Jafar
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends Defines
Classes | Public Types | Public Member Functions | Private Member Functions | Private Attributes
jafar::fdetect::HarrisDetector Class Reference

This class is the implements a detector of Harris points as described in : "A Combined Corner and Edge Detector" par C. More...


Detailed Description

This class is the implements a detector of Harris points as described in : "A Combined Corner and Edge Detector" par C.

Harris et M. Stephens en 1988.

Definition at line 22 of file HarrisDetector.hpp.

#include <HarrisDetector.hpp>

Inheritance diagram for jafar::fdetect::HarrisDetector:
Inheritance graph
[legend]

List of all members.

Classes

struct  HData

Public Types

enum  Cornerness { HS_HARRIS, HS_EIGEN }

Public Member Functions

 HarrisDetector (float scale_factor=1.0, float derivationSigma=1.0, float convolutionSigma=2.0, jafar::fdetect::HarrisDetector::Cornerness cornerness=HS_EIGEN, float threshold=0.0, bool fixnumber=true, int desirednum=1000)
virtual
jafar::fdetect::DetectionResult 
detectIn (jafar::image::Image const &image)
virtual bool detectBestPntInRoi (jafar::image::Image const &image, InterestFeature *featPtr, const jafar::image::ROI *roiPtr=0)
float scaleFactor () const
float derivationSigma () const
float convolutionSigma () const
float threshold () const
Cornerness cornerness () const
int desiredNumber () const
int derivationSize () const
int convolutionSize () const
bool isFixNumber () const

Private Member Functions

void gaussianDerivativesSA (jafar::image::Image const &image)
void gaussianDerivatives (jafar::image::Image const &image)
void gaussianDerivatives (jafar::image::Image const &image, const jafar::image::ROI &roi)
void gaussianConvolutionSA (jafar::image::Image const &image)
void gaussianConvolution (jafar::image::Image const &image)
void gaussianConvolution (jafar::image::Image const &image, const jafar::image::ROI &roi)
void computeHarrisCurvature (jafar::image::Image const &image)
void computeEigenCurvature (jafar::image::Image const &image)
void computeEigenCurvatureWithMax (jafar::image::Image const &image, int pixMax[2], float *scoreMax)
void cornerCharacteristics (DetectionResult &points, jafar::image::Image const &image)
template<bool autoThreshold, bool useScale>
DetectionResult extractCorners (jafar::image::Image const &image)

Private Attributes

float m_scale_factor
float m_derivationSigma
float m_convolutionSigma
float m_threshold
Cornerness m_cornerness
int m_desirednum
int m_derivationSize
int m_convolutionSize
bool m_fixnumber
HDatam_data

Constructor & Destructor Documentation

jafar::fdetect::HarrisDetector::HarrisDetector ( float  scale_factor = 1.0,
float  derivationSigma = 1.0,
float  convolutionSigma = 2.0,
jafar::fdetect::HarrisDetector::Cornerness  cornerness = HS_EIGEN,
float  threshold = 0.0,
bool  fixnumber = true,
int  desirednum = 1000 
)
Parameters:
scale_factorthe scale used to search the interest point, it usually set to 1.0, but it might prove usefull to have it set to a different value if you know that there is a scale difference between the two images you are trying to match.
derivationSigmathe sigma used by the derivation kernel
convolutionSigmathe sigma used by the convolution kernel
cornernessthe type of cornerness (either using eigen values or the "harris" cornerness)
thresholduse to select point that are correctly different from their neighbourhood
fixnumberset it to true if you want a given number of point which have the maximum cornerness (note that the threshold is still used to avoid selecting points which won't be good enought)
desirednumthe number of points (if set to -1 and fixnumber is true, an automatic number of points is used)

The documentation for this class was generated from the following file:
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends Defines

Generated on Wed Oct 15 2014 00:37:35 for Jafar by doxygen 1.7.6.1
LAAS-CNRS