Table Of Contents

Stereo Correspondence (SG Block Matching) (G Dataflow)

Last Modified: June 25, 2019

Computes the disparity map for a pair of stereo images using the semi-global block matching method.

This node implements the entire 3D reconstruction process internally until the disparity information is available. The node internally rectifies the left and right images and then use the rectified images to compute the disparity information.

connector_pane_image
datatype_icon

postfilter options

Disparity information obtained from stereo correspondence and how it is improved.

datatype_icon

texture threshold

Minimum sum of absolute value (SAD) at each pixel to be considered to be a valid disparity.

datatype_icon

uniqueness ratio

Input that specifies how unique the match at each pixel should be to considered as valid.

datatype_icon

speckle window size

Maximum allowed deviation in neighboring pixels disparity values.

The speckle window size defines the following interval: [current disparity - (speckle window size x scaling factor), current disparity + (speckle window size x scaling factor)], where the scaling factor equals 16. For example, if the current disparity is 128 and the speckle window size is 5, then the speckle window size defines an interval [48, 208].

datatype_icon

speckle range

Number of neighboring pixel disparities the speckle filter must locate before a given pixel disparity is considered valid.

A pixel disparity is valid if the value falls within the interval defined in speckle window size.

datatype_icon

filter cap

Input that specifies the filter cap.

datatype_icon

binocular stereo session in

Reference to the binocular stereo vision session on which this node operates.

datatype_icon

left image in

Reference to the input image from the left camera in the binocular stereo system.

datatype_icon

right image in

Reference to the input image from the right camera in the binocular stereo system.

datatype_icon

disparity image in

Reference to the image that will contain the disparity information after the node executes.

datatype_icon

score image in

Reference to the image that will contain the score information associated with the disparity image after the node executes.

datatype_icon

error in

Error conditions that occur before this node runs.

The node responds to this input according to standard error behavior.

Standard Error Behavior

Many nodes provide an error in input and an error out output so that the node can respond to and communicate errors that occur while code is running. The value of error in specifies whether an error occurred before the node runs. Most nodes respond to values of error in in a standard, predictable way.

error in does not contain an error error in contains an error
If no error occurred before the node runs, the node begins execution normally.

If no error occurs while the node runs, it returns no error. If an error does occur while the node runs, it returns that error information as error out.

If an error occurred before the node runs, the node does not execute. Instead, it returns the error in value as error out.

Default: No error

datatype_icon

correspondence options

Option that specifies how the block matching algorithm computes the disparity information using stereo correspondence.

datatype_icon

method

Metric used to compute matches between the left and right images.

Only the sum of absolute differences (SAD) is supported.

datatype_icon

window size

Size of the window that is used to compute the matching score at each pixel in the image.

datatype_icon

minimum disparity

Minimum distance between the same point in the left and right rectified images.

datatype_icon

number of disparities

Maximum distance between the same point in the left and right rectified images.

You can use the Get Maximum Disparity node to compute this value if you know the minimum depth of the object under inspection from the camera.

datatype_icon

P1

Cost of changing the disparity value by a unit value during the updating process based on its neighborhood values.

datatype_icon

P2

Amount of smoothing applied to the disparity image.

A high value smooths the disparity image more at the cost of accuracy.

datatype_icon

full dynamic programming

Boolean that specifies if the block matching algorithm runs through multiple iterations using dynamic programming.

datatype_icon

max left right disparity change

Maximum allowed difference in the disparity value of a pixel with the disparity values of pixels immediately to its left and right.

If the difference in disparity is greater than this value, then the pixel is set to have an invalid disparity.

datatype_icon

subpixel

Option that computes the disparities with subpixel accuracy when enabled.

Bilinear interpolation is used to compute disparities with subpixel accuracy.

datatype_icon

invalid disparity replace value

Pixel value assigned to invalid disparity values in the output disparity image.

Use the same value that is used with the Stereo Correspondence (Block Matching) node or the Stereo Correspondence (SG Block Matching) node.

datatype_icon

process band

Boolean that determines how to compute the disparity.

True minimum disparity and number of disparities values are ignored and the node processes the entire image to compute the disparity.
False The overlap area is replaced with the invalid pixel value.

Depending on the overlap between the left image and the right image, the interpolated disparities in the overlap area may not be correct.

Default: False

datatype_icon

binocular stereo session out

Reference to the stereo vision session on which this node operates.

datatype_icon

disparity image out

Reference to the output image that contains the computed disparity values.

datatype_icon

score image out

Reference to the output image that contains the scores associated with disparity values returned in disparity image out.

Each pixel in the image contains a value between 0 and 1000. A value of 1000 indicates full confidence in the computed disparity value at that pixel coordinates. A value of 0 indicates no confidence in the computed disparity value.

datatype_icon

error out

Error information.

The node produces this output according to standard error behavior.

Standard Error Behavior

Many nodes provide an error in input and an error out output so that the node can respond to and communicate errors that occur while code is running. The value of error in specifies whether an error occurred before the node runs. Most nodes respond to values of error in in a standard, predictable way.

error in does not contain an error error in contains an error
If no error occurred before the node runs, the node begins execution normally.

If no error occurs while the node runs, it returns no error. If an error does occur while the node runs, it returns that error information as error out.

If an error occurred before the node runs, the node does not execute. Instead, it returns the error in value as error out.

Where This Node Can Run:

Desktop OS: Windows

FPGA: Not supported

Web Server: Not supported in VIs that run in a web application


Recently Viewed Topics