Package org.apache.sis.coverage.grid

Class PixelTranslation

java.lang.Object

org.apache.sis.util.Static

org.apache.sis.coverage.grid.PixelTranslation

All Implemented Interfaces:

Serializable

public final class PixelTranslation
extends Static
implements Serializable

The translation to apply for different values of PixelOrientation or PixelInCell. The translation are returned by a call to one of the following static methods:

• getPixelTranslation(PixelOrientation) for the two-dimensional case.
• getPixelTranslation(PixelInCell) for the n-dimensional case.

This class provides also a few translate(…) convenience methods, which apply the translation on a given MathTransform instance.

Example: if the following code snippet, gridToCRS is an AffineTransform from grid cell coordinates (typically pixel coordinates) to some arbitrary CRS coordinates. In this example, the transform maps pixels center, while the upper left corner is desired. This code will switch the affine transform from the pixel center to upper left corner convention:

Since:

1.0

Version:

1.0

See Also:

• PixelInCell
• PixelOrientation
• Serialized Form

Field Summary

Fields

Modifier and Type	Field	Description
final double	dx	The translation among the x axis relative to pixel center.
final double	dy	The translation among the y axis relative to pixel center.
final org.opengis.metadata.spatial.PixelOrientation	orientation	The pixel orientation for this translation.
private static final Map<org.opengis.metadata.spatial.PixelOrientation,PixelTranslation>	ORIENTATIONS	The offset for various pixel orientations.
private static final long	serialVersionUID	For cross-version compatibility.
private static final org.opengis.referencing.operation.MathTransform[]	translations	Math transforms created by translate(MathTransform, PixelInCell, PixelInCell) for dimensions 1 to 6.

Constructor Summary

Constructors

Modifier	Constructor	Description
private	PixelTranslation(org.opengis.metadata.spatial.PixelOrientation orientation, double dx, double dy)	Creates a new pixel translation.

Method Summary

Modifier and Type	Method	Description
private static void	add(org.opengis.metadata.spatial.PixelOrientation orientation, double dx, double dy)	For ORIENTATIONS construction only.
static org.opengis.metadata.spatial.PixelOrientation	getPixelOrientation(org.opengis.referencing.datum.PixelInCell anchor)	Returns the pixel orientation which is equivalent to the given PixelInCell code.
static PixelTranslation	getPixelTranslation(org.opengis.metadata.spatial.PixelOrientation anchor)	Returns the specified position relative to the pixel center.
static double	getPixelTranslation(org.opengis.referencing.datum.PixelInCell anchor)	Returns the position relative to the cell center.
private static IllegalArgumentException	illegalDimension(String name, Object dimension)	Formats an exception for an illegal dimension.
String	toString()	Returns a string representation of this pixel translation.
static org.opengis.referencing.operation.MathTransform	translate(org.opengis.referencing.operation.MathTransform gridToCRS, org.opengis.metadata.spatial.PixelOrientation current, org.opengis.metadata.spatial.PixelOrientation desired, int xDimension, int yDimension)	Converts a math transform from a "pixel orientation" convention to another "pixel orientation" convention.
static org.opengis.referencing.operation.MathTransform	translate(org.opengis.referencing.operation.MathTransform gridToCRS, org.opengis.referencing.datum.PixelInCell current, org.opengis.referencing.datum.PixelInCell desired)	Converts a math transform from a "pixel in cell" convention to another "pixel in cell" convention.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

Field Details

serialVersionUID

private static final long serialVersionUID

For cross-version compatibility.

See Also:

• Constant Field Values

translations

private static final org.opengis.referencing.operation.MathTransform[] translations

Math transforms created by translate(MathTransform, PixelInCell, PixelInCell) for dimensions 1 to 6. Each element in this array will be created when first needed. Even indices are translations by -0.5 while odd indices are translations by +0.5.

orientation

public final org.opengis.metadata.spatial.PixelOrientation orientation

The pixel orientation for this translation. Most common values are PixelOrientation.UPPER_LEFT and PixelOrientation.CENTER.

dx

public final double dx

The translation among the x axis relative to pixel center. The value is typically −½, 0 or +½.

dy

public final double dy

The translation among the y axis relative to pixel center. The value is typically −½, 0 or +½.

ORIENTATIONS

private static final Map<org.opengis.metadata.spatial.PixelOrientation,PixelTranslation> ORIENTATIONS

The offset for various pixel orientations. Keys must be upper-case names.

Constructor Details

PixelTranslation

private PixelTranslation(org.opengis.metadata.spatial.PixelOrientation orientation,
 double dx,
 double dy)

Creates a new pixel translation.

Method Details

add

private static void add(org.opengis.metadata.spatial.PixelOrientation orientation,
 double dx,
 double dy)

For ORIENTATIONS construction only.

getPixelOrientation

public static org.opengis.metadata.spatial.PixelOrientation getPixelOrientation(org.opengis.referencing.datum.PixelInCell anchor)

Returns the pixel orientation which is equivalent to the given PixelInCell code. This equivalence can be used for converting n-dimensional parameters to the more specific two-dimensional case. This method implements the following mapping:

Pixel orientation equivalences

Pixel in cell	Pixel orientation
CELL_CENTER	CENTER
CELL_CORNER	UPPER_LEFT
null	null

Parameters:

anchor - the PixelInCell code, or null.

Returns:

the corresponding pixel orientation, or null if the argument was null.

Throws:

IllegalArgumentException - if the given anchor is not a known code list value.

getPixelTranslation

public static double getPixelTranslation(org.opengis.referencing.datum.PixelInCell anchor)

Returns the position relative to the cell center. This method is typically used for n-dimensional grids, where the number of dimension is unknown. The translation is determined from the following table, with the same value applied to all dimensions:

Translations

Pixel in cell	offset
CELL_CENTER	0.0
CELL_CORNER	-0.5

Parameters:

anchor - the "pixel in cell" value.

Returns:

the translation for the given "pixel in cell" value.

Throws:

IllegalArgumentException - if the given anchor is not a known code list value.

getPixelTranslation

public static PixelTranslation getPixelTranslation(org.opengis.metadata.spatial.PixelOrientation anchor)

Returns the specified position relative to the pixel center. This method can be used for grid restricted to 2 dimensions. The translation vector is determined from the following table:

Translations

Pixel orientation	dx	dy
CENTER	0.0	0.0
UPPER_LEFT	-0.5	-0.5
UPPER_RIGHT	+0.5	-0.5
LOWER_LEFT	-0.5	+0.5
LOWER_RIGHT	+0.5	+0.5

Parameters:

anchor - the pixel orientation.

Returns:

the position relative to the pixel center.

Throws:

IllegalArgumentException - if the given anchor is not a known code list value.

translate

public static org.opengis.referencing.operation.MathTransform translate(org.opengis.referencing.operation.MathTransform gridToCRS,
 org.opengis.referencing.datum.PixelInCell current,
 org.opengis.referencing.datum.PixelInCell desired)

Converts a math transform from a "pixel in cell" convention to another "pixel in cell" convention. This method concatenates −½, 0 or +½ translations on all dimensions before the given transform. If the two given conventions are the same, then this method returns the given transform unchanged.

Example: if a given gridToCRS transform was mapping the cell corner to "real world" coordinates, then a call to translate(gridToCRS, CELL_CORNER, CELL_CENTER) will return a new transform performing the following steps: first convert grid coordinates from cell center convention (desired) to cell corner convention (current), then concatenate the given gridToCRS transform which was designed for the cell corner convention. The above-cited cell center → cell corner conversion is done by translating the grid coordinates by +½, because the grid coordinates (0,0) relative to cell center is (½,½) relative to cell corner.

If the given gridToCRS is null, then this method ignores all other arguments and returns null. Otherwise current and desired arguments must be non-null.

Parameters:

gridToCRS - a math transform from pixel coordinates to any CRS, or null.

current - the pixel orientation of the given gridToCRS transform.

desired - the pixel orientation of the desired transform.

Returns:

the translation from current to desired, or null if gridToCRS was null.

Throws:

IllegalArgumentException - if current or desired is not a known code list value.

translate

public static org.opengis.referencing.operation.MathTransform translate(org.opengis.referencing.operation.MathTransform gridToCRS,
 org.opengis.metadata.spatial.PixelOrientation current,
 org.opengis.metadata.spatial.PixelOrientation desired,
 int xDimension,
 int yDimension)

Converts a math transform from a "pixel orientation" convention to another "pixel orientation" convention. This method concatenates −½, 0 or +½ translations on two dimensions before the given transform. The given transform can have any number of input and output dimensions, but only two of them will be converted.

Example: if a given gridToCRS transform was mapping the upper-left corner to "real world" coordinates, then a call to translate(gridToCRS, UPPER_LEFT, CENTER, 0, 1) will return a new transform translating grid coordinates by +0.5 before to apply the given gridToCRS transform. See example in above translate method for more details.

If the given gridToCRS is null, then this method ignores all other arguments and returns null. Otherwise current and desired arguments must be non-null.

Parameters:

gridToCRS - a math transform from pixel coordinates to any CRS, or null.

current - the pixel orientation of the given gridToCRS transform.

desired - the pixel orientation of the desired transform.

xDimension - the dimension of x coordinates (pixel columns). Often 0.

yDimension - the dimension of y coordinates (pixel rows). Often 1.

Returns:

the translation from current to desired, or null if gridToCRS was null.

Throws:

IllegalArgumentException - if current or desired is not a known code list value.

illegalDimension

private static IllegalArgumentException illegalDimension(String name,
 Object dimension)

Formats an exception for an illegal dimension.

toString

public String toString()

Returns a string representation of this pixel translation.

Overrides:

toString in class Object