NAME

By default, it does so by dividing each value by the maximum value of all cells in the input map. If the input data consists only of positive values, then this will result in a rescaling of data to the range "0" (inclusive) to "1" (inclusive). Normalization is not a very well-defined concept when negative values are involved. By default, "r.normalize" will rescale such data to the range "-1" (inclusive) to "1" (inclusive). Use the -a (absolute) flag to rescale such data into the "0" (inclusive) to "1" (inclusive) range instead.

Alternatively, by setting the -l (local) flag, it is possible to perform local normalization. In local normalization mode, the maximum value will be searched for within a limited neighborhood instead of the entire raster input map. The neighborhood is defined as the set of all cells within a user-defined radius around the cell to be normalized.

USAGE DETAILS

Flags:

-a

Absolute: Rescales all data, including negative cell values to the range "0" (inclusive) to "1" (inclusive).

-l

Local mode: Maximum value will be searched for within a limited neighborhood instead of the entire raster input map.

-r

SquaRe neighborhood (only useful in conjunction with the -l flag): Use a square instead of a circular neighborhood for faster but less accurate computation.

-s

Single precision output: Always store output raster map with single floating point precision, no matter what the precision of the input map is.

-q

Disable all progress and status message displays.

input=name

Name of input raster map with values to normalize.

radius=value

Radius of neighborhood to use for computation of local normalization, expressed as number of cells. The larger, the more processing time will be needed.

Local normalization is subject to the "modifiable areal unit" (MAU) effect. Towards the edges of the input raster, results may become unreliable.

AUTHOR

Last changed: 2014-10-04