org.geotoolkit.image.jai
Interface Combine.Transform

Enclosing class:

Combine

public static interface Combine.Transform

Transforms the sample values for one pixel during a "Combine" operation. The method transformSamples(double[]) is invoked by Combine.computeRect(...) just before the sample values are combined as:

values[0]*row[0] + values[1]*row[1] + values[2]*row[2] + ... + row[sourceBands]

This interface provides a hook where non-linear transformations can be performed before the linear one. For example, the transformSamples method could substitutes some values by their logarithm.

Since:

2.1

Version:

3.00

Author:

Martin Desruisseaux (IRD)

Module:

coverage/geotk-coverage (download)

View source code for this class

Method Summary

boolean	isSeparable() Returns true if the transformation performed by transformSamples(double[]) does not depend on the ordering of samples in the values array.
void	transformSamples(double[] values) Transforms the sample values for one pixel before the linear combination.

Method Detail

transformSamples

void transformSamples(double[] values)

Transforms the sample values for one pixel before the linear combination.

Parameters:

values - The sample values to transformation. Transformation are performed in-place.

isSeparable

boolean isSeparable()

Returns true if the transformation performed by transformSamples(double[]) does not depend on the ordering of samples in the values array. This method can returns true if the transformSamples(double[]) implementation meet the following conditions:

• The transformation is separable, i.e. the output value values[i] depends only on the input value values[i] for all i.
• The transformation do not depends on the value of the index i.

For example, the following implementations meets the above mentioned conditions:

for (int i=0; i<values.length; i++) {
    values[i] = someFunction(values[i]);
}

A true value will allows some optimizations inside the Combine.computeRect(...) method. This method may conservatly returns false if this information is unknown.

Returns:

true if the transformation does not depend on the ordering of sample values.