Functions
cvbres_t	OptFlowCreateColorfulOrientationPlot (IMG ImgV, float MaxVectorLength, cvbval_t ScaleSize, IMG &COPImage)
	Creates colorful orientation plot. More...

cvbres_t	OptFlowLKCalculate (OPTFLOWLK OpticalFlowHandle, IMG Img1, cvbdim_t Img1Index, IMG Img2, cvbdim_t Img2Index, PIXELLIST &result)
	Calculates optical flow for an image pair. More...

cvbres_t	OptFlowLKCalculateSeq (OPTFLOWLK OpticalFlowHandle, IMG Img, cvbdim_t ImgIndex, PIXELLIST &result)
	Calculates optical flow. More...

OPTFLOWLK	OptFlowLKCreate (cvbdim_t Width, cvbdim_t Height)
	Creates a handle for calculating the optical flow based on the pyramidal Lucas-Kanade algorithm. More...

cvbres_t	OptFlowLKGetDistanceThreshold (OPTFLOWLK OpticalFlowHandle, float &Threshold)
	Gets the distance threshold for maximal movement vector length. More...

cvbres_t	OptFlowLKGetFeatureThreshold (OPTFLOWLK OpticalFlowHandle, float &Threshold)
	Gets minimal Eigen value for a point being considered as a feature. More...

cvbres_t	OptFlowLKGetNumGaussPyramidLevels (OPTFLOWLK OpticalFlowHandle, cvbval_t &PyrLevels)
	Gets the number of levels for the Gaussian pyramid. More...

cvbres_t	OptFlowLKGetNumIterationsPerPyramidLevel (OPTFLOWLK OpticalFlowHandle, cvbval_t &IterationsPerLevel)
	Gets the maximal number of iterations to be calculated on every level of the Gaussian-pyramid. More...

cvbres_t	OptFlowLKGetWindowSize (OPTFLOWLK OpticalFlowHandle, cvbdim_t &WindowSize)
	Gets the search window size for tracking moved feature points. More...

cvbres_t	OptFlowLKRecalculateFeaturePoints (OPTFLOWLK OpticalFlowHandle)
	Recalculates the feature points that are used for the optical flow calculation. More...

cvbres_t	OptFlowLKSetDistanceThreshold (OPTFLOWLK OpticalFlowHandle, float Threshold)
	Sets the distance threshold for the optical flow calculation. More...

cvbres_t	OptFlowLKSetFeatureSeparation (OPTFLOWLK OpticalFlowHandle, cvbval_t Separation)
	Sets the feature separation in pixels. More...

cvbres_t	OptFlowLKSetFeatureThreshold (OPTFLOWLK OpticalFlowHandle, float Threshold)
	Sets minimal Eigen value for a point being considered as a feature. More...

cvbres_t	OptFlowLKSetNumGaussPyramidLevels (OPTFLOWLK OpticalFlowHandle, cvbval_t PyrLevels)
	Sets the number Gaussian-pyramid levels to use. More...

cvbres_t	OptFlowLKSetNumIterationsPerPyramidLevel (OPTFLOWLK OpticalFlowHandle, cvbval_t IterationsPerLevel)
	Sets the maximal number of iterations to be calculated on every level of the Gaussian-pyramid. More...

cvbres_t	OptFlowLKSetWindowSize (OPTFLOWLK OpticalFlowHandle, cvbdim_t WindowSize)
	Sets the search window size for tracking moved feature points. More...

Detailed Description

Function Documentation

◆ OptFlowCreateColorfulOrientationPlot()

cvbres_t OptFlowCreateColorfulOrientationPlot	(	IMG	ImgV,
		float	MaxVectorLength,
		cvbval_t	ScaleSize,
		IMG &	COPImage
	)

Creates colorful orientation plot.

Calculates a colorful orientation plot to display the vector-fields of optical flow calculation. The angle of each vector is set to the hue-channel in HSV-color-model and the magnitude is set to value-channel. Afterwards the result-image is converted to the RGB color model.

Parameters

[in]	ImgV	The handle velocity vector image.
[in]	MaxVectorLength	Maximal vector length. Crops vector length with this value (> 0).
[in]	ScaleSize	The width of the color scale around the image in pixels. Must be >= 0.
[out]	COPImage	Handle to receive to the image that contains the colorful orientation plot of the vector field. Remember that this image will need to be released with ReleaseObject() once it is no longer needed.

Returns: A value >= 0 indicates success. All other return values are taken from the definitions in CVCError.h

◆ OptFlowLKCalculate()

cvbres_t OptFlowLKCalculate	(	OPTFLOWLK	OpticalFlowHandle,
		IMG	Img1,
		cvbdim_t	Img1Index,
		IMG	Img2,
		cvbdim_t	Img2Index,
		PIXELLIST &	result
	)

Calculates optical flow for an image pair.

This function calculates the optical flow between two images Img1 and Img2. It starts a new sequence with Img1, thus new feature points are calculated on Img1.

Parameters

[in]	OpticalFlowHandle	Handle of the optical flow state object.
[in]	Img1	The first image. The image plane must consist of integer or floating point pixels.
[in]	Img1Index	Plane index of ImgIn to process.
[in]	Img2	The second image (e.g., the one following Img1) to which the optical flow is calculated. The same restrictions as for ImgIn apply.
[in]	Img2Index	Plane index of ImgRef to process.
[out]	result	Four-dimensional PIXELLIST containing first the key-point from Img1 and then the motion vector to the matched key-point in Img2. Release the returned list via ReleaseObject when not needed anymore.

Returns: A value >= 0 indicates success. All other return values are taken from the definitions in CVCError.h

◆ OptFlowLKCalculateSeq()

cvbres_t OptFlowLKCalculateSeq	(	OPTFLOWLK	OpticalFlowHandle,
		IMG	Img,
		cvbdim_t	ImgIndex,
		PIXELLIST &	result
	)

Calculates optical flow.

This function determines the optical flow between the last processed image and a new Img.

No recalculation of the feature points is done, but the old optical flow is used to estimate the new locations.

After each calculation it may happen that some feature points are not found anymore and thus are removed. Use OptFlowLKRecalculateFeaturePoints to do a full recalculation to get a fresh list of feature points.

Parameters

[in]	OpticalFlowHandle	Handle of the optical flow state object.
[in]	Img	The current image. The image plane must consist of integer or floating point pixels.
[in]	ImgIndex	Plane index of ImgIn to process.
[out]	result	Four-dimensional PIXELLIST containing first the key-point from the last processed image and then the motion vector to the matched key-point in Img. Release the returned list via ReleaseObject when not needed anymore.

Returns: A value >= 0 indicates success. All other return values are taken from the definitions in CVCError.h

◆ OptFlowLKCreate()

OPTFLOWLK OptFlowLKCreate	(	cvbdim_t	Width,
		cvbdim_t	Height
	)

Creates a handle for calculating the optical flow based on the pyramidal Lucas-Kanade algorithm.

Creates an optical flow state object that can afterwards be used with the functions of the optical flow module of the Foundation Package (all functions working on the OPTFLOWLK type). Note that the object created by this function must be disposed by calling ReleaseObject from the CVCImg.dll on it.

Attention: After creating the object two new images need to be set until a result is available!

Parameters

[in]	Width	Width of the image to process in pixels.
[in]	Height	Height of the image to process in pixels.

Returns: Returns the reference to the new handle; NULL on error.

◆ OptFlowLKGetDistanceThreshold()

cvbres_t OptFlowLKGetDistanceThreshold	(	OPTFLOWLK	OpticalFlowHandle,
		float &	Threshold
	)

Gets the distance threshold for maximal movement vector length.

Parameters

[in]	OpticalFlowHandle	Handle of the optical flow state object.
[out]	Threshold	Threshold for maximal movement vector length per level.

Returns: A value >= 0 indicates success. All other return values are taken from the definitions in CVCError.h

◆ OptFlowLKGetFeatureThreshold()

cvbres_t OptFlowLKGetFeatureThreshold	(	OPTFLOWLK	OpticalFlowHandle,
		float &	Threshold
	)

Gets minimal Eigen value for a point being considered as a feature.

Parameters

[in]	OpticalFlowHandle	Handle of the optical flow state object.
[out]	Threshold	Minimal Eigen value for a feature point.

Returns: A value >= 0 indicates success. All other return values are taken from the definitions in CVCError.h

◆ OptFlowLKGetNumGaussPyramidLevels()

cvbres_t OptFlowLKGetNumGaussPyramidLevels	(	OPTFLOWLK	OpticalFlowHandle,
		cvbval_t &	PyrLevels
	)

Gets the number of levels for the Gaussian pyramid.

For detecting large optical-flow vectors it is necessary to use a larger number of Gaussian pyramid levels. Every additional level doubles the maximum magnitude that can be detected. Without additional levels, only motion vectors up to about 1 to 2 pixels can be detected (depending on the window size used for polynomial expansion).

Parameters

[in]	OpticalFlowHandle	The handle of the optical flow state object.
[out]	PyrLevels	Number of scale-levels. 0 means no additional level.

Returns: A value >= 0 indicates success. All other return values are taken from the definitions in CVCError.h

◆ OptFlowLKGetNumIterationsPerPyramidLevel()

cvbres_t OptFlowLKGetNumIterationsPerPyramidLevel	(	OPTFLOWLK	OpticalFlowHandle,
		cvbval_t &	IterationsPerLevel
	)

Gets the maximal number of iterations to be calculated on every level of the Gaussian-pyramid.

The estimation of optical-flow-vectors can be improved by repeating the calculation fore some iterations.

A higher iteration count may increase processing time during optical flow calculation.

Parameters

[in]	OpticalFlowHandle	Handle of the optical flow state object.
[out]	IterationsPerLevel	Maximal number of iterations per pyramid-level.

Returns: A value >= 0 indicates success. All other return values are taken from the definitions in CVCError.h

◆ OptFlowLKGetWindowSize()

cvbres_t OptFlowLKGetWindowSize	(	OPTFLOWLK	OpticalFlowHandle,
		cvbdim_t &	WindowSize
	)

Gets the search window size for tracking moved feature points.

Parameters

[in]	OpticalFlowHandle	Handle of the optical flow state object.
[out]	WindowSize	Search window size.

Returns: A value >= 0 indicates success. All other return values are taken from the definitions in CVCError.h

◆ OptFlowLKRecalculateFeaturePoints()

cvbres_t OptFlowLKRecalculateFeaturePoints ( OPTFLOWLK OpticalFlowHandle )

Recalculates the feature points that are used for the optical flow calculation.

New feature points are always calculated at the start of a new sequence (e.g. if OptFlowLKCalculate is called or one of the settings is changed). When OptFlowLKCalculateSeq is called no recalculation of the feature points is done, but the old optical flow is used to estimate the new locations.

After each calculation it may happen that some feature points are not found anymore and thus are removed. Use this function to do a full recalculation to get a fresh list of feature points.

Parameters

[in] OpticalFlowHandle Handle of the optical flow state object.

Returns: A value >= 0 indicates success. All other return values are taken from the definitions in CVCError.h

See also: OptFlowLKSetFeatureThreshold, OptFlowLKSetFeatureSeparation

◆ OptFlowLKSetDistanceThreshold()

cvbres_t OptFlowLKSetDistanceThreshold	(	OPTFLOWLK	OpticalFlowHandle,
		float	Threshold
	)

Sets the distance threshold for the optical flow calculation.

Attention: When setting new values for these two parameters the optical flow state is reset (two new images need to be set until a result is available)!

Parameters

[in]	OpticalFlowHandle	Handle of the optical flow state object.
[in]	Threshold	Threshold for maximal distance between old and new feature point per level.

Returns: A value >= 0 indicates success. All other return values are taken from the definitions in CVCError.h

◆ OptFlowLKSetFeatureSeparation()

cvbres_t OptFlowLKSetFeatureSeparation	(	OPTFLOWLK	OpticalFlowHandle,
		cvbval_t	Separation
	)

Sets the feature separation in pixels.

The feature separation defines the minimal distance between feature points used for the optical flow. Locations with corners normally have a cluster of high Eigen values. To reduce the number of key points and thus to speed up processing use this value.

This setting takes effect only when a new optical flow sequence is started or OptFlowLKRecalculateFeaturePoints is called.

Attention: When setting new values for these two parameters the optical flow state is reset (two new images need to be set until a result is available)!

Parameters

[in]	OpticalFlowHandle	Handle of the optical flow state object.
[in]	Separation	The feature separation in pixels; must be larger or equal to 0. Uses the cheaper L1 distance.

Returns: A value >= 0 indicates success. All other return values are taken from the definitions in CVCError.h

◆ OptFlowLKSetFeatureThreshold()

cvbres_t OptFlowLKSetFeatureThreshold	(	OPTFLOWLK	OpticalFlowHandle,
		float	Threshold
	)

Sets minimal Eigen value for a point being considered as a feature.

High Eigen values indicate corner points in the image where the optical flow can be measured well.

This setting takes effect only when a new optical flow sequence is started or OptFlowLKRecalculateFeaturePoints is called.

Attention: When setting new values for these two parameters the optical flow state is reset (two new images need to be set until a result is available)!

Parameters

[in]	OpticalFlowHandle	Handle of the optical flow state object.
[in]	Threshold	Minimal Eigen value for a feature point (>= 0).

Returns: A value >= 0 indicates success. All other return values are taken from the definitions in CVCError.h

◆ OptFlowLKSetNumGaussPyramidLevels()

cvbres_t OptFlowLKSetNumGaussPyramidLevels	(	OPTFLOWLK	OpticalFlowHandle,
		cvbval_t	PyrLevels
	)

Sets the number Gaussian-pyramid levels to use.

For detecting large optical-flow vectors it is necessary to use a larger number of Gaussian pyramid levels. Every additional level doubles the maximum magnitude that can be detected. Without additional levels, only motion vectors up to about 1 to 2 pixels can be detected (depending on the window size used for polynomial expansion.

Attention: The more pyramid-levels you use, the bigger the influence of the error-prone border region will become, i.e. the smaller the region with reliable results will become.; Additionally, errors may result if the objects that are moving are small and a high number pyramid levels are being used (objects may be averaged out in the process).; When setting new values for this property the optical flow state is reset (two new images need to be set until a result is available)!

Parameters

[in]	OpticalFlowHandle	The handle of the optical flow state object.
[in]	PyrLevels	Number of scale-levels. 0 means no additional level. Maximum is limited by the kernel and image size.

Returns: A value >= 0 indicates success. All other return values are taken from the definitions in CVCError.h

◆ OptFlowLKSetNumIterationsPerPyramidLevel()

cvbres_t OptFlowLKSetNumIterationsPerPyramidLevel	(	OPTFLOWLK	OpticalFlowHandle,
		cvbval_t	IterationsPerLevel
	)

Sets the maximal number of iterations to be calculated on every level of the Gaussian-pyramid.

The estimation of optical-flow-vectors can be improved by repeating the calculation. Typically, one to three iterations are sufficient.

A higher iteration count will increase processing time during optical flow calculation.

Attention: When setting new values for these two parameters the optical flow state is reset (two new images need to be set until a result is available)!

Parameters

[in]	OpticalFlowHandle	Handle of the optical flow state object.
[in]	IterationsPerLevel	Number of iterations per pyramid-level. Must be >= 1.

Returns: A value >= 0 indicates success. All other return values are taken from the definitions in CVCError.h

◆ OptFlowLKSetWindowSize()

cvbres_t OptFlowLKSetWindowSize	(	OPTFLOWLK	OpticalFlowHandle,
		cvbdim_t	WindowSize
	)

Sets the search window size for tracking moved feature points.

Larger windows increase the processing time and reduce the accuracy, but allow larger motion vectors and may be more stable in noisy images. Use OptFlowLKSetNumIterationsPerPyramidLevel for longer motion vectors and high accuracy.

Attention: When setting new values for this property the optical flow state is reset (two new images need to be set until a result is available)!

Parameters

[in]	OpticalFlowHandle	Handle of the optical flow state object.
[in]	WindowSize	Search window size.

Returns: A value >= 0 indicates success. All other return values are taken from the definitions in CVCError.h


Contact Legal Privacy Statement	Copyright © 2023.
	All rights reserved.

Functions

Detailed Description

Function Documentation

◆ OptFlowCreateColorfulOrientationPlot()

◆ OptFlowLKCalculate()

◆ OptFlowLKCalculateSeq()

◆ OptFlowLKCreate()

◆ OptFlowLKGetDistanceThreshold()

◆ OptFlowLKGetFeatureThreshold()

◆ OptFlowLKGetNumGaussPyramidLevels()

◆ OptFlowLKGetNumIterationsPerPyramidLevel()

◆ OptFlowLKGetWindowSize()

◆ OptFlowLKRecalculateFeaturePoints()

◆ OptFlowLKSetDistanceThreshold()

◆ OptFlowLKSetFeatureSeparation()

◆ OptFlowLKSetFeatureThreshold()

◆ OptFlowLKSetNumGaussPyramidLevels()

◆ OptFlowLKSetNumIterationsPerPyramidLevel()

◆ OptFlowLKSetWindowSize()