FlowCalculatorCpu Class Reference

The FlowCalculatorCpu class is a hardware specific implementation of FlowCalculator. The algorithm takes two pictures and calculates a flow matrix using the Farneback algorithm. Than the flow matrix is evalutated and reduced to one single flow value. More...

#include <flowcalculatorcpu.h>

Inheritance diagram for FlowCalculatorCpu:

Public Member Functions
	FlowCalculatorCpu (int numLevels=5, double pyrScale=0.5, bool fastPyramids=false, int winSize=13, int numIters=10, int polyN=5, double polySigma=1.1, int flags=0)
	FlowCalculatorCpu Constructor creates a Farneback object which is used to execute the farneback algortihm later.
	~FlowCalculatorCpu ()
	FlowCalculatorCpu Destructor cleans up left over data and pointers.
double	calculateFlow (cv::Mat fromMat, cv::Mat toMat) override
	calculateFlow computes the flow between two images and reduces the resulting flow-matrix to a single floating point value
	FlowCalculatorCpu (int numLevels=5, double pyrScale=0.5, bool fastPyramids=false, int winSize=13, int numIters=10, int polyN=5, double polySigma=1.1, int flags=0)
	FlowCalculatorCpu Constructor creates a Farneback object which is used to execute the farneback algortihm later.
	~FlowCalculatorCpu ()
	FlowCalculatorCpu Destructor cleans up left over data and pointers.
double	calculateFlow (cv::Mat fromMat, cv::Mat toMat) override
	calculateFlow computes the flow between two images and reduces the resulting flow-matrix to a single floating point value

Additional Inherited Members
Protected Member Functions inherited from FlowCalculator
double	flowMatToDouble (cv::Mat mat)
	flowMatToDouble reduces a flow-matrix to a single floating point value, if that mat was empty return -1.0
double	flowMatToDouble (cv::Mat mat)
	flowMatToDouble reduces a flow-matrix to a single floating point value, if that mat was empty return -1.0

Detailed Description

The FlowCalculatorCpu class is a hardware specific implementation of FlowCalculator. The algorithm takes two pictures and calculates a flow matrix using the Farneback algorithm. Than the flow matrix is evalutated and reduced to one single flow value.

Author

Dominic Zahn

Date

2022/04/12

Constructor & Destructor Documentation

FlowCalculatorCpu() [1/2]

FlowCalculatorCpu::FlowCalculatorCpu	(	int	numLevels = 5,
		double	pyrScale = 0.5,
		bool	fastPyramids = false,
		int	winSize = 13,
		int	numIters = 10,
		int	polyN = 5,
		double	polySigma = 1.1,
		int	flags = 0
	)

FlowCalculatorCpu Constructor creates a Farneback object which is used to execute the farneback algortihm later.

Parameters

numLevels	number of pyramid layers including the initial image; levels=1 means that no extra layers are created and only the original images are used
pyrScale	parameter, specifying the image scale (<1) to build pyramids for each image; pyr_scale=0.5 means a classical pyramid, where each next layer is twice smaller than the previous one.
fastPyramids	uses a faster calculation for
winSize	averaging window size; larger values increase the algorithm robustness to image noise and give more chances for fast motion detection, but yield more blurred motion field.
numIters	number of iterations the algorithm does at each pyramid level
polyN	size of the pixel neighborhood used to find polynomial expansion in each pixel; larger values mean that the image will be approximated with smoother surfaces, yielding more robust algorithm and more blurred motion field, typically poly_n =5 or 7.
polySigma	standard deviation of the Gaussian that is used to smooth derivatives used as a basis for the polynomial expansion; for poly_n=5, you can set poly_sigma=1.1, for poly_n=7, a good value would be poly_sigma=1.5.
flags	- OPTFLOW_USE_INITIAL_FLOW uses the input flow as an initial flow approximation. OPTFLOW_FARNEBACK_GAUSSIAN uses the Gaussian \texttt{winsize}\times\texttt{winsize} filter instead of a box filter of the same size for optical flow estimation

FlowCalculatorCpu() [2/2]

FlowCalculatorCpu::FlowCalculatorCpu	(	int	numLevels = 5,
		double	pyrScale = 0.5,
		bool	fastPyramids = false,
		int	winSize = 13,
		int	numIters = 10,
		int	polyN = 5,
		double	polySigma = 1.1,
		int	flags = 0
	)

FlowCalculatorCpu Constructor creates a Farneback object which is used to execute the farneback algortihm later.

Parameters

numLevels	number of pyramid layers including the initial image; levels=1 means that no extra layers are created and only the original images are used
pyrScale	parameter, specifying the image scale (<1) to build pyramids for each image; pyr_scale=0.5 means a classical pyramid, where each next layer is twice smaller than the previous one.
fastPyramids	uses a faster calculation for
winSize	averaging window size; larger values increase the algorithm robustness to image noise and give more chances for fast motion detection, but yield more blurred motion field.
numIters	number of iterations the algorithm does at each pyramid level
polyN	size of the pixel neighborhood used to find polynomial expansion in each pixel; larger values mean that the image will be approximated with smoother surfaces, yielding more robust algorithm and more blurred motion field, typically poly_n =5 or 7.
polySigma	standard deviation of the Gaussian that is used to smooth derivatives used as a basis for the polynomial expansion; for poly_n=5, you can set poly_sigma=1.1, for poly_n=7, a good value would be poly_sigma=1.5.
flags	- OPTFLOW_USE_INITIAL_FLOW uses the input flow as an initial flow approximation. OPTFLOW_FARNEBACK_GAUSSIAN uses the Gaussian \texttt{winsize}\times\texttt{winsize} filter instead of a box filter of the same size for optical flow estimation

Member Function Documentation

calculateFlow() [1/2]

double FlowCalculatorCpu::calculateFlow ( cv::Mat  fromMat, cv::Mat  toMat  )

overridevirtual

calculateFlow computes the flow between two images and reduces the resulting flow-matrix to a single floating point value

Parameters

fromMat	cv::Mat which is the image before the posible movement
toMat	cv::Mat which is the image after the posible movement

Returns

doubel which represents the movement between the images

Implements FlowCalculator.

calculateFlow() [2/2]

double FlowCalculatorCpu::calculateFlow ( cv::Mat  fromMat, cv::Mat  toMat  )

overridevirtual

calculateFlow computes the flow between two images and reduces the resulting flow-matrix to a single floating point value

Parameters

fromMat	cv::Mat which is the image before the posible movement
toMat	cv::Mat which is the image after the posible movement

Returns

doubel which represents the movement between the images

Implements FlowCalculator.

---

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

• iVS3D/src/iVS3D-smoothCameraMovementPlugin/flowcalculatorcpu.h
• iVS3D/src/iVS3D-stationaryCameraPlugin/flowcalculatorcpu.h
• iVS3D/src/iVS3D-smoothCameraMovementPlugin/flowcalculatorcpu.cpp
• iVS3D/src/iVS3D-stationaryCameraPlugin/flowcalculatorcpu.cpp