pcl::GaussianKernel Class Reference

Class GaussianKernel assembles all the method for computing, convolving, smoothing, gradients computing an image using a gaussian kernel. More...

#include <pcl/common/gaussian.h>

Public Member Functions
void	compute (float sigma, Eigen::VectorXf &kernel, unsigned kernel_width=MAX_KERNEL_WIDTH) const
	Computes the gaussian kernel and dervative associated to sigma. More...
void	compute (float sigma, Eigen::VectorXf &kernel, Eigen::VectorXf &derivative, unsigned kernel_width=MAX_KERNEL_WIDTH) const
	Computes the gaussian kernel and dervative associated to sigma. More...
void	convolveRows (const pcl::PointCloud< float > &input, const Eigen::VectorXf &kernel, pcl::PointCloud< float > &output) const
	Convolve a float image rows by a given kernel. More...
template<typename PointT >
void	convolveRows (const pcl::PointCloud< PointT > &input, std::function< float(const PointT &p)> field_accessor, const Eigen::VectorXf &kernel, pcl::PointCloud< float > &output) const
	Convolve a float image rows by a given kernel. More...
void	convolveCols (const pcl::PointCloud< float > &input, const Eigen::VectorXf &kernel, pcl::PointCloud< float > &output) const
	Convolve a float image columns by a given kernel. More...
template<typename PointT >
void	convolveCols (const pcl::PointCloud< PointT > &input, std::function< float(const PointT &p)> field_accessor, const Eigen::VectorXf &kernel, pcl::PointCloud< float > &output) const
	Convolve a float image columns by a given kernel. More...
void	convolve (const pcl::PointCloud< float > &input, const Eigen::VectorXf &horiz_kernel, const Eigen::VectorXf &vert_kernel, pcl::PointCloud< float > &output) const
	Convolve a float image in the 2 directions. More...
template<typename PointT >
void	convolve (const pcl::PointCloud< PointT > &input, std::function< float(const PointT &p)> field_accessor, const Eigen::VectorXf &horiz_kernel, const Eigen::VectorXf &vert_kernel, pcl::PointCloud< float > &output) const
	Convolve a float image in the 2 directions. More...
void	computeGradients (const pcl::PointCloud< float > &input, const Eigen::VectorXf &gaussian_kernel, const Eigen::VectorXf &gaussian_kernel_derivative, pcl::PointCloud< float > &grad_x, pcl::PointCloud< float > &grad_y) const
	Computes float image gradients using a gaussian kernel and gaussian kernel derivative. More...
template<typename PointT >
void	computeGradients (const pcl::PointCloud< PointT > &input, std::function< float(const PointT &p)> field_accessor, const Eigen::VectorXf &gaussian_kernel, const Eigen::VectorXf &gaussian_kernel_derivative, pcl::PointCloud< float > &grad_x, pcl::PointCloud< float > &grad_y) const
	Computes float image gradients using a gaussian kernel and gaussian kernel derivative. More...
void	smooth (const pcl::PointCloud< float > &input, const Eigen::VectorXf &gaussian_kernel, pcl::PointCloud< float > &output) const
	Smooth image using a gaussian kernel. More...
template<typename PointT >
void	smooth (const pcl::PointCloud< PointT > &input, std::function< float(const PointT &p)> field_accessor, const Eigen::VectorXf &gaussian_kernel, pcl::PointCloud< float > &output) const
	Smooth image using a gaussian kernel. More...

Static Public Attributes
static const unsigned	MAX_KERNEL_WIDTH = 71

Detailed Description

Class GaussianKernel assembles all the method for computing, convolving, smoothing, gradients computing an image using a gaussian kernel.

The image is stored in point cloud elements intensity member or rgb or...

Author

Nizar Sallem

Definition at line 57 of file gaussian.h.

Member Function Documentation

compute() [1/2]

void pcl::GaussianKernel::compute	(	float	sigma,
		Eigen::VectorXf &	kernel,
		Eigen::VectorXf &	derivative,
		unsigned	kernel_width = MAX_KERNEL_WIDTH
	)		const

Computes the gaussian kernel and dervative associated to sigma.

The kernel and derivative width are adjusted according.

Parameters

[in]	sigma
[out]	kernel	the computed gaussian kernel
[out]	derivative	the computed kernel derivative
[in]	kernel_width	the desired kernel width upper bond

Exceptions

pcl::KernelWidthTooSmallException

compute() [2/2]

void pcl::GaussianKernel::compute	(	float	sigma,
		Eigen::VectorXf &	kernel,
		unsigned	kernel_width = MAX_KERNEL_WIDTH
	)		const

Computes the gaussian kernel and dervative associated to sigma.

The kernel and derivative width are adjusted according.

Parameters

[in]	sigma
[out]	kernel	the computed gaussian kernel
[in]	kernel_width	the desired kernel width upper bond

Exceptions

pcl::KernelWidthTooSmallException

computeGradients() [1/2]

void pcl::GaussianKernel::computeGradients ( const pcl::PointCloud< float > &  input, const Eigen::VectorXf &  gaussian_kernel, const Eigen::VectorXf &  gaussian_kernel_derivative, pcl::PointCloud< float > &  grad_x, pcl::PointCloud< float > &  grad_y  ) const

inline

Computes float image gradients using a gaussian kernel and gaussian kernel derivative.

Parameters

[in]	input	image to compute gardients for
[in]	gaussian_kernel	the gaussian kernel to be used
[in]	gaussian_kernel_derivative	the associated derivative
[out]	grad_x	gradient along X direction
[out]	grad_y	gradient along Y direction

Note

if output doesn't fit in input i.e. output.rows () < input.rows () or output.cols () < input.cols () then output is resized to input sizes.

Definition at line 195 of file gaussian.h.

computeGradients() [2/2]

template<typename PointT >

void pcl::GaussianKernel::computeGradients ( const pcl::PointCloud< PointT > &  input, std::function< float(const PointT &p)>  field_accessor, const Eigen::VectorXf &  gaussian_kernel, const Eigen::VectorXf &  gaussian_kernel_derivative, pcl::PointCloud< float > &  grad_x, pcl::PointCloud< float > &  grad_y  ) const

inline

Computes float image gradients using a gaussian kernel and gaussian kernel derivative.

Parameters

[in]	input	image to compute gardients for
[in]	field_accessor	a field accessor
[in]	gaussian_kernel	the gaussian kernel to be used
[in]	gaussian_kernel_derivative	the associated derivative
[out]	grad_x	gradient along X direction
[out]	grad_y	gradient along Y direction

Note

if output doesn't fit in input i.e. output.rows () < input.rows () or output.cols () < input.cols () then output is resized to input sizes.

Definition at line 217 of file gaussian.h.

convolve() [1/2]

void pcl::GaussianKernel::convolve ( const pcl::PointCloud< float > &  input, const Eigen::VectorXf &  horiz_kernel, const Eigen::VectorXf &  vert_kernel, pcl::PointCloud< float > &  output  ) const

inline

Convolve a float image in the 2 directions.

Parameters

[in]	horiz_kernel	kernel for convolving rows
[in]	vert_kernel	kernel for convolving columns
[in]	input	image to convolve
[out]	output	the convolved image

Note

if output doesn't fit in input i.e. output.rows () < input.rows () or output.cols () < input.cols () then output is resized to input sizes.

Definition at line 149 of file gaussian.h.

References pcl::PointCloud< PointT >::height, and pcl::PointCloud< PointT >::width.

convolve() [2/2]

template<typename PointT >

void pcl::GaussianKernel::convolve ( const pcl::PointCloud< PointT > &  input, std::function< float(const PointT &p)>  field_accessor, const Eigen::VectorXf &  horiz_kernel, const Eigen::VectorXf &  vert_kernel, pcl::PointCloud< float > &  output  ) const

inline

Convolve a float image in the 2 directions.

Parameters

[in]	input	image to convolve
[in]	field_accessor	a field accessor
[in]	horiz_kernel	kernel for convolving rows
[in]	vert_kernel	kernel for convolving columns
[out]	output	the convolved image

Note

if output doesn't fit in input i.e. output.rows () < input.rows () or output.cols () < input.cols () then output is resized to input sizes.

Definition at line 171 of file gaussian.h.

References pcl::PointCloud< PointT >::height, and pcl::PointCloud< PointT >::width.

convolveCols() [1/2]

void pcl::GaussianKernel::convolveCols	(	const pcl::PointCloud< float > &	input,
		const Eigen::VectorXf &	kernel,
		pcl::PointCloud< float > &	output
	)		const

Convolve a float image columns by a given kernel.

Parameters

[in]	input	the image to convolve
[in]	kernel	convolution kernel
[out]	output	the convolved image

Note

if output doesn't fit in input i.e. output.rows () < input.rows () or output.cols () < input.cols () then output is resized to input sizes.

convolveCols() [2/2]

template<typename PointT >

void pcl::GaussianKernel::convolveCols	(	const pcl::PointCloud< PointT > &	input,
		std::function< float(const PointT &p)>	field_accessor,
		const Eigen::VectorXf &	kernel,
		pcl::PointCloud< float > &	output
	)		const

Convolve a float image columns by a given kernel.

Parameters

[in]	input	the image to convolve
[in]	field_accessor	a field accessor
[in]	kernel	convolution kernel
[out]	output	the convolved image

Note

if output doesn't fit in input i.e. output.rows () < input.rows () or output.cols () < input.cols () then output is resized to input sizes.

Definition at line 82 of file gaussian.hpp.

References pcl::PointCloud< PointT >::height, pcl::PointCloud< PointT >::resize(), and pcl::PointCloud< PointT >::width.

convolveRows() [1/2]

void pcl::GaussianKernel::convolveRows	(	const pcl::PointCloud< float > &	input,
		const Eigen::VectorXf &	kernel,
		pcl::PointCloud< float > &	output
	)		const

Convolve a float image rows by a given kernel.

Parameters

[in]	kernel	convolution kernel
[in]	input	the image to convolve
[out]	output	the convolved image

Note

if output doesn't fit in input i.e. output.rows () < input.rows () or output.cols () < input.cols () then output is resized to input sizes.

convolveRows() [2/2]

template<typename PointT >

void pcl::GaussianKernel::convolveRows	(	const pcl::PointCloud< PointT > &	input,
		std::function< float(const PointT &p)>	field_accessor,
		const Eigen::VectorXf &	kernel,
		pcl::PointCloud< float > &	output
	)		const

Convolve a float image rows by a given kernel.

Parameters

[in]	input	the image to convolve
[in]	field_accessor	a field accessor
[in]	kernel	convolution kernel
[out]	output	the convolved image

Note

if output doesn't fit in input i.e. output.rows () < input.rows () or output.cols () < input.cols () then output is resized to input sizes.

Definition at line 49 of file gaussian.hpp.

References pcl::PointCloud< PointT >::height, pcl::PointCloud< PointT >::resize(), and pcl::PointCloud< PointT >::width.

smooth() [1/2]

void pcl::GaussianKernel::smooth ( const pcl::PointCloud< float > &  input, const Eigen::VectorXf &  gaussian_kernel, pcl::PointCloud< float > &  output  ) const

inline

Smooth image using a gaussian kernel.

Parameters

[in]	input	image
[in]	gaussian_kernel	the gaussian kernel to be used
[out]	output	the smoothed image

Note

if output doesn't fit in input i.e. output.rows () < input.rows () or output.cols () < input.cols () then output is resized to input sizes.

Definition at line 236 of file gaussian.h.

smooth() [2/2]

template<typename PointT >

void pcl::GaussianKernel::smooth ( const pcl::PointCloud< PointT > &  input, std::function< float(const PointT &p)>  field_accessor, const Eigen::VectorXf &  gaussian_kernel, pcl::PointCloud< float > &  output  ) const

inline

Smooth image using a gaussian kernel.

Parameters

[in]	input	image
[in]	field_accessor	a field accessor
[in]	gaussian_kernel	the gaussian kernel to be used
[out]	output	the smoothed image

Note

if output doesn't fit in input i.e. output.rows () < input.rows () or output.cols () < input.cols () then output is resized to input sizes.

Definition at line 252 of file gaussian.h.

Member Data Documentation

MAX_KERNEL_WIDTH

const unsigned pcl::GaussianKernel::MAX_KERNEL_WIDTH = 71

static

Definition at line 61 of file gaussian.h.

---

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

• pcl/common/gaussian.h
• pcl/common/impl/gaussian.hpp