#include <img.h>

Include dependency graph for imgfilter.h:

This graph shows which files directly or indirectly include this file:

Functions
int	imgFillGaussKernel (float **kernel, float stdev, int size)
int	imgConvolute2D (float *data, float buffer, int frame, int width, int height, float *kernel, int size, int border, int verbose, char errmsg)
int	imgGaussianFilter (IMG img, float gauss_sd, int size, int border, int verbose, char errmsg)
float **	mallocMatrix (float w, float h)
float **	imgGaussKernel (int size)
	Make a gaussian convolution kernel.
void	imgFreeKernel (float **kernel, int size)
	Free a convolution kernel.
void	imgConvoluteData (float *data, float buffer, int frame, int width, int height, float **kernel, int size)
	Perform a convolution operation on float data.
int	imgConvolute (IMG img, int frame, int plane, float *kernel, int size)
	Perform a convolution operation on image.

Function Documentation

int imgConvolute	(	IMG *	img,
		int	frame,
		int	plane,
		float **	kernel,
		int	size
	)

Perform a convolution operation on image.

The bordering size/2 pixels of the image do not containg valid data. imgConvolute is quite memory efficient. Only (img->dimx+size)*size*sizeof(float) bytes are consumed by a temporary buffer. When processing multiple planes/frames/images with equal width and kernel size, efficiency can be improved further by allocating the temporary buffer yourself and calling imgConvoluteData directly.

Parameters:

img	image
frame	frame to perform operation on
plane	plane to perform operation on
kernel	convolution kernel
size	convolution kernel size

Precondition:: size must be smaller than img->dimx and img->dimy

Returns:: nonzero in case of (malloc) error

Definition at line 491 of file imgfilter.c.

References imgConvoluteData(), and mallocMatrix().

Here is the call graph for this function:

int imgConvolute2D	(	float ***	data,
		float **	buffer,
		int	frame,
		int	width,
		int	height,
		float **	kernel,
		int	size,
		int	border,
		int	verbose,
		char *	errmsg
	)

Perform a convolution operation on one image matrix.

Returns:: Returns 0 when successful, otherwise <>0.

Parameters:

data	Image matrix data[height][width][frame]
buffer	Temporary preallocated memory buffer[size][width+size-1]
frame	frame [0..dimt] of data matrix which is processed
width	Width of image matrix (dimx)
height	Height of image matrix (dimy)
kernel	Convolution kernel[size][size] matrix
size	Convolution kernel size; must be an odd number and >=3
border	Fill borders with zero (0) or with closest image pixel (<>0)
verbose	Verbose level; if zero, then only warnings are printed into stderr
errmsg	Pointer to error message, at least 128 characters; NULL, if not needed

Definition at line 101 of file imgfilter.c.

Referenced by imgGaussianFilter().

void imgConvoluteData	(	float ***	data,
		float **	buffer,
		int	frame,
		int	width,
		int	height,
		float **	kernel,
		int	size
	)

Perform a convolution operation on float data.

Parameters:

data	data in format float[y][x][frame]
buffer	temporary buffer to use. Size must be float[size][data width+size]
frame	frame to extract from data
width	width of the image
height	height of the image
kernel	convolution kernel
size	convolution kernel size

Definition at line 418 of file imgfilter.c.

Referenced by imgConvolute().

int imgFillGaussKernel	(	float **	kernel,
		float	stdev,
		int	size
	)

Make a gaussian convolution kernel.

Returns:: Returns 0 when successful, otherwise <>0.

Parameters:

kernel	Gaussian convolution kernel matrix[size][size], filled here
stdev	Gaussian S.D. in pixels (decimals are ok)
size	Kernel dimension

Definition at line 52 of file imgfilter.c.

Referenced by imgGaussianFilter().

void imgFreeKernel	(	float **	kernel,
		int	size
	)

Free a convolution kernel.

Parameters:

kernel	the convolution kernel
size	size of the kernel

Definition at line 401 of file imgfilter.c.

int imgGaussianFilter	(	IMG *	img,
		float	gauss_sd,
		int	size,
		int	border,
		int	verbose,
		char *	errmsg
	)

Apply Gaussian filter to whole dynamic image in IMG struct.

Returns:: If an error is encountered, function returns a non-zero value. Otherwise 0 is returned.

Parameters:

img	Image data to be processed; data is overwritten with filtered image
gauss_sd	Gaussian S.D. in pixels (decimals are ok)
size	Gaussian kernel size (must be odd number); typically 6*gauss_sd is sufficient
border	In filtering, fill borders with zero (0) or with closest image pixel (<>0)
verbose	Verbose level; if zero, then only warnings are printed into stderr
errmsg	Pointer to error message, at least 128 characters; NULL, if not needed

Definition at line 251 of file imgfilter.c.

References imgConvolute2D(), and imgFillGaussKernel().

Here is the call graph for this function:

float** imgGaussKernel ( int size )

Make a gaussian convolution kernel.

Parameters:

size	kernel size

Returns:: the convolution kernel. NULL on error

Definition at line 376 of file imgfilter.c.

References mallocMatrix().

Here is the call graph for this function:

float** mallocMatrix	(	float	w,
		float	h
	)

Definition at line 352 of file imgfilter.c.

Referenced by imgConvolute(), and imgGaussKernel().

Functions

Function Documentation