latest version v1.9 - last update 24 Nov 2005

lti::gradientFunctor Class Reference
[Linear filtering]

The gradient is a simple wrapper for the convolution functor with some convenience parameterization to choose between different common gradient kernels. More...

#include <ltiGradientFunctor.h>

Inheritance diagram for lti::gradientFunctor:

[legend]Collaboration diagram for lti::gradientFunctor:

[legend]List of all members.

Public Member Functions
	gradientFunctor (const bool initializeParameters=true)
	gradientFunctor (const parameters::eOutputFormat &format, const int gradKernelSize=0)
	gradientFunctor (const parameters &par)
	gradientFunctor (const gradientFunctor &other)
virtual	~gradientFunctor ()
virtual const char *	getTypeName () const
bool	apply (const channel &src, channel &xOrMag, channel &yOrArg) const
bool	apply (const channel &src, channel &mag) const
bool	apply (channel &srcdest) const
bool	apply (const channel8 &src, channel &xOrMag, channel &yOrArg) const
gradientFunctor &	copy (const gradientFunctor &other)
gradientFunctor &	operator= (const gradientFunctor &other)
virtual functor *	clone () const
const parameters &	getParameters () const
Protected Member Functions
void	cartToPolar (const int dx, const int dy, float &mag, float &angle) const
One channel gradient computation
bool	computeGradientCart (const channel &src, channel &dx, channel &dy) const
bool	xyDifferentiateImageCart (const channel &src, channel &dx, channel &dy) const
bool	xyDifferentiateImageCartOpt (const channel &src, channel &dx, channel &dy) const
bool	xyDifferentiateImageCartClassicSep (const channel &src, channel &dx, channel &dy) const
bool	xyDifferentiateImageCartClassic2D (const channel &src, channel &dx, channel &dy) const
bool	xyDifferentiateImageCartInterPixel (const channel &src, channel &dx, channel &dy) const
bool	xyDifferentiateImage (const channel8 &src, channel &gradAbs, channel &gradPhi) const
Protected Attributes
arctanLUT	atan2
Classes
class	parameters
	the parameters for the class gradientFunctor More...

---

Detailed Description

The gradient is a simple wrapper for the convolution functor with some convenience parameterization to choose between different common gradient kernels.

Not only the classical simple difference computation (right minus left for the x direction or bottom minus top for the y direction) and the classical Sobel, Prewitt, Robinson, Roberts and Kirsch kernels can be used, but the more sophisticated Ando optimal kernels (see lti::gradientKernelX) and the approximation using oriented gaussian derivatives can be used.

The kernel computes the gradient for lti::channel objects only. For color images you can use one of the derived classes.

See also:

lti::gradientFunctor,lti::gradientKernelX,lti::colorContrastGradient

---

Constructor & Destructor Documentation

lti::gradientFunctor::gradientFunctor (  const bool  initializeParameters = true  )

Default constructor.

lti::gradientFunctor::gradientFunctor (  const parameters::eOutputFormat &  format, const int  gradKernelSize = 0 )

Construct a functor using gradient kernels with the specified values. Parameters: format  specify the format to be use: Cartesic or Polar gradKernelSize  size for the gradient kernel to be used.

lti::gradientFunctor::gradientFunctor (  const parameters &  par  )

Construct a functor using the given parameters.

lti::gradientFunctor::gradientFunctor (  const gradientFunctor &  other  )

copy constructor Parameters: other  the object to be copied

virtual lti::gradientFunctor::~gradientFunctor (   )  [virtual]

destructor

---

Member Function Documentation

bool lti::gradientFunctor::apply (  const channel8 &  src, channel &  xOrMag, channel &  yOrArg )  const

Computes the gradient of the given channel. Parameters: src  channel8 with the source data. xOrMag  the resulting x component or magnitude of the gradient will be left here. yOrArg  the resulting y component or argument (orientation) of the gradient will be left here. Returns: true if apply successful or false otherwise. Reimplemented in lti::colorContrastGradient.

bool lti::gradientFunctor::apply (  channel &  srcdest  )  const

Computes the gradient magnitude of the given channel. Parameters: srcdest  channel with source data, the resulting magnitude will be left here too. Returns: true if apply successful or false otherwise.

bool lti::gradientFunctor::apply (  const channel &  src, channel &  mag )  const

Computes the gradient magnitude of the given channel. Parameters: src  channel with the source data. mag  the resulting magnitude of the gradient will be left here. Returns: true if apply successful or false otherwise.

bool lti::gradientFunctor::apply (  const channel &  src, channel &  xOrMag, channel &  yOrArg )  const

Computes the gradient of the given channel. Parameters: src  channel with the source data. xOrMag  the resulting x component or magnitude of the gradient will be left here. yOrArg  the resulting y component or argument (orientation) of the gradient will be left here. Returns: true if apply successful or false otherwise. Reimplemented in lti::colorContrastGradient.

void lti::gradientFunctor::cartToPolar (  const int  dx, const int  dy, float &  mag, float &  angle )  const [inline, protected]

convert a pair of values in cartesic system into a pair in polar system The LUT takes care that the returned value lies between 0 and 2*Pi

virtual functor* lti::gradientFunctor::clone (   )  const [virtual]

returns a pointer to a clone of this functor. Implements lti::functor. Reimplemented in lti::colorContrastGradient.

bool lti::gradientFunctor::computeGradientCart (  const channel &  src, channel &  dx, channel &  dy )  const [protected]

compute the gradient of the src channel8 using the adecuate method xyDifferentiateXYZ

gradientFunctor& lti::gradientFunctor::copy (  const gradientFunctor &  other  )

copy data of "other" functor. Parameters: other  the functor to be copied Returns: a reference to this functor object

const parameters& lti::gradientFunctor::getParameters (   )  const

returns used parameters Reimplemented from lti::transform. Reimplemented in lti::colorContrastGradient.

virtual const char* lti::gradientFunctor::getTypeName (   )  const [virtual]

returns the name of this type ("gradientFunctor") Reimplemented from lti::transform. Reimplemented in lti::colorContrastGradient.

gradientFunctor& lti::gradientFunctor::operator= (  const gradientFunctor &  other  )

alias for copy member Parameters: other  the functor to be copied Returns: a reference to this functor object

bool lti::gradientFunctor::xyDifferentiateImage (  const channel8 &  src, channel &  gradAbs, channel &  gradPhi )  const [protected]

Differentiate the image in X and Y directions using operator (-1 0 1) The result will be directly in polar coordinates. Parameters: src  source channel gradAbs  magnitude of the gradient. gradPhi  argument of the gradient (angle). Returns: true if successful, false otherwise

bool lti::gradientFunctor::xyDifferentiateImageCart (  const channel &  src, channel &  dx, channel &  dy )  const [protected]

differentiate the image in X and Y directions using operator (-1 0 1) Parameters: src  source channel dx  x component of the gradient dy  y component of the gradient Returns: true if successful, false otherwise

bool lti::gradientFunctor::xyDifferentiateImageCartClassic2D (  const channel &  src, channel &  dx, channel &  dy )  const [protected]

differentiate the image in X and Y directions using classic gradients. Parameters: src  source channel dx  x component of the gradient dy  y component of the gradient Returns: true if successful, false otherwise

bool lti::gradientFunctor::xyDifferentiateImageCartClassicSep (  const channel &  src, channel &  dx, channel &  dy )  const [protected]

differentiate the image in X and Y directions using optimal operators. Parameters: src  source channel dx  x component of the gradient dy  y component of the gradient Returns: true if successful, false otherwise

bool lti::gradientFunctor::xyDifferentiateImageCartInterPixel (  const channel &  src, channel &  dx, channel &  dy )  const [protected]

Roberts operator. calculates the gradient at the center of a 2x2 mask. The position marked with "S" is the position where the gradient is stored at in the output channels: ( S +) ( + +) (The origin (0,0) is at the bottom left of the image) Parameters: src  the already smoothed image dx  x component of the gradient dy  y component of the gradient Returns: returns NULL if successful, a textual error message otherwise

bool lti::gradientFunctor::xyDifferentiateImageCartOpt (  const channel &  src, channel &  dx, channel &  dy )  const [protected]

differentiate the image in X and Y directions using optimal operators. Parameters: src  source channel dx  x component of the gradient dy  y component of the gradient Returns: true if successful, false otherwise

---

Member Data Documentation

arctanLUT lti::gradientFunctor::atan2 [protected]

Look-Up Table for computation of arc tangent.

---

The documentation for this class was generated from the following file:

• ltiGradientFunctor.h

---