cv::cuda::SURF_CUDA Class Reference

Class used for extracting Speeded Up Robust Features (SURF) from an image. : More...

#include <opencv2/xfeatures2d/cuda.hpp>

Collaboration diagram for cv::cuda::SURF_CUDA:

Public Types
enum	KeypointLayout {   X_ROW = 0 ,   Y_ROW ,   LAPLACIAN_ROW ,   OCTAVE_ROW ,   SIZE_ROW ,   ANGLE_ROW ,   HESSIAN_ROW ,   ROWS_COUNT }

Public Member Functions
	SURF_CUDA ()
	the default constructor
	SURF_CUDA (double _hessianThreshold, int _nOctaves=4, int _nOctaveLayers=2, bool _extended=false, float _keypointsRatio=0.01f, bool _upright=false)
	the full constructor taking all the necessary parameters
int	defaultNorm () const
	returns the default norm type
int	descriptorSize () const
	returns the descriptor size in float's (64 or 128)
void	detect (const GpuMat &img, const GpuMat &mask, GpuMat &keypoints)
	Finds the keypoints using fast hessian detector used in SURF.
void	detectWithDescriptors (const GpuMat &img, const GpuMat &mask, GpuMat &keypoints, GpuMat &descriptors, bool useProvidedKeypoints=false)
	Finds the keypoints and computes their descriptors using fast hessian detector used in SURF.
void	downloadDescriptors (const GpuMat &descriptorsGPU, std::vector< float > &descriptors)
	download descriptors from device to host memory
void	downloadKeypoints (const GpuMat &keypointsGPU, std::vector< KeyPoint > &keypoints)
	download keypoints from device to host memory
void	operator() (const GpuMat &img, const GpuMat &mask, GpuMat &keypoints)
void	operator() (const GpuMat &img, const GpuMat &mask, GpuMat &keypoints, GpuMat &descriptors, bool useProvidedKeypoints=false)
void	operator() (const GpuMat &img, const GpuMat &mask, std::vector< KeyPoint > &keypoints)
void	operator() (const GpuMat &img, const GpuMat &mask, std::vector< KeyPoint > &keypoints, GpuMat &descriptors, bool useProvidedKeypoints=false)
void	operator() (const GpuMat &img, const GpuMat &mask, std::vector< KeyPoint > &keypoints, std::vector< float > &descriptors, bool useProvidedKeypoints=false)
void	releaseMemory ()
void	uploadKeypoints (const std::vector< KeyPoint > &keypoints, GpuMat &keypointsGPU)
	upload host keypoints to device memory

Static Public Member Functions
static Ptr< SURF_CUDA >	create (double _hessianThreshold, int _nOctaves=4, int _nOctaveLayers=2, bool _extended=false, float _keypointsRatio=0.01f, bool _upright=false)

Public Attributes
GpuMat	det
bool	extended
double	hessianThreshold
float	keypointsRatio
	max keypoints = min(keypointsRatio * img.size().area(), 65535)
GpuMat	mask1
GpuMat	maskSum
GpuMat	maxPosBuffer
int	nOctaveLayers
int	nOctaves
GpuMat	sum
GpuMat	trace
bool	upright

Detailed Description

Class used for extracting Speeded Up Robust Features (SURF) from an image. :

The class SURF_CUDA implements Speeded Up Robust Features descriptor. There is a fast multi-scale Hessian keypoint detector that can be used to find the keypoints (which is the default option). But the descriptors can also be computed for the user-specified keypoints. Only 8-bit grayscale images are supported.

The class SURF_CUDA can store results in the GPU and CPU memory. It provides functions to convert results between CPU and GPU version ( uploadKeypoints, downloadKeypoints, downloadDescriptors ). The format of CPU results is the same as SURF results. GPU results are stored in GpuMat. The keypoints matrix is \(\texttt{nFeatures} \times 7\) matrix with the CV_32FC1 type.

• keypoints.ptr<float>(X_ROW)[i] contains x coordinate of the i-th feature.
• keypoints.ptr<float>(Y_ROW)[i] contains y coordinate of the i-th feature.
• keypoints.ptr<float>(LAPLACIAN_ROW)[i] contains the laplacian sign of the i-th feature.
• keypoints.ptr<float>(OCTAVE_ROW)[i] contains the octave of the i-th feature.
• keypoints.ptr<float>(SIZE_ROW)[i] contains the size of the i-th feature.
• keypoints.ptr<float>(ANGLE_ROW)[i] contain orientation of the i-th feature.
• keypoints.ptr<float>(HESSIAN_ROW)[i] contains the response of the i-th feature.

The descriptors matrix is \(\texttt{nFeatures} \times \texttt{descriptorSize}\) matrix with the CV_32FC1 type.

The class SURF_CUDA uses some buffers and provides access to it. All buffers can be safely released between function calls.

See also

SURF

Note

• An example for using the SURF keypoint matcher on GPU can be found at opencv_source_code/samples/gpu/surf_keypoint_matcher.cpp

Member Enumeration Documentation

KeypointLayout

enum cv::cuda::SURF_CUDA::KeypointLayout

Enumerator
X_ROW
Y_ROW
LAPLACIAN_ROW
OCTAVE_ROW
SIZE_ROW
ANGLE_ROW
HESSIAN_ROW
ROWS_COUNT

Constructor & Destructor Documentation

SURF_CUDA() [1/2]

cv::cuda::SURF_CUDA::SURF_CUDA

(

)

the default constructor

SURF_CUDA() [2/2]

cv::cuda::SURF_CUDA::SURF_CUDA ( double  _hessianThreshold, int  _nOctaves = 4, int  _nOctaveLayers = 2, bool  _extended = false, float  _keypointsRatio = 0.01f, bool  _upright = false  )

explicit

the full constructor taking all the necessary parameters

Member Function Documentation

create()

static Ptr< SURF_CUDA > cv::cuda::SURF_CUDA::create ( double  _hessianThreshold, int  _nOctaves = 4, int  _nOctaveLayers = 2, bool  _extended = false, float  _keypointsRatio = 0.01f, bool  _upright = false  )

static

Python:
	cv.cuda.SURF_CUDA.create(	_hessianThreshold[, _nOctaves[, _nOctaveLayers[, _extended[, _keypointsRatio[, _upright]]]]]	) ->	retval
	cv.cuda.SURF_CUDA_create(	_hessianThreshold[, _nOctaves[, _nOctaveLayers[, _extended[, _keypointsRatio[, _upright]]]]]	) ->	retval

Parameters

_hessianThreshold	Threshold for hessian keypoint detector used in SURF.
_nOctaves	Number of pyramid octaves the keypoint detector will use.
_nOctaveLayers	Number of octave layers within each octave.
_extended	Extended descriptor flag (true - use extended 128-element descriptors; false - use 64-element descriptors).
_keypointsRatio	Limits a maximum number of features
_upright	Up-right or rotated features flag (true - do not compute orientation of features; false - compute orientation).

defaultNorm()

int cv::cuda::SURF_CUDA::defaultNorm

(

)

const

Python:
	cv.cuda.SURF_CUDA.defaultNorm(		) ->	retval

returns the default norm type

descriptorSize()

int cv::cuda::SURF_CUDA::descriptorSize

(

)

const

Python:
	cv.cuda.SURF_CUDA.descriptorSize(		) ->	retval

returns the descriptor size in float's (64 or 128)

detect()

void cv::cuda::SURF_CUDA::detect ( const GpuMat &  img, const GpuMat &  mask, GpuMat &  keypoints  )

inline

Python:
	cv.cuda.SURF_CUDA.detect(	img, mask[, keypoints]	) ->	keypoints

Finds the keypoints using fast hessian detector used in SURF.

Parameters

img	Source image, currently supports only CV_8UC1 images.
mask	A mask image same size as src and of type CV_8UC1.
keypoints	Detected keypoints.

detectWithDescriptors()

void cv::cuda::SURF_CUDA::detectWithDescriptors ( const GpuMat &  img, const GpuMat &  mask, GpuMat &  keypoints, GpuMat &  descriptors, bool  useProvidedKeypoints = false  )

inline

Python:
	cv.cuda.SURF_CUDA.detectWithDescriptors(	img, mask[, keypoints[, descriptors[, useProvidedKeypoints]]]	) ->	keypoints, descriptors

Finds the keypoints and computes their descriptors using fast hessian detector used in SURF.

Parameters

img	Source image, currently supports only CV_8UC1 images.
mask	A mask image same size as src and of type CV_8UC1.
keypoints	Detected keypoints.
descriptors	Keypoint descriptors.
useProvidedKeypoints	Compute descriptors for the user-provided keypoints and recompute keypoints direction.

downloadDescriptors()

void cv::cuda::SURF_CUDA::downloadDescriptors	(	const GpuMat &	descriptorsGPU,
		std::vector< float > &	descriptors
	)

download descriptors from device to host memory

downloadKeypoints()

void cv::cuda::SURF_CUDA::downloadKeypoints	(	const GpuMat &	keypointsGPU,
		std::vector< KeyPoint > &	keypoints
	)

Python:
	cv.cuda.SURF_CUDA.downloadKeypoints(	keypointsGPU	) ->	keypoints

download keypoints from device to host memory

operator()() [1/5]

void cv::cuda::SURF_CUDA::operator()	(	const GpuMat &	img,
		const GpuMat &	mask,
		GpuMat &	keypoints
	)

finds the keypoints using fast hessian detector used in SURF supports CV_8UC1 images keypoints will have nFeature cols and 6 rows keypoints.ptr<float>(X_ROW)[i] will contain x coordinate of i'th feature keypoints.ptr<float>(Y_ROW)[i] will contain y coordinate of i'th feature keypoints.ptr<float>(LAPLACIAN_ROW)[i] will contain laplacian sign of i'th feature keypoints.ptr<float>(OCTAVE_ROW)[i] will contain octave of i'th feature keypoints.ptr<float>(SIZE_ROW)[i] will contain size of i'th feature keypoints.ptr<float>(ANGLE_ROW)[i] will contain orientation of i'th feature keypoints.ptr<float>(HESSIAN_ROW)[i] will contain response of i'th feature

operator()() [2/5]

void cv::cuda::SURF_CUDA::operator()	(	const GpuMat &	img,
		const GpuMat &	mask,
		GpuMat &	keypoints,
		GpuMat &	descriptors,
		bool	useProvidedKeypoints = false
	)

finds the keypoints and computes their descriptors. Optionally it can compute descriptors for the user-provided keypoints and recompute keypoints direction

operator()() [3/5]

void cv::cuda::SURF_CUDA::operator()	(	const GpuMat &	img,
		const GpuMat &	mask,
		std::vector< KeyPoint > &	keypoints
	)

operator()() [4/5]

void cv::cuda::SURF_CUDA::operator()	(	const GpuMat &	img,
		const GpuMat &	mask,
		std::vector< KeyPoint > &	keypoints,
		GpuMat &	descriptors,
		bool	useProvidedKeypoints = false
	)

operator()() [5/5]

void cv::cuda::SURF_CUDA::operator()	(	const GpuMat &	img,
		const GpuMat &	mask,
		std::vector< KeyPoint > &	keypoints,
		std::vector< float > &	descriptors,
		bool	useProvidedKeypoints = false
	)

releaseMemory()

void cv::cuda::SURF_CUDA::releaseMemory

(

)

uploadKeypoints()

void cv::cuda::SURF_CUDA::uploadKeypoints	(	const std::vector< KeyPoint > &	keypoints,
		GpuMat &	keypointsGPU
	)

upload host keypoints to device memory

Member Data Documentation

det

GpuMat cv::cuda::SURF_CUDA::det

extended

bool cv::cuda::SURF_CUDA::extended

hessianThreshold

double cv::cuda::SURF_CUDA::hessianThreshold

keypointsRatio

float cv::cuda::SURF_CUDA::keypointsRatio

max keypoints = min(keypointsRatio * img.size().area(), 65535)

mask1

GpuMat cv::cuda::SURF_CUDA::mask1

maskSum

GpuMat cv::cuda::SURF_CUDA::maskSum

maxPosBuffer

GpuMat cv::cuda::SURF_CUDA::maxPosBuffer

nOctaveLayers

int cv::cuda::SURF_CUDA::nOctaveLayers

nOctaves

int cv::cuda::SURF_CUDA::nOctaves

sum

GpuMat cv::cuda::SURF_CUDA::sum

trace

GpuMat cv::cuda::SURF_CUDA::trace

upright

bool cv::cuda::SURF_CUDA::upright

---

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

• opencv2/xfeatures2d/cuda.hpp