Stereo Correspondence

Detailed Description

Classes
class	cv::cuda::DisparityBilateralFilter
	Class refining a disparity map using joint bilateral filtering. : More...
class	cv::cuda::StereoBeliefPropagation
	Class computing stereo correspondence using the belief propagation algorithm. : More...
class	cv::cuda::StereoBM
	Class computing stereo correspondence (disparity map) using the block matching algorithm. : More...
class	cv::cuda::StereoConstantSpaceBP
	Class computing stereo correspondence using the constant space belief propagation algorithm. : More...
class	cv::cuda::StereoSGM
	The class implements the modified H. Hirschmuller algorithm [126]. Limitation and difference are as follows: More...

Functions
Ptr< cuda::DisparityBilateralFilter >	cv::cuda::createDisparityBilateralFilter (int ndisp=64, int radius=3, int iters=1)
	Creates DisparityBilateralFilter object.
Ptr< cuda::StereoBeliefPropagation >	cv::cuda::createStereoBeliefPropagation (int ndisp=64, int iters=5, int levels=5, int msg_type=CV_32F)
	Creates StereoBeliefPropagation object.
Ptr< cuda::StereoBM >	cv::cuda::createStereoBM (int numDisparities=64, int blockSize=19)
	Creates StereoBM object.
Ptr< cuda::StereoConstantSpaceBP >	cv::cuda::createStereoConstantSpaceBP (int ndisp=128, int iters=8, int levels=4, int nr_plane=4, int msg_type=CV_32F)
	Creates StereoConstantSpaceBP object.
Ptr< cuda::StereoSGM >	cv::cuda::createStereoSGM (int minDisparity=0, int numDisparities=128, int P1=10, int P2=120, int uniquenessRatio=5, int mode=cv::cuda::StereoSGM::MODE_HH4)
	Creates StereoSGM object.
void	cv::cuda::drawColorDisp (InputArray src_disp, OutputArray dst_disp, int ndisp, Stream &stream=Stream::Null())
	Colors a disparity image.
void	cv::cuda::reprojectImageTo3D (GpuMat disp, GpuMat &xyzw, Mat Q, int dst_cn=4, Stream &stream=Stream::Null())
void	cv::cuda::reprojectImageTo3D (InputArray disp, OutputArray xyzw, InputArray Q, int dst_cn=4, Stream &stream=Stream::Null())
	Reprojects a disparity image to 3D space.

Function Documentation

createDisparityBilateralFilter()

Ptr< cuda::DisparityBilateralFilter > cv::cuda::createDisparityBilateralFilter	(	int	ndisp = 64,
		int	radius = 3,
		int	iters = 1
	)

#include <opencv2/cudastereo.hpp>

Creates DisparityBilateralFilter object.

Parameters

ndisp	Number of disparities.
radius	Filter radius.
iters	Number of iterations.

createStereoBeliefPropagation()

Ptr< cuda::StereoBeliefPropagation > cv::cuda::createStereoBeliefPropagation	(	int	ndisp = 64,
		int	iters = 5,
		int	levels = 5,
		int	msg_type = CV_32F
	)

#include <opencv2/cudastereo.hpp>

Creates StereoBeliefPropagation object.

Parameters

ndisp	Number of disparities.
iters	Number of BP iterations on each level.
levels	Number of levels.
msg_type	Type for messages. CV_16SC1 and CV_32FC1 types are supported.

createStereoBM()

Ptr< cuda::StereoBM > cv::cuda::createStereoBM	(	int	numDisparities = 64,
		int	blockSize = 19
	)

#include <opencv2/cudastereo.hpp>

Creates StereoBM object.

Parameters

numDisparities	the disparity search range. For each pixel algorithm will find the best disparity from 0 (default minimum disparity) to numDisparities. The search range can then be shifted by changing the minimum disparity.
blockSize	the linear size of the blocks compared by the algorithm. The size should be odd (as the block is centered at the current pixel). Larger block size implies smoother, though less accurate disparity map. Smaller block size gives more detailed disparity map, but there is higher chance for algorithm to find a wrong correspondence.

createStereoConstantSpaceBP()

Ptr< cuda::StereoConstantSpaceBP > cv::cuda::createStereoConstantSpaceBP	(	int	ndisp = 128,
		int	iters = 8,
		int	levels = 4,
		int	nr_plane = 4,
		int	msg_type = CV_32F
	)

#include <opencv2/cudastereo.hpp>

Creates StereoConstantSpaceBP object.

Parameters

ndisp	Number of disparities.
iters	Number of BP iterations on each level.
levels	Number of levels.
nr_plane	Number of disparity levels on the first level.
msg_type	Type for messages. CV_16SC1 and CV_32FC1 types are supported.

createStereoSGM()

Ptr< cuda::StereoSGM > cv::cuda::createStereoSGM	(	int	minDisparity = 0,
		int	numDisparities = 128,
		int	P1 = 10,
		int	P2 = 120,
		int	uniquenessRatio = 5,
		int	mode = cv::cuda::StereoSGM::MODE_HH4
	)

#include <opencv2/cudastereo.hpp>

Creates StereoSGM object.

Parameters

minDisparity	Minimum possible disparity value. Normally, it is zero but sometimes rectification algorithms can shift images, so this parameter needs to be adjusted accordingly.
numDisparities	Maximum disparity minus minimum disparity. The value must be 64, 128 or 256.
P1	The first parameter controlling the disparity smoothness.This parameter is used for the case of slanted surfaces (not fronto parallel).
P2	The second parameter controlling the disparity smoothness.This parameter is used for "solving" the depth discontinuities problem.
uniquenessRatio	Margin in percentage by which the best (minimum) computed cost function value should "win" the second best value to consider the found match correct. Normally, a value within the 5-15 range is good enough.
mode	Set it to StereoSGM::MODE_HH to run the full-scale two-pass dynamic programming algorithm. It will consume O(W*H*numDisparities) bytes. By default, it is set to StereoSGM::MODE_HH4.

drawColorDisp()

void cv::cuda::drawColorDisp	(	InputArray	src_disp,
		OutputArray	dst_disp,
		int	ndisp,
		Stream &	stream = Stream::Null()
	)

#include <opencv2/cudastereo.hpp>

Colors a disparity image.

Parameters

src_disp	Input single-channel 8-bit unsigned, 16-bit signed, 32-bit signed or 32-bit floating-point disparity image. If 16-bit signed format is used, the values are assumed to have no fractional bits.
dst_disp	Output disparity image. It has the same size as src_disp. The type is CV_8UC4 in BGRA format (alpha = 255).
ndisp	Number of disparities.
stream	Stream for the asynchronous version.

This function draws a colored disparity map by converting disparity values from [0..ndisp) interval first to HSV color space (where different disparity values correspond to different hues) and then converting the pixels to RGB for visualization.

reprojectImageTo3D() [1/2]

void cv::cuda::reprojectImageTo3D ( GpuMat  disp, GpuMat &  xyzw, Mat  Q, int  dst_cn = 4, Stream &  stream = Stream::Null()  )

inline

#include <opencv2/cudastereo.hpp>

Here is the call graph for this function:

reprojectImageTo3D() [2/2]

void cv::cuda::reprojectImageTo3D	(	InputArray	disp,
		OutputArray	xyzw,
		InputArray	Q,
		int	dst_cn = 4,
		Stream &	stream = Stream::Null()
	)

#include <opencv2/cudastereo.hpp>

Reprojects a disparity image to 3D space.

Parameters

disp	Input single-channel 8-bit unsigned, 16-bit signed, 32-bit signed or 32-bit floating-point disparity image. If 16-bit signed format is used, the values are assumed to have no fractional bits.
xyzw	Output 3- or 4-channel floating-point image of the same size as disp . Each element of xyzw(x,y) contains 3D coordinates (x,y,z) or (x,y,z,1) of the point (x,y) , computed from the disparity map.
Q	\(4 \times 4\) perspective transformation matrix that can be obtained via stereoRectify .
dst_cn	The number of channels for output image. Can be 3 or 4.
stream	Stream for the asynchronous version.

See also

reprojectImageTo3D