jafar::lines::LsExtractor Class Reference

Class with functions to extract line segments of an image. More...

---

Detailed Description

Class with functions to extract line segments of an image.

This class provides a set of functions to extract line segments of an image. The idea is to extract all line segments contained in an image (global extraction). To re-extract a line from a prediction of that line (local extraction) have a look at LineSegment::fitLineOrientation() function and its Usage in the LsTracker class.

To simply extract the set of lines of an image, the extractLineSegments() function can be used. It combines the extraction steps with the currently best performance.

The functions of LsExtractor class works all with the same scheme: calculate gradient image or Canny image and find contour points by maxima suppression with hysteresis and find lines either with polygonalisation There is a combination of the two OpenCV function for that task, but the contour representation that result by OpenCv cvFindContours runs a lot of contour points for multiple times if there are branched contours, thus the contours are very long. Therefore the well tested Calife functions for that task can be used alternatively. Some documentation of the Calife functions is contained in ctdef.hpp file. The OpenCv polygonalization algorithm seems to be a little bit buggy (CVS version avaiable at January 2008), therefore a simple implementation of Douglas-Peucker Algorithm is provided.

Further there is a function to apply a Hough-transformation to find lines.

Definition at line 28 of file lsExtractor.hpp.

#include <lsExtractor.hpp>

List of all members.

Public Member Functions
bool	procCanny (jafar::image::Image *image, jafar::image::Image *cannyIm)
	Procced cvCanny on _image with member variables as Canny parameters.
void	extractLineSegments (jafar::image::Image *image, jafar::image::Image *cannyIm, LineSegmentSet &segmentStorage, double thresh=15)
	Extraction of line segments with memeber variables as parameters.
bool	procHough (jafar::image::Image *image, LineSegmentSet &newSegments)
	Procced cvHoughLines2 on image and gives resulting set of lines.
bool	procCvFindContour (jafar::image::Image *image, LineSegmentSet &newSegments, double thresh=15)
	Proceed cvFindContours and cvApproxPoly on image Therefore the image has be a binary image (e.g.
void	setCannyLowerThresh (int lowerThresh)
void	setCannyHigherThresh (int higherThresh)
void	setCannyAp (int ap)
void	setCannyPara (int lowerThresh, int higherThresh, int ap)
void	setHoughMethod (int method)
void	setHoughRho (double rho)
void	setHoughTheta (double theta)
void	setHoughThreshold (int thresh)
void	setHoughParam1 (double param)
void	setHoughParam2 (double param)
void	setHoughPara (int method, double rho, double theta, int thresh, int param1, int param2)
Static Public Member Functions
static bool	procCannySt (jafar::image::Image *image, jafar::image::Image *cannyIm, int lowerThresh, int higherThresh, int ap)
	Procced cvCanny on _image with function parameters as Canny parameters.
static void	extractLineSegmentsSt (jafar::image::Image *image, jafar::image::Image *cannyIm, LineSegmentSet &segmentStorage, double thresh=15, int lowerThresh=75, int higherThresh=150, int ap=3)
	Extraction of line segments with function parameters as parameters.
static void	findLinesDP (jafar::image::Image *image, LineSegmentSet &newSegments, double thresh)
	Input image should either be canny or gradient image, attention: an image with height or width which is not a multiple of 4 slows down the function (because the whole image data must be copied) The content of the image is changed.
static void	findLinesCalife (jafar::image::Image *image, LineSegmentSet &newSegments, double thresh=15)
	Input image should either be canny or gradient image, attention: an image with height or width which is not a multiple of 4 slows down the function (because the whole image data must be copied) The content of the image is changed.
static int	dp (short *x, short *y, int startIdx, int endIdx, std::vector< CvPoint > &endpoints, double thresh=1)
	Recursive function for polygonal approximation of a set of points.
Protected Attributes
int	cannyLowerThresh
	Parameter for cvCanny.
int	cannyHigherThresh
	Parameter for cvCanny.
int	cannyAp
	Parameter for cvCanny.
int	houghMethod
	Parameter for cvHoughLines2.
double	houghRho
	Parameter for cvHoughLines2.
double	houghTheta
	Parameter for cvHoughLines2.
int	houghThreshold
	Parameter for cvHoughLines2.
double	houghParam1
	Parameter for cvHoughLines2.
double	houghParam2
	Parameter for cvHoughLines2.

---

Member Function Documentation

static int jafar::lines::LsExtractor::dp	(	short *	x,
		short *	y,
		int	startIdx,
		int	endIdx,
		std::vector< CvPoint > &	endpoints,
		double	thresh = 1
	)		[static]

Recursive function for polygonal approximation of a set of points.

The points are given by x and y coordinates. The used algorithm is Douglas-Peucker. Designed for use with ExtractContours1 function.

void jafar::lines::LsExtractor::extractLineSegments	(	jafar::image::Image *	image,
		jafar::image::Image *	cannyIm,
		LineSegmentSet &	segmentStorage,
		double	thresh = 15
	)

Extraction of line segments with memeber variables as parameters.

Extracts lines from image and appends result to _segmentStorage. If only the resulting lines are desired to be contained in _segmentStorage, then _segmentStorage should be cleared before calling this function (LineSegmentSet::clear()). The _image is used as container for the Canny image, therefore it is changed within this function. As parameters for Canny the member variables are taken.

Parameters:

image	Greyvalue image whose lines are extracted, image content is changed
cannyIm	Storage for the resulting image of Canny edge detection
segmentStorage	Storage for resulting line segments
thresh	Minimum length of accepted lines.

static void jafar::lines::LsExtractor::extractLineSegmentsSt	(	jafar::image::Image *	image,
		jafar::image::Image *	cannyIm,
		LineSegmentSet &	segmentStorage,
		double	thresh = 15,
		int	lowerThresh = 75,
		int	higherThresh = 150,
		int	ap = 3
	)		[static]

Extraction of line segments with function parameters as parameters.

Extracts lines from image and appends result to _segmentStorage. If only the resulting lines are desired to be contained in _segmentStorage, then _segmentStorage should be cleared before calling this function (LineSegmentSet::clear()). The _image is used as container for the Canny image, therefore it is changed within this function. As parameters for Canny the function parameters are taken.

Parameters:

image	Greyvalue image whose lines should be extracted, image content is changed
cannyIm	Storage for the resulting image of Canny edeg detection
segmentStorage	Storage for resulting line segments
thresh	Minimum length of accepted lines.
lowerThresh	Lower threshold for Canny edge detection
higherThresh	Higher threshold for Canny edge detection
ap	Aperture size for Canny edge detection

static void jafar::lines::LsExtractor::findLinesCalife	(	jafar::image::Image *	image,
		LineSegmentSet &	newSegments,
		double	thresh = 15
	)		[static]

Input image should either be canny or gradient image, attention: an image with height or width which is not a multiple of 4 slows down the function (because the whole image data must be copied) The content of the image is changed.

All pixels, that are contained in a taken contour are set to 127.

static void jafar::lines::LsExtractor::findLinesDP	(	jafar::image::Image *	image,
		LineSegmentSet &	newSegments,
		double	thresh
	)		[static]

Input image should either be canny or gradient image, attention: an image with height or width which is not a multiple of 4 slows down the function (because the whole image data must be copied) The content of the image is changed.

All pixels, that are contained in a taken contour are set to 127. The polygonal approximation is done with db() function, witch implements Douglas-Peucker algorithm.

Parameters:

image	Greyvalue image whose lines are extracted, image content is changed
newSegments	Storage for resulting line segments
thresh	Minimum length of accepted lines.

bool jafar::lines::LsExtractor::procCanny	(	jafar::image::Image *	image,
		jafar::image::Image *	cannyIm
	)

Procced cvCanny on _image with member variables as Canny parameters.

Applies Canny edge detection on image and saves the result in cannyIm. The member variables are taken as parameters for Canny.

static bool jafar::lines::LsExtractor::procCannySt	(	jafar::image::Image *	image,
		jafar::image::Image *	cannyIm,
		int	lowerThresh,
		int	higherThresh,
		int	ap
	)		[static]

Procced cvCanny on _image with function parameters as Canny parameters.

Applies Canny edge detection on image and saves the result in cannyIm. The function parameters are taken as parameters for Canny

bool jafar::lines::LsExtractor::procCvFindContour	(	jafar::image::Image *	image,
		LineSegmentSet &	newSegments,
		double	thresh = 15
	)

Proceed cvFindContours and cvApproxPoly on image Therefore the image has be a binary image (e.g.

the result of cvCanny, have a look at the procCanny function) To accept only large lineSegments a threshold is used

bool jafar::lines::LsExtractor::procHough	(	jafar::image::Image *	image,
		LineSegmentSet &	newSegments
	)

Procced cvHoughLines2 on image and gives resulting set of lines.

Applies Hough on image and saves the result in image.

---

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

• modules/lines/include/lines/lsExtractor.hpp