icl::filter::ImageRectification< T > Class Template Reference

The ImageRectification class can be use to rectify an image quadrangle (IPP accellerated) More...

#include <ImageRectification.h>

Public Member Functions
	ImageRectification (bool validateAndSortPoints=true)
	create an ImageRectification instance with given value for validateAndSortPoints More...
const core::Img< T > &	apply (const utils::Point32f ps[4], const core::Img< T > &src, const utils::Size &resultSize, math::FixedMatrix< float, 3, 3 > *hom=0, math::FixedMatrix< float, 2, 2 > *Q=0, math::FixedMatrix< float, 2, 2 > *R=0, float maxTilt=0, bool advanedAlgorithm=true, const utils::Rect *resultROI=0, const core::scalemode eScaleMode=core::interpolateLIN)
	applies the image rectification from given source image quadrangle into a rectangular image More...

Private Attributes
core::Img< T >	buffer
	internal image buffer More...
bool	validateAndSortPoints
	internal flag More...

Detailed Description

template<class T>
class icl::filter::ImageRectification< T >

The ImageRectification class can be use to rectify an image quadrangle (IPP accellerated)

The ImageRectification can be set up for automatic sorting of the given points (which is default). If this behaviour is disabled, the given points are not checked for their correct order.

It is not implemented as a UnaryOp, because we see no straight forward way to apply source image ROI handling

TODO this could still be implemented as a UnaryOp instance

Constructor & Destructor Documentation

ImageRectification()

template<class T >

icl::filter::ImageRectification< T >::ImageRectification ( bool  validateAndSortPoints = true )

inline

create an ImageRectification instance with given value for validateAndSortPoints

Member Function Documentation

apply()

template<class T >

const core::Img<T>& icl::filter::ImageRectification< T >::apply	(	const utils::Point32f	ps[4],
		const core::Img< T > &	src,
		const utils::Size &	resultSize,
		math::FixedMatrix< float, 3, 3 > *	hom = 0,
		math::FixedMatrix< float, 2, 2 > *	Q = 0,
		math::FixedMatrix< float, 2, 2 > *	R = 0,
		float	maxTilt = 0,
		bool	advanedAlgorithm = true,
		const utils::Rect *	resultROI = 0,
		const core::scalemode	eScaleMode = core::interpolateLIN
	)

applies the image rectification from given source image quadrangle into a rectangular image

Parameters

ps	source image points (must define a linestrip around the source image quadrangular regions that needs to be rectified)
src	source image
resultSize	resulting rectified image size
hom	optionally, you can pass a matrix here, that is filled with the homography matrix that is used to transform the pixels
Q	this can optionally be filled with the Q part of the QR-decomposition of hom The Q part (which is an othogonal matrix) usually contains the homographies rotation
R	this can optionally be filled with the R part of the QR-decomposition of hom The R part (which is an upper right triangluar matrix) usually contains the homographies scale and shear information)
maxTilt	can optionally be given. If it is greater than zero, it is used as maximum ratio between R's diagonal elements.
advanedAlgorithm	selects the algorithm that is used for the creation of the homography internally (

See also

Homography2D)

Parameters

resultROI

if this parameter is given, the final homography is only evaluated within the resulting images ROI

Member Data Documentation

buffer

template<class T >

core::Img<T> icl::filter::ImageRectification< T >::buffer

private

internal image buffer

validateAndSortPoints

template<class T >

bool icl::filter::ImageRectification< T >::validateAndSortPoints

private

internal flag

---

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

• /Users/alneuman/vm/icl/ICLFilter/src/ICLFilter/ImageRectification.h