster maps can be exported to VTK using r3.out.vtk. VTK files can be visualized with the
VTK Toolkit, Paraview and MayaVi. Moreover, GRASS GIS 2D raster maps can be exported to VTK with
r.out.vtk and GRASS GIS vector maps can be exported to VTK with v.out.vtk.

Alternatively, GRASS 3D raster maps can be imported and exported from/to Vis5D (r3.in.v5d, r3.out.v5d).
Note that Vis5D is limited in the number of supported volumes.

3D raster data types
3D raster’s single-precision data type is most often called "FCELL" or "float", and the double-precision
one "DCELL" or "double".

See also
• Introduction into raster data processing

• Introduction into vector data processing

• Introduction into image processing

• Temporal data processing

• Projections and spatial transformations

• wxGUI 3D View Mode

• m.nviz.image

Main index | raster3d index | Topics index | Keywords index | Full index