A view dependent, quadtree based terrain. More...
#include <TerrainNode.h>
Public Member Functions | |
| TerrainNode (ptr< Deformation > deform, ptr< TerrainQuad > root, float splitFactor, int maxLevel) | |
| Creates a new TerrainNode. | |
| virtual | ~TerrainNode () |
| Deletes this TerrainNode. | |
| vec3d | getDeformedCamera () const |
| Returns the current viewer position in the deformed terrain space (see deform). | |
| const vec4d * | getDeformedFrustumPlanes () const |
| Returns the current viewer frustum planes in the deformed terrain space (see deform). | |
| vec3d | getLocalCamera () const |
| Returns the current viewer position in the local terrain space (see deform). | |
| float | getCameraDist (const box3d &localBox) const |
| Returns the distance between the current viewer position and the given bounding box. | |
| SceneManager::visibility | getVisibility (const box3d &localBox) const |
| Returns the visibility of the given bounding box from the current viewer position. | |
| float | getSplitDistance () const |
| Returns the viewer distance at which a quad is subdivided, relatively to the quad size. | |
| float | getDistFactor () const |
| Returns the ratio between local and deformed lengths at getLocalCamera(). | |
| void | update (ptr< SceneNode > owner) |
| Updates the terrain quadtree based on the current viewer position. | |
| bool | addOccluder (const box3d &occluder) |
| Adds the given bounding box as an occluder. | |
| bool | isOccluded (const box3d &box) |
| Returns true if the given bounding box is occluded by the bounding boxes previously added by addOccluder(). | |
Public Attributes | |
| ptr< Deformation > | deform |
| The deformation of this terrain. | |
| ptr< TerrainQuad > | root |
| The root of the terrain quadtree. | |
| float | splitFactor |
| Describes how the terrain quadtree must be subdivided based on the viewer distance. | |
| bool | splitInvisibleQuads |
| True to subdivide invisible quads based on distance, like visible ones. | |
| bool | horizonCulling |
| True to perform horizon occlusion culling tests. | |
| int | maxLevel |
| The maximum level at which the terrain quadtree must be subdivided (inclusive). | |
Static Public Attributes | |
| static float | groundHeightAtCamera |
| The terrain elevation below the current viewer position. | |
| static float | nextGroundHeightAtCamera |
| The value groundHeightAtCamera will have at the next frame. | |
Protected Member Functions | |
| TerrainNode () | |
| Creates an uninitialized TerrainNode. | |
| void | init (ptr< Deformation > deform, ptr< TerrainQuad > root, float splitFactor, int maxLevel) |
| Initializes this TerrainNode. | |
Private Attributes | |
| vec3d | deformedCameraPos |
| The current viewer position in the deformed terrain space (see deform). | |
| vec4d | deformedFrustumPlanes [6] |
| The current viewer frustum planes in the deformed terrain space (see deform). | |
| vec3d | localCameraPos |
| The current viewer position in the local terrain space (see deform). | |
| float | splitDist |
| The viewer distance at which a quad is subdivided, relatively to the quad size. | |
| float | distFactor |
| The ratio between local and deformed lengths at localCameraPos. | |
| mat2f | localCameraDir |
| Local reference frame used to compute horizon occlusion culling. | |
| float * | horizon |
| Rasterized horizon elevation angle for each azimuth angle. | |
A view dependent, quadtree based terrain.
This class provides access to the terrain quadtree, defines the terrain deformation (can be used to get planet sized terrains), and defines how the terrain quadtree must be subdivided based on the viewer position. This class does not give any direct or indirect access to the terrain data (elevations, normals, texture, etc). The terrain data must be managed by proland::TileProducer, and stored in proland::TileStorage. The link between with the terrain quadtree is provided by the TileSampler class.
| proland::TerrainNode::TerrainNode | ( | ptr< Deformation > | deform, | |
| ptr< TerrainQuad > | root, | |||
| float | splitFactor, | |||
| int | maxLevel | |||
| ) |
Creates a new TerrainNode.
| deform | the terrain deformation. | |
| root | the root of the terrain quadtree. | |
| splitFactor | how the terrain quadtree must be subdivided (see splitFactor). | |
| maxLevel | the maximum level at which the terrain quadtree must be subdivided (inclusive). |
| virtual proland::TerrainNode::~TerrainNode | ( | ) | [virtual] |
Deletes this TerrainNode.
| proland::TerrainNode::TerrainNode | ( | ) | [protected] |
Creates an uninitialized TerrainNode.
| bool proland::TerrainNode::addOccluder | ( | const box3d & | occluder | ) |
Adds the given bounding box as an occluder.
The bounding boxes must be added in front to back order.
| occluder | a bounding box in local (i.e. non deformed) coordinates. |
| float proland::TerrainNode::getCameraDist | ( | const box3d & | localBox | ) | const |
Returns the distance between the current viewer position and the given bounding box.
This distance is measured in the local terrain space (with Deformation::getLocalDist), with altitudes divided by getDistFactor() to take deformations into account.
| vec3d proland::TerrainNode::getDeformedCamera | ( | ) | const |
| const vec4d* proland::TerrainNode::getDeformedFrustumPlanes | ( | ) | const |
| float proland::TerrainNode::getDistFactor | ( | ) | const |
Returns the ratio between local and deformed lengths at getLocalCamera().
| vec3d proland::TerrainNode::getLocalCamera | ( | ) | const |
| float proland::TerrainNode::getSplitDistance | ( | ) | const |
Returns the viewer distance at which a quad is subdivided, relatively to the quad size.
This relative distance is equal to splitFactor for a field of view of 80 degrees and a viewport width of 1024 pixels. It is larger for smaller field of view angles and/or larger viewports.
| SceneManager::visibility proland::TerrainNode::getVisibility | ( | const box3d & | localBox | ) | const |
Returns the visibility of the given bounding box from the current viewer position.
This visibility is computed with Deformation::getVisbility.
| void proland::TerrainNode::init | ( | ptr< Deformation > | deform, | |
| ptr< TerrainQuad > | root, | |||
| float | splitFactor, | |||
| int | maxLevel | |||
| ) | [protected] |
Initializes this TerrainNode.
| deform | the terrain deformation. | |
| root | the root of the terrain quadtree. | |
| splitFactor | how the terrain quadtree must be subdivided (see splitFactor). | |
| maxLevel | the maximum level at which the terrain quadtree must be subdivided (inclusive). |
| bool proland::TerrainNode::isOccluded | ( | const box3d & | box | ) |
Returns true if the given bounding box is occluded by the bounding boxes previously added by addOccluder().
| box | a bounding box in local (i.e. non deformed) coordinates. |
| void proland::TerrainNode::update | ( | ptr< SceneNode > | owner | ) |
Updates the terrain quadtree based on the current viewer position.
The viewer position relatively to the local and deformed terrain spaces is computed based on the given SceneNode, which represents the terrain position in the scene graph (which also contains the current viewer position).
| owner | the SceneNode representing the terrain position in the global scene graph. |
The deformation of this terrain.
In the terrain local space the terrain sea level surface is flat. In the terrain deformed space the sea level surface can be spherical or cylindrical (or flat if the identity deformation is used).
vec3d proland::TerrainNode::deformedCameraPos [private] |
The current viewer position in the deformed terrain space (see deform).
vec4d proland::TerrainNode::deformedFrustumPlanes[6] [private] |
The current viewer frustum planes in the deformed terrain space (see deform).
float proland::TerrainNode::distFactor [private] |
The ratio between local and deformed lengths at localCameraPos.
float proland::TerrainNode::groundHeightAtCamera [static] |
The terrain elevation below the current viewer position.
This field must be updated manually by users (the TileSamplerZ class can do this for you). It is used to compute the 3D distance between the viewer and a quad, to decide whether this quad must be subdivided or not.
float* proland::TerrainNode::horizon [private] |
Rasterized horizon elevation angle for each azimuth angle.
True to perform horizon occlusion culling tests.
mat2f proland::TerrainNode::localCameraDir [private] |
Local reference frame used to compute horizon occlusion culling.
vec3d proland::TerrainNode::localCameraPos [private] |
The current viewer position in the local terrain space (see deform).
The maximum level at which the terrain quadtree must be subdivided (inclusive).
The terrain quadtree will never be subdivided beyond this level, even if the viewer comes very close to the terrain.
float proland::TerrainNode::nextGroundHeightAtCamera [static] |
The value groundHeightAtCamera will have at the next frame.
The root of the terrain quadtree.
This quadtree is subdivided based on the current viewer position by the update method.
float proland::TerrainNode::splitDist [private] |
The viewer distance at which a quad is subdivided, relatively to the quad size.
Describes how the terrain quadtree must be subdivided based on the viewer distance.
For a field of view of 80 degrees, and a viewport width of 1024 pixels, a quad of size L will be subdivided into subquads if the viewer distance is less than splitFactor * L. For a smaller field of view and/or a larger viewport, the quad will be subdivided at a larger distance, so that its size in pixels stays more or less the same. This number must be strictly larger than 1.
True to subdivide invisible quads based on distance, like visible ones.
The default value of this flag is false.
1.6.1