calculate optical flow with EdgeFlow More...

#include <lib/vision/edge_flow.h>

Include dependency graph for edge_flow.c:

Functions
void	calc_previous_frame_nr (struct opticflow_result_t result, struct opticflow_t opticflow, uint8_t current_frame_nr, uint8_t previous_frame_offset, uint8_t previous_frame_nr)
	Calc_previous_frame_nr; adaptive Time Horizon. More...

void	calculate_edge_histogram (struct image_t *img, int32_t edge_histogram[], char direction, uint16_t edge_threshold)
	Calculate a edge/gradient histogram for each dimension of the image. More...

void	calculate_edge_displacement (int32_t edge_histogram, int32_t edge_histogram_prev, int32_t *displacement, uint16_t size, uint8_t window, uint8_t disp_range, int32_t der_shift)
	Calculate_displacement calculates the displacement between two histograms. More...

uint32_t	getMinimum (uint32_t *a, uint32_t n)
	Calculate minimum of an array. More...

void	line_fit (int32_t displacement, int32_t divergence, int32_t *flow, uint32_t size, uint32_t border, uint16_t RES)
	Fits a linear model to an array with pixel displacements with least squares. More...

void	draw_edgeflow_img (struct image_t img, struct edge_flow_t edgeflow, int32_t edge_hist_x_prev, int32_t *edge_hist_x)
	Draws edgehistogram, displacement and linefit directly on the image for debugging (only for edgeflow in horizontal direction!!) More...

uint32_t	getAmountPeaks (int32_t *edgehist, int32_t thres, int32_t size)
	getAmountPeaks, calculates the amount of peaks in a edge histogram More...

Detailed Description

calculate optical flow with EdgeFlow

Edge-histogram matching, implementation by K. N. McGuire Publication: Local Histogram Matching for Efficient Optical Flow Computation Applied to Velocity Estimation on Pocket Drones by K.N. McGuire et al. (2016), ICRA 2016

Definition in file edge_flow.c.

Function Documentation

void calc_previous_frame_nr	(	struct opticflow_result_t *	result,
		struct opticflow_t *	opticflow,
		uint8_t	current_frame_nr,
		uint8_t *	previous_frame_offset,
		uint8_t *	previous_frame_nr
	)

Calc_previous_frame_nr; adaptive Time Horizon.

Parameters

[in]	*opticflow	The opticalflow structure
[in]	*result	The optical flow result
[in]	*current_frame_nr	The current frame number of the circular array of the edge_hist struct
[in]	*previous_frame_offset	previous frame offset of how far the method should compare the edgehistogram
[out]	*previous_frame_nr	previous frame index of the edgehist struct

Definition at line 39 of file edge_flow.c.

References opticflow_result_t::flow_x, opticflow_result_t::flow_y, MAX_HORIZON, and opticflow_t::search_distance.

Referenced by calc_edgeflow_tot().

Here is the caller graph for this function:

void calculate_edge_displacement	(	int32_t *	edge_histogram,
		int32_t *	edge_histogram_prev,
		int32_t *	displacement,
		uint16_t	size,
		uint8_t	window,
		uint8_t	disp_range,
		int32_t	der_shift
	)

Calculate_displacement calculates the displacement between two histograms.

Parameters

[in]	*edge_histogram	The edge histogram from the current frame_step
[in]	*edge_histogram_prev	The edge histogram from the previous frame_step
[out]	*displacement	array with pixel displacement of the sequential edge histograms
[in]	size	Indicating the size of the displacement array
[in]	window	Indicating the search window size
[in]	disp_range	Indicating the maximum disparity range for the block matching
[in]	der_shift	The pixel shift estimated by the angle rate of the IMU

Definition at line 168 of file edge_flow.c.

References D, DISP_RANGE_MAX, and getMinimum().

Referenced by calc_edgeflow_tot().

Here is the call graph for this function:

Here is the caller graph for this function:

void calculate_edge_histogram	(	struct image_t *	img,
		int32_t	edge_histogram[],
		char	direction,
		uint16_t	edge_threshold
	)

Calculate a edge/gradient histogram for each dimension of the image.

Parameters

[in]	*img	The image frame to calculate the edge histogram from
[out]	*edge_histogram	The edge histogram from the current frame_step
[in]	direction	Indicating if the histogram is made in either x or y direction
[in]	edge_threshold	A threshold if a gradient is considered a edge or not

Definition at line 88 of file edge_flow.c.

References image_t::buf.

Referenced by calc_edgeflow_tot().

Here is the caller graph for this function:

void draw_edgeflow_img	(	struct image_t *	img,
		struct edge_flow_t	edgeflow,
		int32_t *	edge_hist_x_prev,
		int32_t *	edge_hist_x
	)

Draws edgehistogram, displacement and linefit directly on the image for debugging (only for edgeflow in horizontal direction!!)

Parameters

[out]	*img	The image structure where will be drawn on
[in]	edgeflow	Information structure for flow information
[in]	Displacement	Pixel wise Displacement array
[in]	*edge_hist_x	Horizontal edge_histogram

Definition at line 297 of file edge_flow.c.

References edge_flow_t::div_x, edge_flow_t::flow_x, image_t::h, image_draw_line(), image_t::w, point_t::x, and point_t::y.

Referenced by calc_edgeflow_tot().

Here is the call graph for this function:

Here is the caller graph for this function:

uint32_t getAmountPeaks	(	int32_t *	edgehist,
		int32_t	thres,
		int32_t	size
	)

getAmountPeaks, calculates the amount of peaks in a edge histogram

Parameters

[in]	*edgehist	Horizontal edge_histogram
[in]	thres	The threshold from which a peak is considered significant peak or not
[in]	size	Size of the array
	return]	amount of peaks

Definition at line 337 of file edge_flow.c.

Referenced by calc_edgeflow_tot().

Here is the caller graph for this function:

uint32_t getMinimum	(	uint32_t *	a,
		uint32_t	n
	)

Calculate minimum of an array.

Parameters

[in]	*a	Array containing values
[in]	*n	The size of the array

Returns: The index of the smallest value of the array

Definition at line 220 of file edge_flow.c.

Referenced by calculate_edge_displacement().

Here is the caller graph for this function:

void line_fit	(	int32_t *	displacement,
		int32_t *	divergence,
		int32_t *	flow,
		uint32_t	size,
		uint32_t	border,
		uint16_t	RES
	)

Fits a linear model to an array with pixel displacements with least squares.

Parameters

[in]	*displacements	Array with Pixel Displacements
[out]	*divergence	Global divergence of pixel displacements
[out]	*flow	Global translational flow from pixel displacements
[in]	*size	Size of displacement array
[in]	border	A border offset of the array that should not be considerd for the line fit
[in]	RES	Resolution to be used for the integer based linefit

Definition at line 247 of file edge_flow.c.

Referenced by calc_edgeflow_tot().

Here is the caller graph for this function:

Functions

Detailed Description

Function Documentation