GPUHeightFieldAnalysis Class Reference

Parallelized heightfield analysis. More...

#include <gputerrainsimu.h>

Public Member Functions
	GPUHeightFieldAnalysis ()
	Constructor.
	~GPUHeightFieldAnalysis ()
	Destroys OpenGl buffers and shader program.
ScalarField2	FractionalLaplacian (const HeightField &, int, double)
	Compute the fractional Laplacian.
ScalarField2	Laplacian (const HeightField &)
	Computes the Laplacian of the heightfield in double precision on the GPU.
void	FractionalGradient (const HeightField &, double, int, VectorField2 &, ScalarField2 &)
	Compute fractional gradient and fractional gradient norm fields.
void	Gradient (const HeightField &, VectorField2 &, ScalarField2 &)
	Computes the gradient vector field and gradient norm scalar field of a given heightfield.
ScalarField2	Accessibility (const HeightField &, const double &, int, int=2256)
	Compute accessibility.
ScalarField2	ClearSky (const HeightField &, double, int, int=2256)
	Clear sky.
ScalarField2	Shadow (const HeightField &, const Vector &, int=2256)
	Compute shadows.
ScalarField2	SoftShadows (const HeightField &, const Vector &, double, int=2256)
	Compute soft shadows. The code is the same as the SmoothShadow() function in LibMaya/Shaders/heightfield_raytrace.glsl file.
ScalarField2	Diffuse (const HeightField &, const Vector &, double)
	Compute diffuse shading.
int	CountPits (const HeightField &)
	Compute the amount of pits in a given GPU heightfield.
int	CountPits (GLuint hfBuffer, int nx, int ny)
	Compute the amount of pits in a given GPU heightfield buffer with single precision. Results may differ from a CPU computation to the floating-point precision.
Vector2	GetRange (const HeightField &)
	Computes the range of a heightfield on the GPU.
Vector2	GetRange (GLuint hfBuffer, int nx, int ny)
	Computes the range of a heightfield given an already initialized GPU heightfield buffer.
ScalarField2	Visibility (const HeightField &, const Vector &, int=2256)
	Computes the visibility of the sun on the terrain.

Protected Member Functions
void	Init (int nx, int ny)
	Init function, used in the context of an already existing GPU buffer from the QOpenGL widget.
void	InitForSingleBuffer (const HeightField &)
	Init function from a given heightfield.
void	InitForDoubleBuffer (const HeightField &)
	Init function from a given heightfield.
void	InitForTexture (const HeightField &)
	Init function from a given heightfield.
double	C (int, const double &) const
	Not used in the class, but here as a reminder for the fractional gradient computation.
ScalarField2	GetDataToScalarField () const
	Get float buffer data from the GPU and convert it to the ScalarField2 double format.

Protected Attributes
GLShader	shader
	Glsl program.
GLuint	hfTextureBuffer = 0
	Heightfield texture.
GLBuffer	inHfBuffer
	Heightfield buffer, used by CountPits(), with single precision.
GLBuffer	inHfBufferDouble
	Heightfield buffer, with double precision.
GLBuffer	outBuffer
	Output scalar buffer.
GLBuffer	outVecBuffer
	Output 2D vector buffer.
GLBuffer	outIntBuffer
	Output integer buffer.
GLBuffer	outDoubleBuffer
	Output double buffer.

Detailed Description

Parallelized heightfield analysis.

Constructor & Destructor Documentation

GPUHeightFieldAnalysis()

GPUHeightFieldAnalysis::GPUHeightFieldAnalysis

(

)

Constructor.

Initialize everthing to zero.

Member Function Documentation

Accessibility()

ScalarField2 GPUHeightFieldAnalysis::Accessibility	(	const HeightField &	hf,
		const double &	r,
		int	n,
		int	maxSteps = 2256 )

Compute accessibility.

See also

HeightField::Accessbility

Parameters

hf	Heightfield
r	Radius.
n	Number of rays.
maxSteps	Number of maximum steps for sphere tracing

ClearSky()

ScalarField2 GPUHeightFieldAnalysis::ClearSky	(	const HeightField &	hf,
		double	r,
		int	n,
		int	maxSteps = 2256 )

Clear sky.

Parameters

hf	Heightfield.
r	maximum distance along a ray.
n	Number of samples per cell.
maxSteps	Number of maximum steps for sphere tracing

CountPits() [1/2]

int GPUHeightFieldAnalysis::CountPits

(

const HeightField &

hf

)

Compute the amount of pits in a given GPU heightfield.

This function uses a double precision buffer, so the results should be equivalent to a CPU computation.

Parameters

hf	Heightfield.

CountPits() [2/2]

int GPUHeightFieldAnalysis::CountPits	(	GLuint	hfBuffer,
		int	nx,
		int	ny )

Compute the amount of pits in a given GPU heightfield buffer with single precision. Results may differ from a CPU computation to the floating-point precision.

Parameters

hfBuffer	Heightfield single precision gpu buffer, should already be initialized.
nx,ny	Heightfield resolution.

Diffuse()

ScalarField2 GPUHeightFieldAnalysis::Diffuse	(	const HeightField &	hf,
		const Vector &	l,
		double	s )

Compute diffuse shading.

Parameters

hf	Heightfield.
l	Light direction, should be normalized
s	Shadow strength

FractionalGradient()

void GPUHeightFieldAnalysis::FractionalGradient	(	const HeightField &	hf,
		double	s,
		int	m,
		VectorField2 &	grad,
		ScalarField2 &	gradNorm )

Compute fractional gradient and fractional gradient norm fields.

Parameters

hf	heightfield
s	Fractional parameter, should be in [0,1].
m	mask size
grad	returned fractional gradient field
gradNorm	returned fractional gradient norm

FractionalLaplacian()

ScalarField2 GPUHeightFieldAnalysis::FractionalLaplacian	(	const HeightField &	hf,
		int	m,
		double	s )

Compute the fractional Laplacian.

Parameters

hf	Heightfield
m	Mask size
s	Fractional parameter, should be in [0,1].

GetRange() [1/2]

Vector2 GPUHeightFieldAnalysis::GetRange

(

const HeightField &

hf

)

Computes the range of a heightfield on the GPU.

Parameters

hf	the HeightField

GetRange() [2/2]

Vector2 GPUHeightFieldAnalysis::GetRange	(	GLuint	hfBuffer,
		int	nx,
		int	ny )

Computes the range of a heightfield given an already initialized GPU heightfield buffer.

The architecture is not the most efficient, but it avoids the need to get the data back on the CPU for computing the elevation range, which is slow for large heightfields.

Parameters

hfBuffer	Heightfield gpu buffer, should be initialized.
nx,ny	Heightfield dimensions.

Gradient()

void GPUHeightFieldAnalysis::Gradient	(	const HeightField &	hf,
		VectorField2 &	grad,
		ScalarField2 &	gradNorm )

Computes the gradient vector field and gradient norm scalar field of a given heightfield.

Parameters

hf	the heightfield
grad	returned gradient vector field
gradNorm	returned gardient norm scalar field

Init()

void GPUHeightFieldAnalysis::Init ( int nx, int ny )

protected

Init function, used in the context of an already existing GPU buffer from the QOpenGL widget.

Parameters

nx,ny

Heightfield dimensions, should match the size of the GPU buffer.

InitForDoubleBuffer()

void GPUHeightFieldAnalysis::InitForDoubleBuffer ( const HeightField & hf )

protected

Init function from a given heightfield.

Initialize a double precision GPU buffer.

Parameters

hf	Heightfield

InitForSingleBuffer()

void GPUHeightFieldAnalysis::InitForSingleBuffer ( const HeightField & hf )

protected

Init function from a given heightfield.

Initialize a single precision GPU buffer (and not a texture).

This function is useful when the single precision is not an issue for the calling function. It also allows for a simple API for functions taking both an existing buffer or a heightfield.

Parameters

hf	Heightfield.

InitForTexture()

void GPUHeightFieldAnalysis::InitForTexture ( const HeightField & hf )

protected

Init function from a given heightfield.

Initialize a GPU texture (and not a buffer).

This function is useful when the single precision is not an issue for the calling function.

Parameters

hf	Heightfield.

Laplacian()

ScalarField2 GPUHeightFieldAnalysis::Laplacian

(

const HeightField &

hf

)

Computes the Laplacian of the heightfield in double precision on the GPU.

Parameters

hf	the heightfield

Shadow()

ScalarField2 GPUHeightFieldAnalysis::Shadow	(	const HeightField &	hf,
		const Vector &	l,
		int	maxSteps = 2256 )

Compute shadows.

Parameters

hf	Heightfield.
l	Light direction, should be normalized
maxSteps	Number of maximum steps for sphere tracing

SoftShadows()

ScalarField2 GPUHeightFieldAnalysis::SoftShadows	(	const HeightField &	hf,
		const Vector &	l,
		double	s,
		int	maxSteps = 2256 )

Compute soft shadows. The code is the same as the SmoothShadow() function in LibMaya/Shaders/heightfield_raytrace.glsl file.

Parameters

hf	Heightfield.
l	Light direction, should be normalized
s	Shadow strength
maxSteps	Number of maximum steps for sphere tracing

Visibility()

ScalarField2 GPUHeightFieldAnalysis::Visibility	(	const HeightField &	hf,
		const Vector &	sun,
		int	maxSteps = 2256 )

Computes the visibility of the sun on the terrain.

Parameters

hf	The HeightField
sun	The sun vector
maxSteps	Number of maximum steps for sphere tracing