Class AbstractCoordinateOperation

  • All Implemented Interfaces:
    Serializable, Deprecable, Lenient­Comparable, Identified­Object, Coordinate­Operation
    Direct Known Subclasses:
    Default­Conversion, Default­Pass­Through­Operation, Default­Transformation

    public class AbstractCoordinateOperation
    extends AbstractIdentifiedObject
    implements CoordinateOperation
    Describes the operation for transforming coordinates in the source CRS to coordinates in the target CRS. Coordinate operations contain a math transform, which does the actual work of transforming coordinates, together with the following information:
    This class is conceptually abstract, even if it is technically possible to instantiate it. Typical applications should create instances of the most specific subclass prefixed by Default instead. An exception to this rule may occur when it is not possible to identify the exact operation type.
    Immutability and thread safety
    This base class is immutable and thus thread-safe if the property values (not necessarily the map itself) given to the constructor are also immutable. Most SIS subclasses and related classes are immutable under similar conditions. This means that unless otherwise noted in the javadoc, Coordinate­Operation instances created using only SIS factories and static constants can be shared by many objects and passed between threads without synchronization.
    See Also:
    Serialized Form

    Defined in the sis-referencing module

    • Method Detail

      • castOrCopy

        public static AbstractCoordinateOperation castOrCopy​(CoordinateOperation object)
        Returns a SIS coordinate operation implementation with the values of the given arbitrary implementation. This method performs the first applicable action in the following choices:
        • If the given object is null, then this method returns null.
        • Otherwise if the given object is an instance of Transformation, Conversion, Projection, Cylindrical­Projection, Conic­Projection, Planar­Projection, Pass­Through­Operation or Concatenated­Operation, then this method delegates to the cast­Or­Copy(…) method of the corresponding SIS subclass. Note that if the given object implements more than one of the above-cited interfaces, then the cast­Or­Copy(…) method to be used is unspecified.
        • Otherwise if the given object is already an instance of Abstract­Coordinate­Operation, then it is returned unchanged.
        • Otherwise a new Abstract­Coordinate­Operation instance is created using the copy constructor and returned. Note that this is a shallow copy operation, since the other properties contained in the given object are not recursively copied.
        object - the object to get as a SIS implementation, or null if none.
        a SIS implementation containing the values of the given object (may be the given object itself), or null if the argument was null.
      • getInterface

        public Class<? extends CoordinateOperation> getInterface()
        Returns the GeoAPI interface implemented by this class. The default implementation returns Coordinate­Operation​.class. Subclasses implementing a more specific GeoAPI interface shall override this method.
        get­Interface in class Abstract­Identified­Object
        the coordinate operation interface implemented by this class.
      • getInterpolationCRS

        public CoordinateReferenceSystem getInterpolationCRS()
        Returns the CRS which is neither the source CRS or target CRS but still required for performing the operation.
        Example: some transformations of vertical coordinates (h) require the horizontal coordinates (φ,λ) in order to interpolate in a grid. This method returns the CRS of the grid where such interpolations are performed.
        the CRS (neither source or target CRS) required for interpolating the values, or null if none.
      • getOperationVersion

        public String getOperationVersion()
        Returns the version of the coordinate operation. Different versions of a coordinate transformation may exist because of the stochastic nature of the parameters. In principle this property is irrelevant to coordinate conversions, but Apache SIS accepts it anyway.
        Specified by:
        get­Operation­Version in interface Coordinate­Operation
        the coordinate operation version, or null in none.
      • getLinearAccuracy

        public double getLinearAccuracy()
        Returns an estimation of positional accuracy in metres, or Na­N if unknown. The default implementation tries to infer a value from the metadata returned by get­Coordinate­Operation­Accuracy() using SIS-specific heuristics.
        Current implementation
        The current implementation uses the heuristic rules listed below. Note that those rules may change in any future SIS version.
        • If at least one quantitative result is found with a linear unit, then returns the largest result value converted to metres.
        • Otherwise if the operation is a conversion, then returns 0 since a conversion is by definition accurate up to rounding errors.
        • Otherwise if the operation is a transformation, then checks if the datum shift were applied with the help of Bursa-Wolf parameters. If a datum shift has been applied, returns 25 meters. If a datum shift should have been applied but has been omitted, returns 3000 meters.
          Note: the 3000 meters value is higher than the highest value (999 meters) found in the EPSG database version 6.7. The 25 meters value is the next highest value found in the EPSG database for a significant number of transformations.
        • Otherwise if the operation is a concatenated operation, returns the sum of the accuracy of all components. This is a conservative scenario where we assume that errors cumulate linearly.
          Note: this is not necessarily the "worst case" scenario since the accuracy could be worst if the math transforms are highly non-linear.
        the accuracy estimation (always in meters), or NaN if unknown.
        See Also:
      • getDomainOfValidity

        public Extent getDomainOfValidity()
        Returns the area or region or timeframe in which this coordinate operation is valid.
        Specified by:
        get­Domain­Of­Validity in interface Coordinate­Operation
        the coordinate operation valid domain, or null if not available.
      • getScope

        public InternationalString getScope()
        Returns a description of domain of usage, or limitations of usage, for which this operation is valid.
        Specified by:
        get­Scope in interface Coordinate­Operation
        a description of domain of usage, or null if none.
      • getMathTransform

        public MathTransform getMathTransform()
        Returns the object for transforming coordinates in the source CRS to coordinates in the target CRS.
        Use with interpolation CRS
        If the interpolation CRS is non-null, then the math transform input coordinates shall by (interpolation, source) tuples: for each value to transform, the interpolation point coordinates shall be first, followed by the source coordinates.
        Example: in a transformation between two vertical CRS, if the source coordinates are (z) values but the coordinate operation additionally requires (x,y) values for interpolation purpose, then the math transform input coordinates shall be (x,y,z) tuples in that order.
        The interpolation coordinates will pass through the operation and appear in the math transform outputs, in the same order than inputs.
        Specified by:
        get­Math­Transform in interface Coordinate­Operation
        the transform from source to target CRS, or null if not applicable.
      • getWrapAroundChanges

        public Set<Integer> getWrapAroundChanges()
        Returns the indices of target dimensions where "wrap around" may happen as a result of this coordinate operation. If such change exists, then this is usually the longitude axis when the source CRS uses the [-180 … +180]° range and the target CRS uses the [0 … 360]° range, or the converse. If there is no change, then this is an empty set.
        Inverse relationship: sometime the target dimensions returned by this method can be mapped directly to wraparound axes in source CRS, but this is not always the case. For example consider the following operation chain:
        source projected CRS ⟶ base CRS ⟶ target geographic CRS
        In this example, a wraparound axis in the target CRS (the longitude) can be mapped to a wraparound axis in the base CRS. But there is no corresponding wraparound axis in the source CRS because the easting axis in projected CRS does not have a wraparound range meaning. We could argue that axis directions match, but such matching is not guaranteed to exist since Projected­CRS is a special case of General­Derived­CRS and derived CRS can have rotations.

        The default implementation infers this set by inspecting the source and target coordinate system axes. It returns the indices of all target axes having Range­Meaning​.WRAPAROUND and for which the following condition holds: a colinear source axis exists with compatible unit of measurement, and the range (taking unit conversions in account) or range meaning of those source and target axes are not the same.

        indices of target dimensions where "wrap around" may happen as a result of this coordinate operation.