Class SoGetBoundingBoxAction

  • public class SoGetBoundingBoxAction
    extends SoAction
    Computes bounding box of a scene. This class is used to compute a 3D bounding box enclosing objects defined by a scene graph. The box is a rectangular prism. The action also computes the center point, which is defined differently for different objects. (For example, the center of an SoFaceSet is the average of its vertices' coordinates.) For a group, the center point is defined as the average of the centers of all shapes in it.

    Each bounding box is calculated as an SbXfBox3f, where the transformation matrix is defined so that the bounding box can be stored in the object space of the SoShape. When two bounding boxes are combined by a group node, the combination is performed so as to produce the smaller object-space box. Note that this does not necessarily yield the smallest world-space-aligned box, especially in the case of few SoShape nodes. The application can accomodate this by using SoBBox nodes.
    The result of the calculation by the action can be returned as an SbXfBox3f or as a world-space-aligned SbBox3f.

    To calculate the bounding box of a subgraph bounded by two paths, specify the left edge of the subgraph with setResetPath(), and apply the action to the path that defines the right edge of the subgraph. The accumulated bounding box and transformation will be reset when the tail of the reset path is traversed.

    If the subgraph being traversed does not contain any shapes, the returned bounding box will be empty (that is, SbBox3f.isEmpty() will return true).

    The SoBBox node can be used to specify the bounding box for a shape or portion of the scene graph. This is useful, for example, to avoid the cost of computing the bounding box of a large (or frequently changing) geometry. The SoBBox node can also be used to specify that a shape or portion of the scene graph should not be included in the bounding box calculation. This is useful, for example, to avoid considering screen aligned "annotation" graphics like legends.

    Bounding boxes are used internally by Open Inventor to optimize rendering, picking, culling and other operations. For example, the standard viewer classes apply an SoGetBoundingBoxAction to the scene graph when automatic adjustment of the near and far clipping planes is enabled (which is the default).

    Bounding boxes are cached by some grouping nodes, primarily SoSeparator. When this action is applied to a scene graph, it will use the cached bounding box whenever possible. So querying the bounding box of a (relatively) static scene graph is normally very fast. Open Inventor automatically detects when bounding caches become invalid and they are re-computed the next time an SoGetBoundingBoxAction is applied.


    • The bounding box will only include shapes that are actually traversed.
      For example the bounding box will not include shapes under an SoSwitch with whichChild set to SO_SWITCH_NONE.
    • The action does not consider the visibility of shapes that are traversed.
      In other words the bounding box will include shapes that are invisible because of an SoDrawStyle, shapes that are clipped out because of an SoClipPlane, etc. Use SoBBox if necessary to exclude shapes from the bounding box computation.
    • Using the correct viewport region is critical when computing the bounding box of SoText2 (and a few other screen space related) nodes.
      However the bounding box of most shape nodes does not depend on the size of the viewport region and an approximate value may be passed to the constructor.

    Sets: SoViewportRegionElement


     // Get bounding box using current viewport on screen
     SoGetBoundingBoxAction action = new SoGetBoundingBoxAction( viewer.getArea().getViewportRegion() );
     action.apply( root );
     SbBox3f bbox = action.getBoundingBox();

    See Also:
    SbBox3f, SbXfBox3f, SoGetMatrixAction, SoBBox
    • Constructor Detail

      • SoGetBoundingBoxAction

        public SoGetBoundingBoxAction​(SbViewportRegion viewportRegion)
        Even though the bounding box computation may not involve a window per se, some nodes need the size of the viewport to determine their size in 3D coordinates. For example the size of an SoText2 is a fixed value in device coordinates, but the corresponding size in 3D depends on the current camera.

        Usually it is convenient to get the actual viewport region from the viewer as shown in the example above. However if the scene graph does not contain any viewport dependent nodes then any values can be used.

    • Method Detail

      • setResetPath

        public void setResetPath​(SoPath path)
        Calls setResetPath(path, true, SoGetBoundingBoxAction.ResetTypes.valueOf( SoGetBoundingBoxAction.ResetTypes.ALL.getValue() )).
      • setResetPath

        public void setResetPath​(SoPath path,
                                 boolean resetBefore)
        Calls setResetPath(path, resetBefore, SoGetBoundingBoxAction.ResetTypes.valueOf( SoGetBoundingBoxAction.ResetTypes.ALL.getValue() )).
      • getCenter

        public SbVec3f getCenter()
        Returns computed center point in world space.
      • setInCameraSpace

        public void setInCameraSpace​(boolean flag)
        Set this flag to true if you want the returned bounding box to be in the space of whatever camera is in the graph. Camera space is defined to have the viewpoint at the origin, with the direction of view along the negative z axis. This space can be used to determine distances of objects from the camera.
      • isInCameraSpace

        public boolean isInCameraSpace()
        Returns camera space flag.
      • resetCenter

        public void resetCenter()
      • setResetPath

        public void setResetPath​(SoPath path,
                                 boolean resetBefore,
                                 SoGetBoundingBoxAction.ResetTypes what)
        If a non-NULL path is specified, the action will reset the computed bounding box to be empty and/or the current transformation to identity. The resetBefore flag indicates whether to perform the reset before or after the tail node of the path is traversed.
      • getXfBoundingBox

        public SbXfBox3f getXfBoundingBox()
        Returns computed bounding box before transformation into world space.
      • setViewportRegion

        public void setViewportRegion​(SbViewportRegion newRegion)
        Sets current viewport region to use for action.
      • getViewportRegion

        public SbViewportRegion getViewportRegion()
        Returns current viewport region to use for action.
      • getBoundingBox

        public SbBox3f getBoundingBox()
        Returns computed bounding box in world space.
      • enableElement

        public static void enableElement​(java.lang.Class<? extends Inventor> t,
                                         int stkIndex)
      • getResetPath

        public SoPath getResetPath()
        Returns the current reset path, or NULL.
      • extendBy

        public void extendBy​(SbXfBox3f box)
      • extendBy

        public void extendBy​(SbBox3d box)
      • extendBy

        public void extendBy​(SbXfBox3d box)
      • checkResetAfter

        public void checkResetAfter()
      • checkResetBefore

        public void checkResetBefore()
      • extendBy

        public void extendBy​(SbBox3f box)
      • setCenter

        public void setCenter​(SbVec3d center,
                              boolean transformCenter)
      • isResetPath

        public boolean isResetPath()
        Returns true if the current reset path is not NULL.
      • isCenterSet

        public boolean isCenterSet()
      • setCenter

        public void setCenter​(SbVec3f center,
                              boolean transformCenter)
      • isResetBefore

        public boolean isResetBefore()
        Returns true if the resetBefore flag was specified for the reset path.