r.reclass - Creates a new map layer whose category values are based upon the user's reclassification of categories in an existing raster map layer.

r.reclass

r.reclass help

r.reclass input=string output=string [title=string]

Parameters:

r.reclass creates an output map layer based on an input 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 files 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.

Suppose we want to reclassify the raster map layer roads, consisting of five categories, into the three new categories: paved roads, unpaved roads, and railroad tracks. The user is asked whether the reclass table is to be established with each category value initially set to 0, or with each category value initially set to its own value. A screen like that shown below then appears, listing the categories of the roads raster map layer to be reclassified and prompting the user for the new category values to be assigned them.

ENTER NEW CATEGORY NUMBERS FOR THESE CATEGORIES

OLD CATEGORY NAME OLD NEW

NUM NUM

no data 0 0___

Hard Surface, 2 lanes 1 0___

Loose Surface, 1 lane 2 0___

Improved Dirt 3 0___

Unimproved Dirt Trail 4 0___

Railroad, single track 5 0___

AFTER COMPLETING ALL ANSWERS, HIT <ESC> TO CONTINUE

(OR <Ctrl-C> TO CANCEL)

In the following screen the new category values have been entered beside the appropriate old category names. Cells assigned category values 2, 3, and 4 in the old raster map layer are now assigned the new category value 2 in the reclassed map; cell data formerly assigned to category value 5 in the old raster map map are now assigned the new category value 3 in the reclassed map.

ENTER NEW CATEGORY NUMBERS FOR THESE CATEGORIES

OLD CATEGORY NAME OLD NEW

NUM NUM

no data 0 0___

Hard Surface, 2 lanes 1 1___

Loose Surface, 1 lane 2 2___

Improved Dirt 3 2___

Unimproved Dirt Trail 4 2___

Railroad, single track 5 3___

AFTER COMPLETING ALL ANSWERS, HIT <ESC> TO CONTINUE

(OR <Ctrl-C> TO CANCEL)

Hitting the escape key will bring up the following screen, which prompts the user to enter a new TITLE and category label for the newly reclassed categories.

ENTER NEW CATEGORY NAMES FOR THESE CATEGORIES

TITLE: Roads Reclassified

CAT NEW CATEGORY NAME

NUM

0 no data

1 Paved Roads

2 Unpaved Roads

3 Railroad, single track

AFTER COMPLETING ALL ANSWERS, HIT <ESC> TO CONTINUE

(OR <Ctrl-C> TO CANCEL)

Based upon the information supplied by the user in the above sample screens, the new output map, supporting category, color, history, and header files are created.

In non-interactive program use, the names of an input map, output map, and output map TITLE are given on the command line. However, the reclass rules are still read from standard input (i.e., from the keyboard, redirected from a file, or piped through another program).

Once the user has specified an input raster map layer, output map layer name, and (optionally) output map layer TITLE by typing r.reclass input=name output=name [TITLE=name]

Each line of input must have the following format: input_categories=output_category [label]

where the input lines specify 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.

No data have to be spcified with NULL.

A line containing only the word end terminates the input.

The following examples may help clarify the reclass rules.

"

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 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

3. 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

4. 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

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.

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. 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 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 (an ASCII file found under the cellhd directory, or viewable by running r.resample.

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

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 a table referencing some original raster map layer rather than creating a 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.

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.

r.resample, r.rescale, r.recode

James Westervelt,

Michael Shapiro,

U.S.Army Construction Engineering Research Laboratory

Last changed: $Date: 2003/05/06 15:19:41 $

Help Index