mds::img::CCanny< I > Class Template Reference

Canny edge detector. More...

#include <mdsCanny.h>

Inheritance diagram for mds::img::CCanny< I >:

List of all members.

Public Types
enum	{ TEMPLATE_PARAMETER_IS_NOT_IMAGE = I::CLASS_IMAGE }
	Check that I is an image. More...
typedef CImageEdgeDetector< I >	base
	Image edge detector base.
typedef base::tImage	tImage
	Image type.
typedef base::tPixel	tPixel
	Image pixel type.
Public Member Functions
	CCanny (double dSigma, double dT1, double dT2)
	Default constructor.
double	getSigma () const
	Returns smoothing filter standard deviation.
void	getThresholds (double &dT1, double &dT2)
	Returns thresholds of the non-maximal suppression algorithm.
bool	operator() (const tImage &SrcImage, tImage &DstImage)
	Edge detection in a given image.
void	setSigma (double dSigma)
	Sets the smoothing filter standard deviation.
void	setThresholds (double dT1, double dT2)
	Sets thresholds of the non-maximal suppression algorithm.
	~CCanny ()
	Destructor.
Protected Member Functions
bool	checkNeighbours (tImage &Image, tSize x, tSize y, tPixel T)
	Returns true if any neighbouring pixel has greater value than a given threshold.
void	hysteresis (tImage &MagnitudeImage, tImage &Image, tPixel T1, tPixel T2)
	Hystersis controlled by two given thresholds.
void	nonMaxSuppression (const tImage &GradImageX, const tImage &GradImageY, tImage &MagnitudeImage, tImage &Image)
	Non-maximal suppression.
Protected Attributes
double	m_dT1
	Thresholds used by hysteresis algorithm.
double	m_dT2
CGaussFilter< tImage >	m_GaussFilter
	Gaussian smoothing filter.
CSobelX< tImage, CShiftZero2Gray >	m_SobelX
	Sobel operators.
CSobelY< tImage, CShiftZero2Gray >	m_SobelY

---

Detailed Description

template<class I>
class mds::img::CCanny< I >

Canny edge detector.

• Parameter I is a used image type.

---

Member Typedef Documentation

template<class I >

typedef CImageEdgeDetector<I> mds::img::CCanny< I >::base

Image edge detector base.

template<class I >

typedef base::tImage mds::img::CCanny< I >::tImage

Image type.

Reimplemented from mds::img::CImageEdgeDetector< I >.

template<class I >

typedef base::tPixel mds::img::CCanny< I >::tPixel

Image pixel type.

Reimplemented from mds::img::CImageEdgeDetector< I >.

---

Member Enumeration Documentation

template<class I >

anonymous enum [inherited]

Check that I is an image.

You will see name of this enum somewhere in compiler error message if the type I is not image.

Enumerator:

TEMPLATE_PARAMETER_IS_NOT_IMAGE

---

Constructor & Destructor Documentation

template<class I >

mds::img::CCanny< I >::CCanny	(	double	dSigma,
		double	dT1,
		double	dT2
	)

Default constructor.

• Standard deviation ("sigma" parameter) of the Gaussian filter which is used for the smoothing.
• Thresholds for the hysteresis (0..1).

template<class I >

mds::img::CCanny< I >::~CCanny

(

)

Destructor.

---

Member Function Documentation

template<class I >

bool CCanny::checkNeighbours	(	tImage &	Image,
		tSize	x,
		tSize	y,
		tPixel	T
	)		[protected]

Returns true if any neighbouring pixel has greater value than a given threshold.

template<class I >

double mds::img::CCanny< I >::getSigma

(

)

const

Returns smoothing filter standard deviation.

template<class I >

void CCanny::getThresholds	(	double &	dT1,
		double &	dT2
	)

Returns thresholds of the non-maximal suppression algorithm.

template<class I >

void CCanny::hysteresis	(	tImage &	MagnitudeImage,
		tImage &	Image,
		tPixel	T1,
		tPixel	T2
	)		[protected]

Hystersis controlled by two given thresholds.

• Applies the hysteresis to the provided slice. It checks each pixel against an upper threshhold 'T1'. If the pixel value is equal or higher, the current pixel is an edge pixel. Then all direct neighbours of the pixel are compared to a lower threshhold 'T2' and if their values are equal or higher than this threshhold, they become edge pixels as well.
• In practice the threshold T1 = (2-3) * T2.

template<class I >

void CCanny::nonMaxSuppression	(	const tImage &	GradImageX,
		const tImage &	GradImageY,
		tImage &	MagnitudeImage,
		tImage &	Image
	)		[protected]

Non-maximal suppression.

template<class I >

bool CCanny::operator()	(	const tImage &	SrcImage,
		tImage &	DstImage
	)		[virtual]

Edge detection in a given image.

• Magnitude image is produced.
• Returns false on failure.

Implements mds::img::CImageEdgeDetector< I >.

template<class I >

void mds::img::CCanny< I >::setSigma

(

double

dSigma

)

Sets the smoothing filter standard deviation.

template<class I >

void CCanny::setThresholds	(	double	dT1,
		double	dT2
	)

Sets thresholds of the non-maximal suppression algorithm.

---

Member Data Documentation

template<class I >

double mds::img::CCanny< I >::m_dT1 [protected]

Thresholds used by hysteresis algorithm.

template<class I >

double mds::img::CCanny< I >::m_dT2 [protected]

template<class I >

CGaussFilter<tImage> mds::img::CCanny< I >::m_GaussFilter [protected]

Gaussian smoothing filter.

template<class I >

CSobelX<tImage, CShiftZero2Gray> mds::img::CCanny< I >::m_SobelX [protected]

Sobel operators.

template<class I >

CSobelY<tImage, CShiftZero2Gray> mds::img::CCanny< I >::m_SobelY [protected]

---

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

• EdgeDetection/mdsCanny.h
• EdgeDetection/mdsCanny.hxx