ebl::camera< Tdata > Class Template Reference

#include <camera.h>

Inheritance diagram for ebl::camera< Tdata >:

List of all members.

Public Member Functions
	camera (int height=-1, int width=-1, std::ostream &out=std::cout, std::ostream &err=std::cerr)
virtual	~camera ()
	Destructor.
virtual idx< Tdata >	grab ()=0
	Return a new frame.
virtual string	grab_filename ()
virtual void	next ()
virtual void	previous ()
	Move to the previous frame, without returning the frame.
virtual bool	empty ()
	Return true if no frames available, false otherwise.
virtual void	skip (uint n)
	Skip n frames.
virtual void	set_input_narrow (int dim, int size, int off)
virtual void	set_grayscale ()
	This will force the camera to output grayscale images.
virtual bool	start_recording (uint window_id=0, const char *name=NULL)
virtual bool	record_frame ()
	Dump all frames declared to be recorded by start_recording().
virtual bool	stop_recording (float fps, const char *root=NULL)
virtual float	fps ()
	Return the number of frames per second obtained via grab().
virtual uint	frame_id ()
	Return the id for current frame.
virtual string	frame_name ()
	Return a name for current frame.
virtual string	frame_fullname ()
	Return the complete filename for current frame (including path).
virtual int	remaining ()
	Return the number of frames left to process, -1 if unknown.
virtual int	size ()
Protected Member Functions
virtual idx< Tdata >	postprocess ()
Protected Attributes
idx< Tdata >	frame
	frame buffer
int	height
	resize input
int	width
	resize input
bool	bresize
	resize or not
bool	mresize
	mean resize or not
uint	resize_mode
	resize mode: 0, 1 or 2
uint	frame_id_
	frame counter
string	frame_name_
	frame name
bool	grabbed
	false if no frame grabbed yet
uint	wid
	window to record from
string	recording_name
	name of video
uint	record_cnt
	frame counter for recording
float	fps_grab
	frames per second of grabbing
string	audio_filename
	filename of audio file
std::ostream &	out
	stream where to output info
std::ostream &	err
	stream where to output info
timer	tfps
	timer for computing fps.
long	fps_ms_elapsed
	ms elapsed since last reset
uint	cntfps
	counter for computing fps.
int	narrow_dim
int	narrow_size
int	narrow_off
	Narrow parameters.
bool	grayscale
	If true, output grayscale images.

---

Detailed Description

template<typename Tdata>
class ebl::camera< Tdata >

The camera class is an abstract class that serves as an interface to different camera implementation, such as camera_opencv, etc. It allows to grab images from camera in the idx format, and also to save gui outputs into video files.

---

Constructor & Destructor Documentation

template<typename Tdata >

ebl::camera< Tdata >::camera	(	int	height = -1,
		int	width = -1,
		std::ostream &	out = std::cout,
		std::ostream &	err = std::cerr
	)

Initialize a camera. height and width are optional parameters that resize the input image to those dimensions if given (different than -1). One may want to decrease the input resolution first to speed up operations, for example when computing multiple resolutions.

Parameters:

height	Resize input frame to this height if different than -1.
width	Resize input frame to this width if different than -1.

---

Member Function Documentation

template<typename Tdata >

string ebl::camera< Tdata >::grab_filename

(

)

[virtual]

Do not read the file, instead return the filename to be grabbed. This method should only be used by file-based cameras.

Reimplemented in ebl::camera_directory< Tdata >.

template<typename Tdata >

void ebl::camera< Tdata >::next

(

)

[virtual]

Move to the next frame, without returning the frame. This is called by grab before grabbing. This can be used to get frames infos without grabbing.

Reimplemented in ebl::camera_directory< Tdata >.

template<typename Tdata >

idx< Tdata > ebl::camera< Tdata >::postprocess

(

)

[inline, protected, virtual]

Return post processed frame after grabbing it, e.g. resize frame to target height and width (if specified). This also increments frame counter.

template<typename Tdata >

void ebl::camera< Tdata >::set_input_narrow	(	int	dim,
		int	size,
		int	off
	)		[virtual]

This will narrow all images coming out of the camera in dimension 'dim' with size 'size' starting at offset 'off'.

template<typename Tdata >

int ebl::camera< Tdata >::size

(

)

[virtual]

Return the total number of frames to process from the initialization, of the camera, -1 if unknown.

Reimplemented in ebl::camera_directory< Tdata >.

template<typename Tdata >

bool ebl::camera< Tdata >::start_recording	(	uint	window_id = 0,
		const char *	name = NULL
	)		[virtual]

Start recording of frames from window window_id into path. This creates a directory name in path. No frames are actually recorded,

template<typename Tdata >

bool ebl::camera< Tdata >::stop_recording	(	float	fps,
		const char *	root = NULL
	)		[virtual]

Create all videos started with start_recording() using frames dumped with record_frame(), using fps frames per second.

Parameters:

root	Indicates the root where the images are and where to save the video.

---

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

• /home/rex/ebltrunk/tools/libeblearntools/include/camera.h
• /home/rex/ebltrunk/tools/libeblearntools/include/camera.hpp