GRASS GIS manual: r.reclass

Note: A new GRASS GIS stable version has been released: GRASS GIS 7.4, available here.
Updated manual page: here

NAME

r.reclass - Reclassify raster map based on category values.
Creates a new raster map whose category values are based upon a reclassification of the categories in an existing raster map.

KEYWORDS

raster, reclassification

SYNOPSIS

r.reclass

r.reclass --help

r.reclass input=name output=name rules=name [title=string] [--overwrite] [--help] [--verbose] [--quiet] [--ui]

Flags:

--overwrite: Allow output files to overwrite existing files
--help: Print usage summary
--verbose: Verbose module output
--quiet: Quiet module output
--ui: Force launching GUI dialog

Parameters:

input=name [required]: Name of raster map to be reclassified
output=name [required]: Name for output raster map
rules=name [required]: File containing reclass rules; '-' for standard input
title=string: Title for output raster map

DESCRIPTION
NOTES
- Reclass Rules
EXAMPLES
- Reclass rules examples
- Usage example
SEE ALSO
AUTHORS

DESCRIPTION

r.reclass creates an output map layer based on an input integer raster map layer. The output map layer will be a reclassification of the input map layer based on reclass rules input to r.reclass, and can be treated in much the same way that raster maps are treated. A TITLE for the output map layer may be (optionally) specified by the user.

The reclass rules are read from standard input (i.e., from the keyboard, redirected from a file, or piped through another program).

Before using r.reclass the user must know the following:

The new categories desired; and, which old categories fit into which new categories.
The names of the new categories.

NOTES

In fact, the r.reclass program does not generate any new raster map layers (in the interests of disk space conservation). Instead, a reclass table is stored which will be used to reclassify the original raster map layer each time the new (reclassed) map name is requested. As far as the user (and programmer) is concerned, that raster map has been created.

r.reclass only works on an integer input raster map; if the input map is instead floating point data, you must multiply the input data by some factor to achieve whole number input data, otherwise r.reclass will round the raster values down to the next integer.

Also note that although the user can generate a r.reclass map which is based on another r.reclass map, the new r.reclass map will be stored in GRASS as a reclass of the original raster map on which the first reclassed map was based. Therefore, while GRASS allows the user to provide r.reclass map layer information which is based on an already reclassified map (for the user's convenience), no r.reclass map layer (i.e., reclass table) will ever be stored as a r.reclass of a r.reclass.

To convert a reclass map to a regular raster map layer, set your geographic region settings to match the settings in the header for the reclass map (with "g.region raster=reclass_map", or viewable by running r.info) and then run r.resample.

r.mapcalc can be used to convert a reclass map to a regular raster map layer as well:

  r.mapcalc "raster_map = reclass_map"

where raster_map is the name to be given to the new raster map, and reclass_map is an existing reclass map.

Because r.reclass generates internally simply a table by referencing some original raster map layer rather than creating a full new reclassed raster map layer, a r.reclass map layer will no longer be accessible if the original raster map layer, upon which it was based, is later removed. Therefore, attempting to remove a raster map layer from which a r.reclass has been derived is only possible if the original map is removed first. Alternatively, a r.reclass map can be removed including its base map by using

g.remove's -b flag.

A r.reclass map is not a true raster map layer. Rather, it is a table of reclassification values which reference the input raster map layer. Therefore, users who wish to retain reclassified map layers must also save the original input raster map layers from which they were generated. Alternatively r.recode can be used.

Category values which are not explicitly reclassified to a new value by the user will be reclassified to NULL.

Reclass Rules

Each line of input must have the following format:

input_categories=output_category [label]

where each line of input specifies the category values in the input raster map layer to be reclassified to the new output_category category value. Specification of a label to be associated with the new output map layer category is optional. If specified, it is recorded as the category label for the new category value. The equal sign = is required. The input_category(ies) may consist of single category values or a range of such values in the format "low thru high." The word "thru" must be present.

To include all (remaining) values the asterix "*" can be used. This rule has to be set as last rule. No further rules are accepted after setting this rule. The special rule "* = *" specifies that all categories not expicitly set by one of the above rules should be passed through unaltered instead of being set to NULL.

Categories to become no data are specified by setting the output category value to "NULL".

A line containing only the word end terminates the input.

EXAMPLES

Reclass rules examples

The following examples may help clarify the reclass rules.

1. This example reclassifies categories 1, 2 and 3 in the input raster map layer "roads" to category 1 with category label "good quality" in the output map layer, and reclassifies input raster map layer categories 4 and 5 to category 2 with the label "poor quality" in the output map layer.

    1 2 3   = 1    good quality
    4 5     = 2    poor quality

2. This example reclassifies categories 1, 3 and 5 in the input raster map layer to category 1 with category label "poor quality" in the output map layer, and reclassifies input raster map layer categories 2, 4, and 6 to category 2 with the label "good quality" in the output map layer. All other values are reclassified to NULL.

    1 3 5   = 1    poor quality
    2 4 6   = 2    good quality
    *       = NULL

3. This example reclassifies input raster map layer categories 1 thru 10 to output map layer category 1, input map layer categories 11 thru 20 to output map layer category 2, and input map layer categories 21 thru 30 to output map layer category 3, all without labels. The range from 30 to 40 is reclassified as NULL.

     1 thru 10	= 1
    11 thru 20	= 2
    21 thru 30	= 3
    30 thru 40  = NULL

4. Subsequent rules override previous rules. Therefore, the below example reclassifies input raster map layer categories 1 thru 19 and 51 thru 100 to category 1 in the output map layer, input raster map layer categories 20 thru 24 and 26 thru 50 to the output map layer category 2, and input raster map layer category 25 to the output category 3.

     1 thru 100	= 1    poor quality
    20 thru 50	= 2    medium quality
    25	        = 3    good quality

5. The previous example could also have been entered as:

     1 thru 19  51 thru 100	= 1    poor quality
    20 thru 24  26 thru 50	= 2    medium quality
    25				= 3    good quality

or as:

     1 thru 19	 = 1    poor quality
    51 thru 100	 = 1
    20 thru 24	 = 2
    26 thru 50	 = 2    medium quality
    25		 = 3    good quality

The final example was given to show how the labels are handled. If a new category value appears in more than one rule (as is the case with new category values 1 and 2), the last label which was specified becomes the label for that category. In this case the labels are assigned exactly as in the two previous examples.

Usage example

In this example, the 21 classes of the landuse map (North Carolina sample dataset) are simplified to 7 classes:

r.category landuse96_28m
0	not classified
1	High Intensity Developed
2	Low Intensity Developed
3	Cultivated
[...]
20	Water Bodies
21      Unconsolidated Sediment

# use this command or save rules with editor in textfile "landuserecl.txt"
echo "0 = NULL
1 2     = 1 developed
3       = 2 agriculture
4 6     = 3 herbaceous
7 8 9   = 4 shrubland
10 thru 18 = 5 forest
20      = 6 water
21      = 7 sediment" > landuserecl.txt

r.reclass input=landuse96_28m output=landclass96_recl \
  rules=landuserecl.txt \
  title="Simplified landuse classes 1996"

# verify result
r.category landuse96_recl
1	developed
2	agriculture
3	herbaceous
4	shrubland
5	forest
6	water
7	sediment

AUTHORS

James Westervelt,
Michael Shapiro
U.S.Army Construction Engineering Research Laboratory

Last changed: $Date: 2016-09-19 03:30:49 -0700 (Mon, 19 Sep 2016) $

SOURCE CODE

Available at: r.reclass source code (history)